Engagement of the B-cell antigen receptor (BCR) allows ... - Nature

Gene Therapy (2004) 11, 1416–1424 & 2004 Nature Publishing Group All rights reserved 0969-7128/04 $30.00 www.nature.com/gt

RESEARCH ARTICLE

Engagement of the B-cell antigen receptor (BCR) allows efficient transduction of ZAP-70-positive primary B-CLL cells by recombinant adeno-associated virus (rAAV) vectors DM Kofler1,3,4,5, H Bu¨ning4, C Mayr1, D Bund1, J Baumert2, M Hallek1,3,4,5 and C-M Wendtner1,3,4,5 1

KKG Gene Therapy, GSF-National Research Center for Environment and Health, Munich, Germany; 2Institute of Epidemiology, GSF-National Research Center for Environment and Health, Munich, Germany; 3Medical Clinic III, Klinikum Grosshadern Medical Center (KGMC), Munich, Germany; 4Gene Center, Ludwig-Maximilians-University, Munich, Germany; and 5Medical Clinic I, University of Cologne, Cologne, Germany

Engagement of the B-cell antigen receptor (BCR) by crosslinking of the surface immunoglobulin (sIg) homodimer was studied for recombinant adeno-associated virus (rAAV)mediated gene transfer into B-cell chronic lymphocytic leukaemia (B-CLL) cells. Leukemic cells obtained from 20 patients were stimulated with anti-sIg-directed antibodies and transduced with rAAV vectors coding for enhanced green fluorescent protein (EGFP) (AAV/EGFP) or CD40L (AAV/CD40L). Transduction of B-CLL cells was enhanced after BCR engagement compared to unstimulated controls (P ¼ 0.0356). BCR crosslinking induced a significant, doseand time-dependent upregulation of heparan sulfate proteoglycan (HSPG), the primary receptor for AAV, on B-CLL cells (mean: 38.2 versus 1.7%; P ¼ 0.0006). A correlation of

HSPG expression after BCR crosslinking with transduction efficiency by AAV/EGFP (P ¼ 0.0153) and AAV/CD40L (P ¼ 0.0347) was observed. High expression of zeta-associated protein 70 (ZAP-70) in B-CLL cells correlated with a better transduction efficiency by AAV/EGFP (Po0.0001) and AAV/CD40L (P ¼ 0.002), respectively: 48 h after transduction of ZAP-70-positive samples, transgene expression was seen in a mean of 33.8% (s.e.m. 3.7%) and 28.9% (s.e.m. 6.7%) of cells, respectively, and could be specifically blocked by heparin, a soluble competitor of HSPG (Po0.0001). In summary, engagement of the BCR on ZAP-70 positive B-CLL cells allows efficient rAAV-mediated gene delivery. Gene Therapy (2004) 11, 1416–1424. doi:10.1038/ sj.gt.3302279; Published online 22 July 2004

Keywords: AAV; CLL; BCR; HSPG; ZAP-70

Introduction B-cell chronic lymphocytic leukemia (B-CLL) is characterized by the abnormal expansion of CD5+ B cells arrested in G0/G1 phase of the cell cycle.1 CLL cells fail to undergo apoptosis in vivo, but survive only few days in vitro. Besides bone marrow stromal cells, ‘nurse-like’ cells, and different chemokines and cytokines as interleukin 4 and tumor necrosis factor alpha (TNFa), engagement of the CD40 receptor by its ligand, CD40L, is thought to induce survival via the activation of NFkB.2–5 The survival of leukemic B cells is also enhanced by engagement of their surface immunoglobulins (sIg), usually of IgM subtype (sIgM). B-cell receptor (BCR) crosslinking was shown to promote cell survival by caspase inhibition, induction of NF-kB, and expression of antiapoptotic molecules like bcl-2.6 The purpose of this study was to determine whether BCR crosslinking could be used to overcome the resistance of primary B-CLL cells toward transduction Correspondence: Dr C-M Wendtner, Klinik I fu¨r Innere Medizin, Universita¨t zu Ko¨ln, Joseph-Stelzmann-Strasse 9, Ko¨ln D-50924, Germany Received 22 December 2003; accepted 20 March 2004; published online 22 July 2004

by recombinant adeno-associated virus (rAAV) vectors in order to enable the expression of immunostimulatory transgenes for cellular vaccination strategies. Recently, we have shown that B-CLL cells stimulated by cocultivation on CD40L-expressing feeder cells can be efficiently infected by rAAV vectors.7 To avoid this time-consuming and laborious prestimulation step, we explored whether it could be replaced by the stimulation of B-CLL cells with antibodies directed against the sIg homodimer of the BCR. Here, using freshly isolated cells from patients with CLL, we demonstrate that BCR crosslinking allows upregulation of the primary AAV receptor, heparan sulfate proteoglycan (HSPG). This resulted in an efficient AAV transduction with functional expression of transgenes in primary B-CLL cells without the need for a stimulating CD40L-expressing feeder system, thus facilitating future vaccination efforts for B-CLL.

