High Transdominant Revm10 Protein Levels Are Required to Inhibit HIV-1 Replication in Cell Lines and Primary T Cells: Implication for Gene Therapy of AIDS

High transdominant RevM10 protein levels are required to inhibit HIV-1 replication in cell lines and primary T cells: implication for gene therapy of AIDS

I Plavec1, M Agarwal1,KEHo2, M Pineda1, J Auten1, J Baker1, H Matsuzaki3, S Escaich4, M Bonyhadi1 and E Bo¨ hnlein1 1Progenesys Program, SyStemix Inc, 3155 Porter Drive, Palo Alto, CA 94304, USA

Expression of antiviral genes in CD4+ T cells has been pro- uniformly higher than from internal promoters (eg CMV, posed as a strategy for gene therapy of AIDS. Over the PGK). Analysis of selected vectors in acutely and chron- past years, we and others have developed retroviral vec- ically HIV-infected cell lines suggested that threshold levels tors encoding the RevM10 protein, a dominant-negative of RevM10 expression are required to achieve inhibition of mutant of the HIV-1 Rev trans-activator protein. We could HIV replication. LTR-driven RevM10 expression also demonstrate gene transfer and inhibition of HIV-1 repli- yielded high steady-state protein levels in activated primary cation in cultured T cell lines and primary T cells. However, T cells resulting in inhibition of HIV replication, and there little is known about the levels of the antiviral protein was no apparent difference between the MoMLV, MPSV required to achieve a therapeutic effect, particularly in pri- and MESV-LTR vectors. However, RevM10 expression mary cells. In this report, we compare different vector was down-regulated in resting primary cells and conse- designs with regard to expression of the antiviral gene to quently anti-HIV efﬁcacy was signiﬁcantly reduced. Taken develop an optimal vector for clinical applications. Our together, the data suggest that relatively high steady-state results demonstrate that intracellular steady-state RevM10 levels of RevM10 protein are required to achieve inhibition protein levels expressed from the Moloney murine leuke- of HIV replication and that the MPSV- and MESV-derived mia virus (MoMLV), myeloproliferative sarcoma virus retroviral vectors show no advantage over the MoMLV- (MPSV) or mouse embryonic stem cell virus (MESV) pro- based vectors for expression of anti-HIV genes in human moters located in the long terminal repeat (LTR) were T cells.

Keywords: gene therapy; HIV-1; retroviral vectors; transgene expression; T cells

Introduction of Rev function.9 Furthermore, a number of studies have demonstrated that RevM10 can inhibit viral replication in Human immunodeficiency virus type-1 (HIV-1) encodes HIV-infected T cell lines and primary T cells.10–15 several regulatory genes including Rev (reviewed in Refs Intracellular expression of genes which can suppress 1–3). Through binding to a structured RNA sequence, HIV-1 replication has been proposed as an alternative to termed the Rev-responsive element (RRE), Rev promotes standard therapies for the treatment of AIDS.16,17 A large 4,5 the export of RRE-containing mRNAs from the nucleus number of antiviral genes including RNA decoys, anti- leading to cytoplasmic accumulation of incompletely sense RNA, transdominant mutants (including RevM10), spliced HIV-1 transcripts. First, these transcripts serve as ribozymes and intracellular antibodies (reviewed in Ref. templates for translation of viral proteins gag, pol and env. 16) have been shown to inhibit effectively HIV-1 in vitro. Second, the nonspliced genomic mRNA is needed for The target cell population for delivery of anti-HIV genes infectious particle formation. HIV-1 Rev function requires could be peripheral blood mononuclear cells,18,19 or alter- 6–8 interaction with one or more cellular host proteins. natively hematopoietic stem cells.16 Retroviral vectors are RevM10, a mutant of Rev retains the ability to bind to presently the most widely used system to deliver genes RRE but it does not transactivate, possibly because of its into those cells.20 Irrespective of the target cell popu- 8,9 inability to interact with host factor(s). When co- lation, the success of gene therapy will depend on both expressed with wild-type Rev protein in transient trans- efficient gene transfer and persistent high-level gene fection assays, RevM10 acts as a transdominant inhibitor expression. Initially, efforts in developing retroviral vectors were mainly focused on improving virus titers to increase gene transfer and only recently more work is Correspondence: E Bo¨hnlein 2 being invested in optimizing vectors for efficient trans- Present addresses: Genentech Inc, 460 Point San Bruno Boulevard, South gene expression (reviewed in Refs 21–23). A wide variety San Francisco, California 94080-4990, USA; 3Affymetrix, 3380 Central Expressway, Santa Clara, California 95051, USA; 4Rhoˆne-Poulenc Rorer, of vector constructions have been described including 94403 Vitry-sur-Seine, France different promoters (retroviral LTRs, internal promoters), Received 15 August 1996; accepted 29 October 1996 orientation of the gene of interest (forward, reverse), self- RevM10 expression in cell lines and primary T cells I Plavec et al 129 inactivating LTR vectors and other LTR modifications, driven transcript (Figure 2, lower left panel) resulting in and insertion of an expression cassette into the LTR RevM10 protein levels approximately 10-fold higher than (double-copy vectors) (reviewed in Refs 21, 23). Despite internal promoter vectors (Figure 3). No apparent differ- all these studies, the optimal vector design for the ence between the MoMLV-LTR and MPSV-LTR vectors expression of a particular gene needs to be determined could be detected in CEMSS cells and the different empirically and the final result may depend on the com- primer binding sites also had no measurable impact. To bination of both gene and vector sequences.24–26 further improve expression of the LTR-RevM10 vectors, As part of our efforts to develop clinical protocols for the tk-Neo sequence was replaced by the SV40-Puro tran- gene therapy of AIDS we have tested a number of retro- scription unit (LMSPL vector; Figure 1). This modifi- viral vector designs to optimally express the transdomi- cation resulted in a two- to three-fold increase in RevM10 nant RevM10 protein. In particular, we carefully ana- protein levels in CEMSS cells (compare LMTNL with lyzed vector expression in primary T cells at various LMSPL, Figure 3). stages of activation. We also evaluated the anti-HIV In all vectors (Figure 1; numbers 4–8), the packaging potency of selected vectors in chronically and acutely signal is encoded by a short retroviral sequence13,36 and HIV-infected cell lines as well as in primary T cells. The the RevM10 protein is synthesized from an unspliced results suggest that the extent of inhibition of HIV-1 repli- full-length messenger RNA. The presence of an intron in cation is a function of the relative expression of RevM10 a transcription unit can improve the expression of some protein in those cells, supporting the hypothesis that cDNAs.37 To test this possibility, we constructed RevM10 RevM10 acts as an intracellular competitive inhibitor.9 vectors based on the LXSN38 (construct LMSPL-INT1; Implications of these findings for the development of Figure 1, No. 9) and MFG design39 (construct LMSPL- effective retrovirus vector-based anti-HIV gene therapies INT2; Figure 1, No. 10) in which protein expression relies are discussed. on the efficient splicing of the primary transcript.39,40 The ratio of spliced to unspliced transcripts was higher for the LMSPL-INT2 vector compared to the LMSPL-INT1 Results (Figure 2, lower right panel). This finding was not sur- prising since the splicing in the MFG vector had been Retroviral vector LTR-driven expression yields high optimized for protein expression.39 However, both vec- steady-state levels of RevM10 protein in transduced tors yielded similar steady-state intracellular RevM10 T cells protein levels and these were not significantly higher We started our study by testing vectors in which RevM10 when compared to the intronless LMSPL vector (Figure expression is driven by an internal promoter using the 2). Thus, inclusion of an intron did not further improve immediate–early promoter of the human cytomegalo- RevM10 protein expression. An unexpected short Rev- virus27 (LNCML vector; Figure 1) and the mouse phos- specific transcript was detected in the LMSPL RNA phoglycerate kinase gene promoter28 (LNMPL and (Figure 2). This RNA is a splice product between the LPMPL vectors; Figure 1). Human CEMSS T lymphoid splice donor in the retroviral packaging sequence and a cells were transduced with amphotropic retroviral vec- splice acceptor located in the RevM10 coding sequence tors harvested from PA317-based producer cells and and does not encode functional RevM10 protein (data not selected either with G418 or puromycin to yield homo- shown). As expected, a 10-fold higher virus titer was genous populations of transduced cells. RevM10 protein obtained with the LMSPL-INT1 vector because of the and RNA expression were analyzed by Rev-specific extended packaging sequence38 (Figure 1). immunofluorescence29 and Northern blotting, respectively. Regardless of the internal promoter used, low RevM10 protein threshold level is required for inhibition RevM10 RNA expression levels were observed in trans- of HIV-1 replication in chronically infected T cell lines duced cells (Figure 2) which resulted in intracellular We have used an assay system based on chronically HIV- RevM10 protein levels barely detectable above back- infected A3.01/ProHygro T cells to compare the anti-HIV ground (Figure 3). Since it has been reported that the neo- efficacy of the different vectors expressing various levels mycin phosphotransferase gene (Neo) gene sequence can of the RevM10 protein.13 The A3.01/ProHygro cells were act as a transcriptional silencer,30 we attempted to transduced by co-cultivation with the PA317-derived improve RevM10 expression by replacing the Neo- retroviral vector producer cells followed by selection in resistance gene with the puromycin-resistance gene G418 and hygromycin. Established drug-resistant cell (Puro) (LPMPL vector; Figure 1). This modification populations were analyzed for RevM10 protein increased intracellular RevM10 RNA (Figure 2) and pro- expression and for HIV-1 production kinetics. As nega- tein steady-state levels in CEMSS cells slightly, but the tive control, we used vectors LNC⌬ML, and L⌬MTNL overall expression was still low (Figure 3). which encode a mutated RevM10 gene (⌬M10) lacking Next we designed a series of vectors which use the the methionine initiation codon.13 A3.01/ProHygro cells retroviral LTR promoter to express the RevM10 gene produce relatively high steady-state levels of wild-type (Figure 1). The MoMLV LTR enhancer and tRNA primer Rev protein13 (Table 1, compare mean fluorescence values binding site (PBS) have been implicated as possible nega- in A3.01 and A3.01/ProHygro cells). Since the intracellu- tive elements for LTR-driven expression.31–33 For these lar Rev protein assay does not distinguish the wild-type reasons, we also generated vectors containing the MPSV Rev from the mutant RevM10 protein we have estimated LTR34 and the primer binding site derived from the the relative expression of the RevM10 protein by compar- dl587rev virus35 (Figure 1, vectors 4–7). All four vectors ing the total Rev antigen signals in vector-transduced and were stable and had similar titers on NIH3T3 cells (1– untransduced A3.01/ProHygro cells (Table 1, relative 4 × 106 c.f.u./ml; Figure 1). Furthermore, these vectors RevM10 Ag). We could not detect significant RevM10 expressed comparable steady-state levels of the LTR- steady-state levels above background in cells transduced RevM10 expression in cell lines and primary T cells I Plavec et al 130

Figure 1 Retroviral vector design. The names of the retroviral vectors are indicated in the ﬁrst column. A schematic depiction of the different vectors (not to scale) is indicated in the second column. Arrows indicate initiation of transcription and orientation of the transcript. The third column indicates the G418-resistance titers on NIH-3T3 cells. The fourth column indicates whether a retroviral vector is stable as determined by transcript pattern in Northern blot analyses. The ﬁfth column indicates whether the retroviral producer cells generate replication-competent retrovirus (RCR). ND, not determined; cfu, colony forming unit; CMV, cytomegalovirus immediate–early gene promoter; PGK, phosphoglycerate kinase promoter; Puro, puromycin resistance gene; MPSV, myeloproliferative sarcoma virus; MESV, murine embryonal stem cell virus; LTR, long terminal repeat; SD, splice donor; SA, splice acceptor; Neo, neomycin-phosphotransferase gene; tk, herpes simplex virus thymidine kinase promoter; SV40, SV40 early gene enhancer/promoter.