Results Characterization of B-CLL samples by surface IgM (sIgM), CD38, and zeta-associated protein 70 (ZAP-70) expression Freshly isolated CD5/CD19+ cells from patients with established diagnosis of B-CLL were studied for the

AAV gene transfer into B-CLL by BCR-stimulation DM Kofler et al

expression of sIg, CD38, and intracellular ZAP-70. The study population consisted of 20 patients (15 male, five female). The mean age of the study population was 60.8 years (s.e.m., 2.4 years) with a median age of 62 years. Eight patients presented with early disease (Binet stage A) and 12 with advanced disease (Binet stage B, n ¼ 7; Binet stage C, n ¼ 5). Clinical characteristics of the patients including cytogenetic analysis by FISH are summarized in Table 1. Expression of sIgM, the most common sIg expressed on B-CLL cells,8 and expression of CD38 and ZAP-70, two surrogate markers for the mutational status of the immunoglobulin gene (IgVH),9,10 were analyzed in all patients being entered on this study. Normal B cells isolated from human tonsils (n ¼ 3) were also studied for sIgM, CD38, and ZAP-70. While expression of sIgM was seen in both B-CLL cells (range, 3–97.0%) and tonsillar B cells (range, 55–69%), CD38 (range, 0–38.0%) and ZAP-70 (range, 0–100%) molecules could only be detected in CLL samples.

HSPG upregulation on primary B-CLL cells after CD40L stimulation compared to BCR engagement In the previous work, we have shown that AAV transduction of primary B-CLL cells is possible after stimulation of leukemic B cells by CD40L.7 Besides induction of cell cycle progression, we hypothesized that activation of B-CLL cells by CD40L might induce an upregulation of the primary AAV receptor, HSPG. Therefore, HSPG expression was studied on B-CLL cells 48 h after cocultivation on CD40L-expressing HeLa/SF cells. As shown in Figure 1a, expression of HSPG is significantly increased when CLL cells are stimulated by HeLa/SF cells compared to unstimulated control CLL cells (mean: 54.3%, s.e.m. 15.4% versus mean: 0.5%, s.e.m. 0.2%; P ¼ 0.019). A significant result was also found using mean fluorescence intensity ratio (MFIR) values for this comparison (mean: 22.3, s.e.m. 10.8 versus mean: 1.2, s.e.m. 0.1; P ¼ 0.03).

HSPG expression was examined in CLL samples after stimulation with anti-IgM antibodies (a-IgM) at different concentrations for 72 h. As shown in Figure 1b, after stimulation with a-IgM at a concentration of 10 mg/ml, a mean of 51.9% of cells expressed HSPG on their surface, while no significant expression was detectable without the use of a-IgM. A further increase of HSPG induction was seen using 100 mg/ml of a-IgM with a mean percentage of 82.4% HSPG-positive CLL cells. There was a saturation kinetics observed with a mean of 98.4% HSPG-positive cells at a high concentration of 200 mg/ml a-IgM. Therefore, for all further experiments a concentration of 100 mg/ml a-IgM was used. Samples derived from B-CLL patients (Table 2) were examined for their HSPG expression before and after stimulation with a-IgM. The expression of HSPG was significantly increased on CLL cells after BCR engagement (mean: 38.2%, s.e.m. 7.7% versus mean: 1.7%, s.e.m. 0.7%; P ¼ 0.0006). Evaluation of HSPG expression by MFIR also revealed a significant difference between both groups in favor of a-IgM-stimulated samples (mean: 4.5, s.e.m. 1.4 versus mean: 1.2, s.e.m. 0.08; P ¼ 0.035). HSPG induction by a-IgM was correlated with sIgM expression levels (P ¼ 0.004). Time kinetics of HSPG expression was assessed after cocultivation of B-CLL cells with CD40L-expressing HeLa/SF cells (nos. 8, 14–16, see Table 1). After 24 h, a mean of 62.5% (s.e.m. 15.6%) of cells expressed HSPG, a maximum was seen after 48 h with a mean of 74.6% (s.e.m. 14.3%) of CLL cells positive for HSPG. The expression of HSPG increased nonsignificantly after 96 and 120 h with a mean percentage of 84.7% (s.e.m. 10.1%) and 82.0% (s.e.m. 9.3%) positive cells, respectively (P ¼ 0.10). Samples derived from the same patients were studied for HSPG upregulation over time using a-IgM. After 24 h, a mean HSPG expression of 24.9% (s.e.m. 2.7%) was observed, while a significant enhancement of receptor expression on B-CLL cells was detectable 48 h after