with the LNCML vector while the LMTNL, MPSV- obtaining a minimum RevM10 threshold level was MTNL, MESV-MTNL, L/587-MTNL and LMSPL vectors required for inhibition of HIV replication in this assay produced RevM10 at levels two- to four-fold higher than system. the endogenous wild-type Rev protein (Table 1). Interest- ingly, the LMSPL vector did not produce more RevM10 Antiviral effect of RevM10 is overcome at high HIV-1 protein than the LMTNL vector, which is in contrast to inoculation dose the results in CEMSS cells (compare Figure 3 and Table In the A3.01/ProHygro assay system, HIV-1 is secreted 1). We have observed RevM10 expression patterns simi- at a constant rate and by using the retroviral vectors lar to CEMSS cells in other cells lines (PA317, U937, data described above, we were able to express different not shown). At present, we have no explanation why amounts of the RevM10 antiviral protein. To perform a these vectors yield different RevM10 antigen levels in converse experiment, we inoculated LMTNL-transduced A3.01/ProHygro cells. The antiviral effect in transduced CEMSS cells with various doses of the HIV-1 HXB-3 virus A3.01/ProHygro cells was analyzed by measuring levels (Figure 5). At higher inoculation doses, virus spread fas- of secreted HIV-1 p24 Gag protein in the culture super- ter through the cultures and reached peak replication at natants. Cells expressing detectable levels of RevM10 earlier time points. The extent of inhibition of HIV-1 rep- protein produced 75–90% less p24 Gag antigen per 106 lication in RevM10-expressing cells was inversely pro- cells than the control cultures (Figure 4, Table 1). portional to the HIV-1 inoculation dose and was com- Although the protein assay did not detect RevM10 in the pletely overcome at the highest inoculation dose MOI = LNCML-transduced cells it is likely that low levels of 0.1. These experiments demonstrate that both the relative protein were produced which were not sufficient to level of RevM10 expression and the viral burden are demonstrate any anti-HIV effect. Hence it appears that important factors that influence antiviral efficacy. RevM10 expression in cell lines and primary T cells I Plavec et al 131 healthy donors by depleting CD8+ cells using immunomagnetic beads yielding approximately 90% pure cell populations. Cells were stimulated with PHA, IL-2 and irradiated allogeneic feeder cells for 3–4 days and subsequently transduced with retroviral vectors by centrifugation.43,44 Following this protocol, we detected 4–8% Lyt-2 positive cells (data not shown) and after expansion in vitro, positive cells were further enriched to 70–80% purity using immunomagnetic beads. We have used the CD25 surface protein (low-affinity IL-2 receptor) as a marker for the activation status of T cells. Three to 5 days after stimulation with PHA, IL-2 and irradiated allogeneic feeder cells, CD25 expression was at maximum and starting to decline. By days 11–14, primary T cells ceased to proliferate and the CD25 marker was almost completely down-regulated reflecting the mitotically resting state of these cells (Figure 6b and 7). Interestingly, we have observed high Lyt-2 staining in activated T cells but transgene expression was low in resting cells. An example is shown in Figure 6b. No significant difference between the LMILy, MPSV-MILy, MESV-MILy and L/587-MILy vectors was detected. Lyt- 2 staining in fully activated T cells was comparable to CEMSS cells (data not shown). Furthermore, we analyzed in detail the expression of various vectors relative to the activation status of T cells over the period of 11 days after stimulation. The kinetics of Lyt-2 expression followed exactly the expression pattern of the CD25 marker and again there was no difference between the four retroviral vectors tested (Figure 7).

Figure 2 Northern blot analysis of transduced CEMSS cells. RNA from Low retroviral vector expression in resting primary transduced, drug-resistant CEMSS cell populations was analyzed using T cells results in reduced anti-HIV efficacy the Rev-specific oligonucleotide probe (Materials and methods). The trans- To determine the implication of this fluctuating transgene ducing vectors are indicated on top of each lane (see Figure 1 for more ⌬ information). The genomic size transcripts (LTR) and internal Rev-enco- expression on anti-HIV efficacy, LMILy and L MILy- ding mRNAs (CMV, PGK) are indicated by arrows. Analysis of splicing transduced cells were collected at day 5 and day 8 after vectors (middle right panel) yields two bands per RNA, dots indicate the stimulation and inoculated with the HIV-1 JR-CSF strain. transcript encoding RevM10. The lower panels show rehybridization with Viral replication was monitored over a period of 9 days the cellular ß-actin gene as a control for the amounts of mRNA loaded (Figure 8). To maintain the activated status of T cells dur- per lane. ing the HIV-1 infection experiment, a portion of day 5 harvested cells were given extra feeder cells 72 h after Expression of retroviral vectors is down-regulated in inoculation with HIV (‘day 5 restimulated’). The addition resting primary T cells of feeder cells did not influence HIV replication in the In contrast to cell lines, a majority of primary T cells from cultures (compare ‘day 5’ and ‘day 5 restimulated’ peripheral blood (PBL) are normally mitotically quie- L⌬MILy samples). In fully activated LMILy-expressing scent. We were interested in analyzing expression of the cells (‘day 5 restimulated’), HIV-1 replication was LMTNL, MPSV-MTNL, MESV-MTNL, and L/587-MTNL inhibited between day 5 and 7 and the extent of virus- vectors in PBLs at various stages of activation and pro- inhibition (approximately 1 log) was comparable to the liferation (resting and activated PBLs). We modified the inhibition observed in transduced CEMSS cells (Figure vectors by replacing the drug-resistance neomycin phos- 5). In contrast, the RevM10-mediated anti-HIV effect was photransferase gene with the Lyt-2 ␣′ surface marker41 significantly reduced when transgene expression was not (Figure 6a, vectors LMILy, MPSV-MILy, MESV-MILy, maintained (compare ‘day 5’ to ‘day 5 restimulated’, Fig- L/587-MILy). In this vector design, both genes (RevM10 ure 8). Cells harvested on day 8 after stimulation did not and Lyt-2) are encoded by a single bicistronic mRNA and support virus replication, in agreement with previous the translation of the Lyt-2 is mediated by the internal observations that HIV-1 does not replicate well in non- ribosomal entry site (IRES) of the human encephalomy- activated CD25− T cells.45,46 ocarditis virus (EMCV).42 Double-staining of transduced CEMSS and primary T cells for RevM10 and Lyt-2 Discussion showed that the expression of the two proteins is colinear (data not shown). Therefore, Lyt-2 surface antigen The focus of this study was to identify the optimal retro- expression directly correlates with expression of the viral vector design to support maximal RevM10 protein RevM10 protein, and flow cytometric analysis of the sur- expression in relevant target cells for HIV-1 infection. We face antigen can be used as a surrogate marker for have tested two basic vector structures: one design where RevM10 expression. RevM10 is expressed from internal promoters and CD4+ T cells were enriched from PBLs from normal another based on retroviral LTR-directed RevM10 RevM10 expression in cell lines and primary T cells I Plavec et al 132