1417

Table 1 Patients’ characteristics Patient 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Sex/age (years)

Stage (Binet)

Karyotype (FISH)

sIgM (in %)

CD38 (in %)

ZAP-70 (in %)

M/82 F/54 M/47 F/64 M/62 M/45 F/37 M/54 F/60 M/53 M/72 M/59 M/78 M/63 M/67 M/67 M/60 M/62 F/62 M/67

B A B A B C A A C A B C C C B A A B A B

13q Normal Normal Normal Normal ND 12+ 11q ND 13q 11q, 13q 13q 11q Normal ND ND 13q Normal Normal 11q, 13q

85.5 33.0 59.0 59.0 23.0 43.0 3.0 39.0 97.0 36.0 78.0 52.5 89.0 91.0 87.4 33.8 92.0 92.4 97.0 89.0

7.0 31.8 7.0 28.4 0.1 0 17.7 10.7 14.6 0.6 2.7 0.1 32.9 8.0 32.6 0.8 38.0 4.5 0 0.1

97.7 94.8 100 85.7 97.3 75.6 0 0.4 62.7 98.1 77.6 64.0 97.6 100 98.0 99.6 12.0 0.1 0 14.8

M ¼ male, F ¼ female, FISH ¼ fluorescence in situ hybridization, sIgM ¼ surface IgM, ZAP-70 ¼ zeta-associated protein 70, ND ¼ not determined. Gene Therapy


1418

Table 2 Expression of HSPG and transgenes (EGFP, CD40L) 48 h after BCR crosslinking Patient 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

HSPG (in %)

EGFP (in %)

CD40L (in %)

54.1 31.4 40.8 53.0 53.0 31.0 0 13.0 72.0 72.0 80.5 40.7 21.6 58.0 45.8 32.5 10.0 2.0 0 0

11.0 9.0 31.0 54.0 51.0 46.0 0 0.5 28.0 30.0 36.0 ND 40.8 37.5 29.8 35.1 0 2.0 0 0

ND ND 58.5 ND ND ND 0.9 1.7 54.0 23.0 11.0 13.0 28.0 9.0 34.4 ND 1.0 2.8 0 0.3

ND ¼ not determined.

Figure 1 Expression of membrane-bound HSPG, the primary receptor for AAV-2, on B-CLL cells. Given is the expression of HSPG (%) on B-CLL cells (nos. 4–6,14–16; see Table 1) 48 h after cocultivation with CD40Lexpressing HeLa/SF cells in comparison to unstimulated leukemic B cells (a). Concentration kinetics of HSPG induction on B-CLL cells (nos. 4–6) 72 h after BCR stimulation with different concentrations of a-IgM in mg/ml (b). Shown is the mean percentage of HSPG-positive cells with s.e.m. (error bars).

coincubation (mean: 37.5%, s.e.m. 4.2%; P ¼ 0.20). At 96 and 120 h, HSPG levels were decreasing (mean: 36.6%, s.e.m. 4.0% and mean: 25.9%, s.e.m. 8.0%, respectively; P ¼ 0.398 and 0.148, respectively). Concurrent stimulation of B-CLL cells by HeLa/SF and a-IgM (100 mg/ml) did not increase HSPG expression. Human tonsillar lymphocytes (n ¼ 3) showed an HSPG expression in 7.9 to 22.1% of cells. HSPG could be upregulated by either HeLa/SF (mean: 23.4%) or aIgM (mean: 50.9%), but a high rate of apoptotic cells was observed after BCR ligation.