Figure 3 Comparison of steady-state intracellular RevM10 protein levels. RevM10 protein expression in CEMSS cells, transduced with the retroviral vectors listed on the left was determined by Rev antigen-speciﬁc intracellular FACS analysis. The values shown (mean ﬂuorescence intensity) are average from two to three independent experiments. Bars indicate standard deviation.

Table 1 Comparison of anti-HIV efﬁcacy of various RevM10 vectors in chronically HIV-infected A3.01 cells

Cell line Mean ﬂuorescence intensity Relative RevM10/Ag % p24 reduction (total Rev/antigen)

Exp I A3.01 18 — — A3.01/ProHygro 51 — — A3.01/ProHygro LNC⌬ML 45 0.9 0 A3.01/ProHygro LNCML 41 0.8 0 A3.01/ProHygro L⌬MTNL 40 0.8 0 A3.01/ProHygro MPSV-MTNL 198 3.9 90 Ϯ 1 A3.01/ProHygro L/587-MTNL 102 2.0 84 Ϯ 2 A3.01/ProHygro MESV-MTNL 173 3.4 75 Ϯ 4

ExpII A3.01 20 — — A3.01/ProHygro 72 — — A3.01/ProHygro L⌬MTNL 70 1.0 0 A3.01/ProHygro LMTNL 281 3.9 84 Ϯ 5 A3.01/ProHygro LMSPL 275 3.8 85

Mean fluorescence intensity values for various cell lines were obtained by Rev-specific imunofluorescence staining and subsequent analysis on fluorescence-activated cell analyzer (FACScan). The values indicate total Rev antigen steady-state levels over the background fluorescence of the parental A3.01 cells. Relative RevM10 antigen is a ratio of mean fluorescence between the RevM10-vector-transduced and untransduced A3.01/ProHygro cells. Percentage p24 reduction in the RevM10-vector-transduced relative to untransduced A3.01/ProHygro cells was determined at the peak of accumulation of p24 in tissue culture supernatant (day 4; average values from two independent measurements are shown).

expression. In addition to stability and titer, the RevM10 also in the monocyte–macrophage lineage (L Su, unpub- retroviral vectors were evaluated for steady-state mRNA lished results). This ﬁnding corroborates previous studies and intracellular RevM10 protein levels. Based on our reporting inhibition of HIV-1 replication using LTR- studies, LTR-based RevM10 vectors appear to be the opti- directed RevM10 expression in retrovirus-transduced T mal design for RevM10 expression not only in T cells but cell lines and primary T cells.10,11 A study by Bahner et RevM10 expression in cell lines and primary T cells I Plavec et al 133

Figure 4 Comparison of the antiviral potency of various retroviral vectors in chronically HIV-infected A3.01/ProHygro cells. Growth kinetics of various vector- and control-transduced cells are shown in the upper panels. HIV-1 replication (lower panels) was measured as accumulation of p24 Gag protein in the culture supernatants over time. All values are average from duplicate samples. Vertical bars indicate standard errors (where not shown, the error was below resolution of the graphic program). al15 reported that a retroviral vector expressing the The panel of studied retroviral vectors enabled us to RevM10 gene from an internal CMV promoter yielded evaluate the effect of various RevM10 levels on virus rep- better antiviral efficacy in transduced CEM cells than a lication both in chronically and acutely HIV-infected LTR-RevM10 vector. The actual steady-state RevM10 pro- cells. The concurrent assessment of intracellular RevM10 tein levels were not measured in this report and other protein concentration and antiviral efficacy suggests that than different experimental conditions (HIV isolate, relatively high levels of RevM10 are required to inhibit inoculation dose, cell line), we have no explanation for HIV replication. A3.01/ProHygro cells used in this study 5 this apparent discrepancy. We have previously demon- produce high titers of infectious HIV virus (10 TCID50 strated that CMV-directed RevM10 expression can suffice on CEMSS cells)13 and express wild-type Rev protein at to block HIV-1 replication14 albeit this particular vector levels comparable to acutely infected CEMSS cells expresses significantly lower intracellular RevM10 pro- (unpublished results). Data presented here clearly sug- tein levels than the LTR-RevM10 vectors described in this gest that an excess of RevM10 protein over wild-type Rev study. It is not excluded that promoters other than the is required for significant inhibition of HIV replication, LTR could express efficiently, in particular, in primary T supporting the hypothesis that RevM10 acts as a competi- cells. Indeed, our preliminary data show that the human tive inhibitor of Rev function.9 CD4 gene promoter47 inserted in a retroviral vector HIV-1 replication was inhibited in RevM10-expressing expresses transgenes in primary T cells although at lower activated primary T cells, as previously reported.10,11 levels than the LTR promoters (unpublished results). The However, the expression of retroviral vectors was down- differences in expression between the various retroviral regulated as the cells turned quiescent resulting in vectors could in part be due to differences in proviral reduced ability to suppress HIV-1 replication. We have copy number per cell, primarily because vectors with observed similar down-regulation of expression with higher titers may infect target cells multiple times. How- vectors expressing the human NGF receptor48 surface ever, we have not observed any correlation between the marker suggesting that this result is not specific to Lyt-2 viral titer and RevM10 expression (Figure 1 and Figure or RevM10 proteins (unpublished results). Furthermore, 3). For example, the vector LMSPL had a 20-fold lower preliminary RNA analysis suggests that RevM10 titer than the LMSPL-INT1 vector (Figure 1), but pro- expression is controlled at the transcriptional level (data duced similar RevM10 protein levels (Figure 3). Con- not shown). MLV LTR-directed transgene expression versly, the LMTNL and the LMSPL vectors, which had exclusively relies on cellular transcription factors. In con- comparable virus titers (3 × 106 and 2 × 106 c.f.u. per ml, trast, cellular factors (such as NF-␬B, NF-AT) are only respectively), produced different RevM10 protein levels required for HIV-1 basal gene expression (reviewed in (Figure 3). Ref. 1). Subsequently, the HIV-1-encoded Tat transactiv- RevM10 expression in cell lines and primary T cells I Plavec et al 134