AAV transduction of primary B-CLL cells after BCR engagement Freshly isolated B-CLL cells were incubated with 100 mg/ ml a-IgM and rAAV coding for enhanced green fluorescent protein (EGFP) (AAV/EGFP) and/or CD40L (AAV/CD40L) was added (MOI 50). After 48 h, transgene expression was assessed by flow cytometric analysis. The median expression of EGFP (19 patients) and CD40L (14 patients) was 29.8% (range 0–54%) and 10% (range 0–58.5%), respectively, with a mean expression of 23.2% (s.e.m. 4.4%) and 17.0% (s.e.m. 5.3%), Gene Therapy

respectively (Table 2). Transduction of B-CLL cells was significantly enhanced after treatment of leukemic cells by anti-sIg antibodies compared to unstimulated leukemic cells (mean: 21.8%, s.e.m. 7.0% versus mean: 1.2%, s.e.m. 0.6%) (P ¼ 0.0356). Transduction levels of EGFP and CD40L were not significantly correlated with expression levels of sIgM (P ¼ 0.561 and 0.713, respectively). Importantly, there was a significant correlation of HSPG expression after a-IgM stimulation and transduction efficiency using AAV/EGFP (P ¼ 0.0153) and AAV/ CD40L (P ¼ 0.0347), respectively. There was no correlation seen between transduction efficiency with AAV/ EGFP and AAV/CD40L, respectively, and clinical stage of disease (A versus B/C; P ¼ 0.174 and 0.105, respectively), sex (P ¼ 0.511 and 0.903, respectively) or age (continuous) (P ¼ 0.545 and 0.928, respectively). Specificity of AAV transduction after BCR stimulation was controlled by the incubation of AAV-infected CLL samples by heparin, a soluble receptor analogue similar to HSPG.11 As seen in Figure 2, viral transduction of EGFP after BCR stimulation was significantly decreased by heparin (P ¼ 0.0001), resulting in background levels of EGFP expression (mean: 1.8%, s.e.m. 1.2%). As further control, CLL samples were unspecifically stimulated by antibodies directed against sIgA, which is rarely expressed on CLL cells: only 2.4% (s.e.m. 1.3%) of cells could be transduced by AAV/EGFP. Transgene expression after unspecific stimulation of the BCR was significantly inferior compared to AAV transduction after stimulation of the same samples with a-IgM (P ¼ 0.012). Transgene expression after AAV/EGFP transduction in the context of BCR engagement was compared to rAAV transduction after stimulation of B-CLL cells by CD40L-expressing HeLa/SF cells. Based on different CLL samples studied (nos. 14–16), there was no significant difference detectable in the percentage of EGFP-expressing cells (feeder: mean of 32.7%, s.e.m. 11.1% versus a-IgM: mean of 34.1%, s.e.m. 1.6%;


Figure 2 Transduction of primary B-CLL cells by AAV/EGFP after BCR ligation. CLL cells derived from 19 different patients (nos. 1–11, 13–20; see Table 2) were stimulated with 100 mg/ml a-IgM and infected with AAV/ EGFP (MOI 50). After 48 h, EGFP expression was detected by flow cytometry and mean percentage of positive cells with s.e.m. (error bars) is shown. Samples from CLL patients (nos. 4, 5, 9, 14–16) being infected with AAV/EGFP and stimulated with a-IgM were also incubated with heparin (1250 U/ml) and analyzed by flow cytometry 48 h later. Finally, CLL samples (nos. 1–4, 6, 7, 9) were stimulated with a-IgA (100 mg/ml) before infection with AAV/EGFP. Assessment of EGFP expression was performed 48 h after AAV transduction.

P ¼ 0.913). Furthermore, concurrent stimulation of CLL cells by HeLa/SF cells and a-IgM stimulation did not increase the transduction efficiency (data not shown). In contrast, there was a marked difference of transduction efficiency in tonsillar B-lymphocytes derived from healthy donors after IgM crosslinking and after stimulation by CD40L-expressing feeder cells: the detection of transduced normal B cells was hampered by a high rate of apoptotic cells after IgM stimulation, while feeder stimulation of B-lymphocytes resulted in a high transduction efficiency (mean 42.7%, s.e.m. 4.6%). Next, the same CLL samples were infected by a different rAAV serotype, that is, AAV serotype 1 (AAV-1), in contrast to the common AAV serotype 2 (AAV-2) used in other experiments. AAV-1 is known to enter target cells via so far unknown receptors excluding membranebound HSPGs. Infection by recombinant AAV-1 coding for EGFP (AAV-1/EGFP) together with a-IgM stimulation did not result in a transgene expression beyond the background level (data not shown). Finally, we used an AAV mutant (RGD4C) that contains an Arg-Gly-Asp (RGD) peptide at a specific site of the AAV capsid, thus lacking the specific binding site for the HSPG receptor on target cells.12 This mutant was able to infect HeLa cells (MOI 100) via avb5 integrins as shown previously, but did not infect B-CLL cells (patient nos. 14–16) with (mean 0.4%, s.e.m. 0.3%) or without (mean 3.8%, s.e.m. 3.0%) stimulation of the BCR (P ¼ 0.382).