Figure 5 RevM10 antiviral efﬁcacy is dose-dependent. CEMSS transduced with the LMTNL vector were inoculated with various doses of HIV-1 HXB3 virus, as indicated at the top of each panel. Virus replication was monitored over 14–15 days. A representative experiment is shown. Results were reproduced in two to four individual experiments. Vertical bars indicate standard error (where not shown, the error was below resolution of the graphic program). MOI, multiplicity of infection.

ator protein increases HIV-1 gene expression, resulting in or mouse hematopoietic stem cells.24,49 Our data suggest high steady-state viral RNA levels and protein synthesis that in human primary T cells, MPSV or MESV-derived (reviewed in Ref. 1). A possible explanation for our vectors have no advantage over MoMLV-based vectors results is that as T cells turned quiescent, active cellular with regard to transgene expression. Furthermore, con- transcription factors became less abundant and as a sidering that MoMLV-based vectors are expressed in consequence retroviral vector expression was down- human T cells43 and monocyte–macrophages (Su et al, regulated. In contrast, even low levels of transcription manuscript submitted) derived from transduced hemato- factors could be sufficient to induce basal expression of poietic stem cells in SCID-hu mice, we propose that these HIV-1 LTR supporting Tat-mediated transactivation. Our vectors are good candidates for human stem cell-based results suggest that there is a window in time between gene therapy. MoMLV LTR-based retroviral vectors are fully activated and fully quiescent cells where HIV-1 can commonly used in preclinical experiments and have also still replicate, but recombinant retroviral vector been utilized in phase I clinical studies.20,51 Whether these expression is down-regulated to a level where the anti- vectors will maintain sufficient expression in vivo in dif- viral efficacy was measurably reduced. Although the acti- ferentiated target cells when introduced into human vation status of HIV-infected T cells in vivo is being inves- hematopoietic stem cells in gene therapy protocols needs tigated in many laboratories, our data indicate that to be studied. The results from the phase I human hema- improved transgene expression in nonactivated T cells topoietic stem cell clinical trials will help resolve these will be an important factor to develop effective retroviral critical issues. vector-based anti-HIV gene therapies. Previously, MoMLV-LTR has been suspected as being Materials and methods a poor promoter for expression in hematopoietic cells20,21 and vectors derived from MPSV and MESV viruses were Vector constructions and retrovirus-producing cells proposed as alternatives.24,34,49,50 However, these obser- The structures of recombinant retrovirus vectors are vations are from studies utilizing either stem cell lines34,50 shown in Figure 1. LNCML was constructed by cloning RevM10 expression in cell lines and primary T cells I Plavec et al 135

Figure 6 Expression of retroviral vectors is down-regulated in nonactivated primary CD4ϩ T cells from PBL. (a) Schematic depiction of a bicistronic retroviral vector encoding RevM10 and the Lyt-2 surface marker. (b) Lyt-2-enriched LMILy and MESV-MILy-transduced CD4ϩ T cells were stimulated with PHA and feeder cells and analyzed for CD25 and Lyt-2 expression on days 3 and 11 after stimulation. A representative experiment is shown. Results were reproduced using four separate tissues.

the RevM10 cDNA13 into the polylinker of the LNCX vec- fragment spanning 5′LTR and packaging sequences, the tor.38 LNMPL consist of the murine PGK promoter,28 NcoI–XhoI RevM10 fragment (NcoI site was engineered at RevM10 gene and the murine PGK poly A signal28 the Rev ATG) and the XhoI–SalI fragment from LMSPL inserted in place of the CMV promoter fragment (NruIto spanning the SV40-Puro gene and the 3′LTR. Retroviral HindIII) of LNCX. In LPMPL, the Neo gene fragment was vector plasmid DNAs were transfected into BOSC-23 replaced by the puromycin resistance gene fragment.52 cells as described.54 Forty-eight hours after transfection, The LMTNL vector was described previously.13,43 The BOSC-23 supernatants containing ecotropic retrovirus LMTNL vector carries a variant of the MoMLV LTR were used to inoculate PA317 cells.55 Following transduc- which has one direct repeat in the enhancer region. tion, PA317 cells were selected either in 600 ␮g/ml G418 LMTNL LTR is functionally indistinguishable from the (Gibco-BRL, Gaithersburg, MD, USA) or 10 ␮g/ml puro- MoMLV LTRs (unpublished results). MPSV-MTNL is mycin hydrochloride (Sigma, St Louis, MO, USA) to gen- identical to the LMTNL, only the LTRs (NheI–KpnI erate pools of producer cells. Retroviral titers were fragment) are derived from MPSV (a gift from C Stock- determined on NIH3T3 cells as described.56 All producer ing, Heinrich-Pette-Institut, Hamburg, Germany). The cells were tested for the presence of replication- L/587-MTNL and MESV-MTNL were derived by replac- competent retrovirus by S+L− assay on PG4 cells.57 ing the MoMLV primer binding sites (NheI–SpeI fragment) in LMTNL and MPSV-MTNL, respectively, Analysis of Rev protein and RNA expression in with the primer binding site from the dl587rev virus.35 transduced CEMSS cells LMSPL was derived from LMTNL by replacing the tk- CEMSS cells were transduced by co-cultivation with the Neo gene with the SV40Puro gene from pBABEpuro.52 PA317-based producer cells in 25-cm3 (T25) ﬂasks over- LMSPL-INT1 was constructed by replacing the SV40Neo night in the presence of 2 ␮g/ml polybrene and sub- gene in LXSN for the RevM10–SV40Puro fragment from sequently passaged twice daily for 3 days into new ﬂasks LMSPL. The LMSPL-INT2 vector was constructed using to remove any adherent producer cells. Transduced cells three DNA fragments: the MFG-lacZ-derived53 SalI–NcoI were selected in medium containing either 2.5 ␮g/ml pur- RevM10 expression in cell lines and primary T cells I Plavec et al 136

Figure 7 Lyt-2 surface protein follows CD25 expression kinetics, a marker of T cell activation. PBL-derived CD4ϩ T cells were transduced with various vectors (indicated in the legend) and enriched for Lyt-2 surface protein. Fully resting cells (as determined by staining for CD25 marker) were stimulated with PHA and feeder cells and Lyt-2 and CD25 expression were measured every 2 to 3 days over a period of 11 days. A representative experiment is shown. Results were reproduced using four separate tissues.

omycin hydrochloride (Sigma, St Louis, MO, USA) or 800 temperature for 2 h. The 50% TCID50/ml of cell-free HIV- ␮g/ml G418 (Gibco-BRL). Stable G418 and puromycin- 1 stock was determined by end-point titration as pre- resistant CEMSS pools were generally established 2 to 3 viously described.59 Following virus inoculation, cells weeks after transduction. RevM10 protein expression was were diluted in fresh RPMI-1640 medium, 10% FBS and measured by Rev-specific immunofluorescence as incubated at 37°C. At different time-points thereafter, cell described.13,29 Total cellular RNA was isolated from counts were determined and culture supernatants were transduced CEMSS cells using the RNazol B protocol analyzed for HIV-1 p24 antigen concentration. (Ambion, Austin, TX, USA) and analyzed by Northern blot using radiolabeled Rev-specific oligonucleotide (5′ Transduction of primary T cells CGG GCC TGT CGG GTC CCC TCG GAA TTG GGA Buffy coats containing PBL from normal healthy donors GGT GGG 3′) as described.58 were enriched for CD4+ cells by depleting CD8+ cells using anti-CD8 biotinylated Ab (Becton Dickinson, San HIV-1 assays Jose´, CA, USA) and streptavidin magnetic dynabeads A3.01/ProHygro cells13 were co-cultivated with the (Dynal, Lake Success, NY, USA). This procedure yielded PA317-based producer cells in transwell plates for 1 to 3 90–95% pure CD4+ populations. Cells were stimulated in days in the presence of 2 ␮g/ml polybrene. Following a TOC medium (RPMI supplemented with 1 × MEM vit- transduction, cells were selected in 400 ␮g/ml hygromy- amin solution (Gibco-BRL), insulin-transferrin-sodium cin B (Boehringer, Indianapolis, IN, USA) and 800 ␮g/ml selenite supplement (Sigma) and 10% fetal bovine serum G418 (Gibco-BRL). HIV production in G418 and hygro- (Hyclone, Logan, UT, USA) with PHA (2 ␮g/ml), IL-2 mycin-resistant pools was monitored over a period of 5– (40 U/ml) and irradiated allogeneic peripheral blood 6 days. Cells were washed once with PBS and then mononuclear cells and JY cells as feeders60 for 3–4 days. seeded in duplicate at a density of 2 × 105 cells/ml (5 ml Retroviral vector transduction was performed by centri- total volume) in a six-well plate. Each day, after thorough fugation of 5 × 105 cells with 1 ml of supernatant from pipetting to ensure homogeneous cultures, 1 ml of cell retroviral producer cells supplemented with 8 ␮g/ml suspension was removed and replaced with 1 ml of fresh polybrene for 4 h at 1500 g and 34°C. This procedure was medium until the cells reached confluency (day 5 or 6). repeated on 2 consecutive days. The transduced cells Cell densities were determined using a Coulter Counter were generally enriched by two further rounds of posi- (Model Z1; Coulter, Miami, FL, USA). The supernatants tive selection using anti-Lyt-2 biotinylated Ab were cleared by centrifugation and frozen at −70°C. HIV- (PharMingen, San Diego, CA, USA) and streptavidin 1 p24 antigen concentration in the culture supernatants magnetic dynabeads (Dynal). For analysis of retroviral was determined on serial dilutions using a commercially vector expression, approximately 106 cells were stimu- available ELISA kit (Dupont-NEN, Boston, MA, USA). lated as above and at different time-points thereafter, Polyclonal CEMSS cell lines (2 × 105/ml) were inoculated small aliquots of cells were removed and double-stained with varying doses (TCID50/ml) of HIV-1 HXB3 at room with anti Lyt-2-PE (PharMingen) and anti-human CD25- RevM10 expression in cell lines and primary T cells I Plavec et al 137

Figure 8 Antiviral effect of RevM10 is reduced in nonstimulated primary T cells. Cells were harvested on days 5 and 8 after stimulation with PHA and feeder cells and inoculated with the HIV-1 JR-CSF isolate. A portion of the day 5 harvested cells were restimulated with feeder cells 72 h after inoculation (‘day 5 restimulated’). The cultures were followed over time measuring HIV-1 p24 Gag antigen levels in the culture supernatants. Vertical bars indicate standard error from triplicate samples (where not shown, the error was below resolution of the graphic program).