ZAP-70 expression and transduction efficiency by rAAV after BCR engagement High expression of ZAP-70 in B-CLL cells was shown to be associated with enhanced signal transduction via the BCR complex and a more aggressive clinical course.13,14 We asked whether the expression level of ZAP-70, an

easily detectable indicator of an efficient signaling through the BCR, is predictive for the AAV transduction efficiency into an individual CLL sample after stimulation of the BCR. Based on 20 samples derived from patients with B-CLL (Tables 1 and 2), a high expression of ZAP-70 as assessed by flow cytometric analysis correlated with a significantly enhanced transduction efficiency with AAV/EGFP (Po0.0001) and CD40L, respectively (P ¼ 0.002). Applying a cutoff level of 20% positive cells for definition of ZAP-70 positivity, the association to EGFP (Po0.0001) and CD40L (P ¼ 0.004) transduction levels with ZAP-70 expression was significant. In CLL samples with a ZAP-70 expression below 20%, only 0.4% (s.e.m. 0.3%) and 1.1% (s.e.m. 0.4%) EGFP- and CD40L-positive cells, respectively, were detectable after AAV transduction in the context of BCR stimulation. The mean fraction of EGFP- and CD40Ltransduced cells increased to 33.8% (s.e.m. 3.7%) and 28.9% (s.e.m. 6.7%), respectively, in the cohort of ZAP-70 positive (420%) samples (Figure 3). In all these samples, ZAP-70 expression was neither correlated with CD38 (P ¼ 0.806) nor with sIgM (P ¼ 0.988). HSPG induction after BCR engagement was significantly correlated with the level of ZAP-70 expression (P ¼ 0.0002).

1419

Activation of B-CLL cells after AAV/CD40L gene transfer Since CD40L, a critical molecule for T-cell activation, is downregulated on T cells in patients with CLL and this defect was shown to be corrected by gene transfer of CD40L into leukemic B cells, we wanted to investigate the functionality of B-CLL cells after AAV/CD40L transduction in combination with BCR ligation.7,15–17 To prove that CD40L-transduced CLL cells were activated and become efficient antigen-presenting cells, expression of the costimulatory molecule CD80 was measured before and 120 h after infection with AAV/CD40L. CD80 expression could be induced from a mean of 0.1% (s.e.m. 0.1%) to 17.4% (s.e.m. 4.7%) and was significantly higher in comparison to uninfected control CLL cells (P ¼ 0.0076). Primary CLL cells transduced with wtAAV did not result in any upregulation of CD80 (mean 0.5%, s.e.m. 0.3%) (Figure 4a). It was previously shown that transduction with AAV/ CD40L based on feeder-stimulated CLL cells resulted in the upregulation of CD80 also on noninfected bystander leukemia B cells.7,15 This transactivation capacity of BCR-stimulated and AAV/CD40L-infected CLL cells was also assessed. Transduced CLL cells were labeled with a green fluorescent dye (CellTracker) and used as stimulator cells for equal numbers of nonlabeled CLL B cells derived from the same patient. Uninfected nonlabeled bystander CLL cells were induced to express CD80 after cocultivation with CD40L-transduced CLL cells (mean 5.5%, s.e.m. 1.3%), but not after coincubation with mock-infected (wtAAV) control CLL cells (mean 0.5%, s.e.m. 0.3%; P ¼ 0.0826) (Figure 4b).

Discussion The results presented demonstrate for the first time that gene transfer by rAAV into B-CLL cells can be markedly improved by engagement of the BCR. For ex vivo rAAV-mediated gene transfer in CLL cells, this proceGene Therapy


1420

Figure 4 CD80 upregulation on primary B-CLL cells after AAV/CD40L transduction combined with BCR ligation. CD80 on B-CLL cells (nos. 3, 9, 11–16; see Table 1) was detected before and 120 h after AAV/CD40L transduction (MOI 50) in combination with a-IgM (100 mg/ml) stimulation. CD80 expression was controlled on uninfected or wtAAV-infected CLL cells being stimulated with a-IgM (a). Uninfected nonlabeled bystander CLL cells (nos. 14–16) were assessed for CD80 expression 48 h after coincubation with labeled (CellTracker), CD40L- or wtAAVinfected CLL cells derived from the same patients (b). Shown is the mean percentage of the fraction of CD80-positive CLL cells with corresponding s.e.m. (error bars).

Figure 3 Correlation of the level of expression of ZAP-70 and transduction efficiency by rAAV after BCR engagement. Shown are scatter plots representing the levels of ZAP-70 expression (~: o20% and }: X20%) and the transgene levels after transduction with AAV/EGFP (a) and AAV/CD40L (b), all measured by flow cytometric analysis. In addition, the mean level of EGFP and CD40L expression is shown (horizontal line). When the threshold for categorizing ZAP-70 was established at 20%, two subgroups were clearly delineated: patients with high transgene expression and a high level of ZAP-70 expression and patients with low transgene expression and low levels of ZAP-70 expression: Po0.0001 for AAV/EGFP (a) and P ¼ 0.004 for AAV/ CD40L (b). Flow cytometric analysis of HSPG and transgene (EGFP, CD40L) expression after BCR engagement is shown together with baseline expression of ZAP-70 for four representative patients (nos. 11, 13–15). Within each panel, the curve on the left represents the isotype control and the filled curve (black) on the right shows the expression of ZAP-70, HSPG, EGFP, and CD40L. The percentage of positive cells is defined as the fraction beyond the region of 99% of the control-stained cells (c).

dure allows to replace a time-consuming stimulation process involving the coculture of a feeder cell line transfected with CD40L. Compared to previously published transduction data using the same AAV constructs in the context of feeder stimulation of CLL cells, we Gene Therapy

achieved similar transduction efficiencies for rAAV after BCR ligation and feeder stimulation.7,18 More importantly, we could elucidate a mechanism underlying the phenomenon that AAV transduction in B-CLL cells can be enhanced after BCR crosslinking: the primary AAV receptor, HSPG, was upregulated by BCR engagement. Recently, it was shown that activation of the BCR induces a strong transient expression of HSPGs on human tonsillar B cells, while malignant B cells were not studied.19 HSPGs are critical regulators of growth and differentiation of epithelial and connective tissues for which AAV has a specific tropism.20–24 Moreover, HSPGs function as coreceptors promoting cytokine signaling in normal B cells in the context of antigen-specific B-cell differentiation. Besides activation of the BCR, CD40 was shown to control the expression of HSPGs on tonsillar B cells.19 We provide data that AAV transduction in the context of activation of CLL cells by CD40L-expressing feeder cells is regulated by HSPG expression. Therefore, different signals, that is, stimulation of the BCR and the CD40 receptor on human B-CLL cells, result in the induction of HSPGs on leukemic B cells. On a molecular level, the synergy between CD40 and the B-cell antigen receptor was demonstrated in normal B cells through cooperative signaling by TNF receptor-associated factor 2.25