FITC (Becton Dickinson) antibodies. After stimulation, all 2 gene. The following reagents were obtained through cultures were maintained in TOC medium supplemented the AIDS Research and Reference Reagent Program, only with IL-2. Division of AIDS, NIAID, NIH: CEMSS cells, JR-CSF HIV-1 virus. The Progenesys Program at SyStemix is a HIV infection of primary T cells Research and Development Collaboration jointly spon- Transduced T cells were stimulated with PHA/IL- sored by Sandoz and SyStemix. 2/allogeneic feeders as above and on days 5 and 8 thereafter aliquots of cells were harvested, washed and resus- References pended in TOC medium containing IL-2 alone. Twenty thousand cells in 75 ␮l were mixed with 75 ␮l of an undi- 1 Cullen BR. Regulation of HIV-1 gene expression. FASEB J 1991; 4 5 5: 2361–2368. luted JR-CSF HIV-1 virus stock (10 –10 TCID50/ml) and plated in triplicate in round-bottom 96-well plates. Cells 2 Felber BK, Pavlakis GN. Molecular biology of HIV-1: positive were cultured overnight and on the following day, 125 and negative regulatory elements important for virus expression. AIDS 1993; 7: 51–62. ␮l of medium was removed and replaced with 135 ␮l + 3 Rosen CA. Regulation of HIV gene expression by RNA-protein of fresh TOC medium IL-2. Culture supernatants were interactions. Trends Genet 1991; 7: 9–14. harvested on days 3, 5, 7 and 9 after inoculation. Where 4 Malim MH et al. The HIV-1 rev trans-activator acts through a indicated, 72 h after inoculation with HIV, 135 ␮l TOC structured target sequence to activate nuclear export of medium + IL-2 containing 2.5 × 105 feeder cells/ml and unspliced viral mRNA. Nature 1989; 338: 254–257. PHA (2 ␮g/ml) were added to the cultures. This step was 5 Fischer U et al. The HIV-1 Rev activation domain is a nuclear included to maintain the activation status of T cells. HIV- export signal that accesses an export pathway used by specific 1 p24 antigen concentration in the culture supernatants cellular RNAs. Cell 1995; 82: 475–483. was determined using an ELISA kit (Dupont-NEN). 6 Bogerd HP et al. Identification of a novel cellular cofactor for the Rev/Rex class of retroviral regulatory proteins. Cell 1995; 82: 485–494. Acknowledgements 7 Fritz CC, Zapp ML, Green MR. A human nucleoporin-like protein that specifically interacts with HIV Rev. Nature 1995; 376: The authors thank Creton Kalfoglou, Vincent Suzara, Jo 530–533. Dando, Jingyi Chen, Sean Forestell and Heini Ilves for 8 Ruhl M et al. Eukaryotic initiation factor 5A is a cellular target technical assistance, and Garry P Nolan for the IRES-Lyt- of the human immunodeficiency virus type 1 Rev activation RevM10 expression in cell lines and primary T cells I Plavec et al 138 domain mediating trans-activation. J Cell Biol 1993; 123: 1309– leukemia virus long terminal repeats. Mol Cell Biol 1989; 9: 1320. 739–746. 9 Malim MH et al. Functional dissection of the HIV-1 Rev trans- 33 Berwin B, Barklis E. Retrovirus-mediated insertion of expressed activator – derivation of a trans-dominant repressor of Rev func- and non-expressed genes at identical chromosomal locations. tion. Cell 1989; 58: 205–214. Nucleic Acids Res 1993; 21: 2399–2407. 10 Vandendriesche T et al. Inhibition of clinical human immuno- 34 Beck-Engeser G et al. Retroviral vectors related to the myelopro- deficiency virus (HIV) type 1 isolates in primary CD4+ T lym- liferative sarcoma virus allow efficient expression in hemato- phocytes by retroviral vectors expressing anti-HIV genes. J Virol poietic stem and progenitor cell lines, but retroviral infection is 1995; 69: 4045–4052. reduced in more primitive cells. Hum Gene Ther 1991; 2: 61–70. 11 Woffendin C et al. Nonviral and viral delivery of a human 35 Colicelli J, Goff SP. Isolation of a recombinant murine leukemia immunodeficiency virus protective gene into primary human T virus utilizing a new primer tRNA. J Virol 1986; 57: 37–45. cells. Proc Natl Acad Sci USA 1994; 91: 11581–11585. 36 Cepko CL, Roberts BE, Mulligan RC. Construction and appli- 12 Malim MH et al. Stable expression of transdominant Rev protein cations of a highly transmissible murine retrovirus shuttle vec- in human T cells inhibits human immunodeficiency virus repli- tor. Cell 1984; 37: 1053–1062. cation. J Exp Med 1992; 176: 1197–1201. 37 Palmiter RD et al. Heterologous introns can enhance expression 13 Escaich S et al. RevM10-mediated inhibition of HIV-1 replication of transgenes in mice. Proc Natl Acad Sci USA 1991; 88: 478–482. in chronically infected T cells. Hum Gene Ther 1995; 6: 625–634. 38 Miller AD, Rosman GJ. Improved retroviral vectors for gene 14 Bevec D et al. Inhibition of human immunodeficiency virus type transfer and expression. BioTechniques 1989; 7: 980–990. 1 replication in human T cells by retroviral mediated gene trans- 39 Dranoff G et al. Vaccination with irradiated tumor cells fer of a dominant-negative Rev trans-activator. Proc Natl Acad engineered to secrete murine granulocyte–macrophage colony- Sci USA 1992; 89: 9870–9874. stimulating factor stimulates potent, specific, and long-lasting 15 Bahner I et al. Comparison of trans-dominant inhibitory mutant anti-tumor immunity. Proc Natl Acad Sci USA 1993; 90: 3539– human immunodeficiency virus genes expressed by retroviral 3543. vectors in human T lymphocytes. J Virol 1993; 67: 3199–3207. 40 Bender MA et al. Evidence that the packaging signal of Moloney 16 Gilboa E, Smith C. Gene therapy for infectious diseases: the murine leukemia virus extends into the gag region. J Virol 1987; AIDS model. Trends Genet 1994; 10: 139–144. 61: 1639–1646. 17 Baltimore D. Intracellular immunization. Nature 1988; 335: 41 Tagawa M et al. Formal proof that different-size Lyt-2 polypep- 395–396. tides arise from differential splicing and post-transcriptional 18 Nabel GJ et al. A molecular genetic intervention for AIDS – regulation. Proc Natl Acad Sci USA 1986; 83: 3422–3426. effects of a transdominant negative form of Rev. Hum Gene Ther 42 Jang SK et al. Initiation of protein synthesis by internal entry of ′ 1995; 5: 79–92. ribosomes into 5 nontranslated region of encephalomyocarditis 19 Woffendin C et al. Expression of a protective gene prolongs sur- virus RNA in vivo. J Virol 1989; 63: 1651–1660. vival of T cells in human immunodeficiency virus-infected 43 Plavec I et al. Sustained retroviral gene marking and expression patients. Proc Natl Acad Sci USA 1996; 93: 2889–2894. in lymphoid and myeloid cells derived from transduced hema- 20 Miller AD. Human gene therapy comes of age. Nature 1992; 357: topoietic progenitor cells. Gene Therapy 1996; 3: 717–724. 44 Kotani H et al. Improved method of retroviral vector transduc- 455–460. tion and production for gene therapy. Hum Gene Ther 1994; 5: 21 Naviaux RK, Verma IM. Retroviral vectors for persistent 19–28. expression in vivo. Curr Opin Biotech 1992; 3: 540–547. 45 Ramilo O et al. Role of CD25+ and CD25− T cells in acute HIV 22 Williams DA. Expression of introduced genetic sequences in infection in vitro. J Immunol 1993; 150: 5202–5208. hematopoietic cells following retroviral mediated gene transfer. 46 Zack JA et al. HIV-1 entry into quiescent primary lymphocytes: Hum Gene Ther 1990; 1: 229–239. molecular analysis reveals a labile, latent viral structure. Cell 23 Jolly D. Viral vector system for gene therapy. Cancer Gene Ther 1990; 61: 213–222. 1994; 1: 51–64. 47 Salmon P et al. Characterization of the human CD4 gene pro- 24 Rivie`re I, Brose K, Mulligan RC. Effect of retroviral vector moter: transcription from the CD4 gene core promoter is tissue- design on expression of human adenosine deaminase in murine specific and is activated by Ets proteins. Proc Natl Acad Sci USA bone marrow transplant recipients engrafted with genetically 1993; 90: 7739–7743. modified cells. Proc Natl Acad Sci USA 1995; 92: 6733–6737. 48 Mavilio F et al. Peripheral blood lymphocytes as target cells of 25 Lynch CM et al. Sequences in the coding region of clotting factor retroviral vector-mediated gene transfer. Blood 1994; 83: 1988– VIII act as dominant inhibitors of RNA accumulation and pro- 1997. tein production. Hum Gene Ther 1993; 4: 259–272. 49 Hawley RG et al. Transplantable myeloproliferative disease 26 Apperley JF, Luskey BD, Williams DA. Retroviral gene transfer induced in mice by an interleukin 6 retrovirus. J Exp Med 1992; of human adenosine deaminase in murine hematopoietic cells: 176: 1149–1164. effect of selectable marker sequences on long-term expression. 50 Baum C et al. Novel retroviral vectors for efficient expression of Blood 1991; 78: 310–317. the multidrug resistance (mdr-1) gene in early hematopoietic 27 Boshart M et al. A very strong enhancer is located upstream of cells. J Virol 1995; 69: 7541–7547. an immediate–early gene of human cytomegalovirus. Cell 1985; 51 Brenner MK. Human somatic gene therapy: progress and prob- 41: 521–530. lems. J Int Med 1995; 237: 229–239. 28 Adra CN, Boer PH, McBurney MW. Cloning and expression of 52 Morgenstern JP, Land H. Advanced mammalian gene transfer: the mouse pgk-1 gene and the nucleotide sequence of its pro- high titre retroviral vectors with multiple drug selection mark- moter. Gene 1987; 60: 65–74. ers and a complementary helper-free packaging cell line. Nucleic 29 Rigg RJ et al. Detection of intracellular HIV-1 Rev protein by Acids Res 1990; 18: 3587–3596. flow cytometry. J Immunol Meth 1995; 188: 187–195. 53 Ferry N et al. Retroviral mediated gene transfer into hepatocytes 30 Artelt P et al. The prokaryotic neomycin-resistance-encoding in vitro. Proc Natl Acad Sci USA 1991; 88: 8377–8381. gene acts as a transcriptional silencer in eukaryotic cells. Gene 54 Pear WS et al. Production of high-titer helper-free retroviruses 1991; 99: 249–254. by transient transfection. Proc Natl Acad Sci USA 1993; 90: 31 Challita PM et al. Multiple modifications in cis elements of the 8392–8396. long terminal repeat of retroviral vectors lead to increased 55 Miller AD, Buttimore C. Redesign of retrovirus packaging cell expression and decreased DNA methylation in embryonic carci- lines to avoid recombination leading to helper virus production. noma cells. J Virol 1995; 69: 748–755. Mol Cell Biol 1986; 6: 2895–2902. 32 Flanagan JR et al. Negative control region at the 5′ end of murine 56 Forestell SP, Bo¨ hnlein E, Rigg RJ. Retroviral end-point titer is RevM10 expression in cell lines and primary T cells I Plavec et al 139 not predictive of gene transfer efficiency: implications for vector Lennette EH, Halonen P, Murphy FA (eds). Laboratory Diagnosis production. Gene Therapy 1995; 2: 723–730. of Infectious Diseases: Principles and Practice. Springer-Verlag: 57 Haapala DK et al. Isolation from cats of an endogenous type C New York, 1988, pp 39–59. virus with a novel glycoprotein. J Virol 1985; 53: 827–833. 60 Vandekerckhove BA et al. Human hematopoietic cells and 58 Valent P et al. Interleukin 4 promotes expression of mast cell thymic epithelial cells induce tolerance via different mech- ICAM-1 antigen. Proc Natl Acad Sci USA 1991; 88: 3339–3342. anisms in the SCID-hu mouse thymus. J Exp Med 1992; 175: 59 Leland DS, French MLV. Virus isolation and identification. In: 1033–1043.