HSPGs were previously shown to be involved in integrin triggering and adhesion to endothelial cells during the process of extravasation of activated leukemic cells.26 Transmembrane proteoglycans were proven to be essential for some NHLs, including CLL, to ensure a required connection to the microenvironment.27 It can be speculated that integrin-mediated B-cell adhesion and migration is established by hepatocyte growth factor, which was shown to be crucial for HSPG binding in human tonsillar B cells.19 Another interesting candidate would be stromal cell-derived factor (SDF)-1a,28 although at least on human tonsillar B cells no interaction with HSPGs was shown. On the other hand, it was demonstrated that proteoglycans can establish an SDF-1 gradient in the subendothelial matrix guiding migrating hematopoietic progenitor cells into the bone marrow.29 There is some additional evidence in the literature that HSPG induction is possible after the activation of malignant B cells: it was reported that the addition of lipopolysaccharide (LPS) to Daudi cells, a B-cell leukemia cell line, resulted in increased synthesis of heparan sulfate.30 BCR crosslinking and LPS stimulation trigger both similar signaling molecules, that is, activation of NF-kB in B cells.31 An alternative mechanism facilitating AAV transduction in B-CLL cells after BCR engagement that has to be discussed is an induction of S phase of the cell cycle resulting in an enhanced cell survival of B-CLL cells ex vivo. This cell cycle progression was previously described for malignant B cells stimulated by CD40L.7,32,33 While we could observe a shift of the relative number of cells in S phase after BCR ligation in our study (data not shown), there are some conflicting results with regard to the effects of BCR engagement on survival and cell cycle progression of CLL cells. Engagement of surface IgM was described to elicit a strong survival program in BCLL cells, which is associated with the inhibition of caspase activity and activation of NF-kB.6 In contrast, McConkey et al34 described in a small series of patients that in four of seven cases CLL cells underwent apoptosis upon IgM engagement. It is speculated that differences in antibody reagents and concentration of CLL cells in culture contribute to the contradictory results. Furthermore, different effects were observed for sIgD and sIgM crosslinking in dependence of CD38 expression on CLL cells.35 Response to IgM binding in CLL cells can be modulated by other cofactors like CD6 engagement.36 CD5 is another molecule associated with the human BCR complex in CLL B cells and in a small subpopulation of normal B cells.37 CD5 ligation results in heterogeneous apoptotic responses with death signaling operating via CD79 and CD38.38 In a bovine leukemia model resulting in persistent lymphocytosis, the CD5 molecule was dissociated from the BCR in CD5+ B cells.39 This disrupted CD5–BCR interaction resulted in decreased apoptosis and increased survival after a-sIgM stimulation. Finally, one important mechanism for inhibiting signaling for apoptosis in B-CLL after ligation of the BCR might be an overexpression of the alternative transcript of CD79b (DeltaCD79b).40–42 Resulting survival signals in B-CLL are then mediated via protein kinase C, phosphatidylinositol 3-kinase and the serine/threonine kinase Akt.43,44 Whatever the mechanism of these antiapoptotic effects after BCR ligation might be, it enables an almost selective transduction of malignant

B-CLL cells after BCR stimulation by rAAV, while normal B-lymphocytes were shown to be driven in apoptosis by the same stimulus. In our series of patients, we have observed differing responses to IgM ligation with regard to HSPG upregulation and rAAV transduction efficiency. Recently, it was demonstrated that differential signaling via sIgM is closely associated with the VH gene status: cases with unmutated VH genes showed increased tyrosine phosphorylation including activation of Syk.45 Expression of ZAP-70 was shown to be closely correlated with an unmutated IgVH status in B-CLL and with an unfavorable prognosis of the disease.9,14 In other studies, there was also an association observed between CD38 expression and a-IgM responsiveness.10 Importantly, recent data indicate that ZAP-70 expression is associated with enhanced signal transduction via the BCR complex.13 In our study, high levels of ZAP-70 as an indicator of an efficient BCR signaling were highly predictive for a better transduction efficiency by rAAV when the gene transfer was accomplished by stimulation of the BCR. Therefore, a screening analysis for ZAP-70 expression should allow a prediction whether individual CLL samples will be amenable to AAV-based gene transduction. AAV transduction after CD40L stimulation was shown to be enhanced by the addition of CpG oligodeoxynucleotides (CpG-ODNs); however, CpG-ODNs could not replace the strong stimulatory capacity of CD40L for efficient AAV infection.18 CpG-ODNs were also shown to enhance the capacity of recombinant adenovirusmediated gene transfer in a murine B-cell lymphoma model.46 Interestingly, normal and malignant human Blymphocytes express a distinct toll-like receptor repertoire including TLR9 and TLR10, both of them being induced upon engagement of the BCR.47,48 It can be speculated that combined use of a-sIg and CpG-ODNs might be optimal for responsiveness of B-CLL cells for AAV infection without the need for CD40 receptor activation on leukemic cells. Besides recombinant adenovirus that was already successfully established as a gene transfer vehicle for CD40L in a phase I clinical trial,17 our improved feederfree AAV transduction protocol might become ready for clinical testing in the near future. In contrast to other transduction systems for B-CLL, rAAV has advantageous safety features without the threat of the potential activation of oncogenic sequences as described for EBV-derived transduction systems.49 Viral mutants with modified tropism based on a selection process within a library of AAV clones with randomly modified capsids, the AAV display, allow to generate CLL-specific targeting vectors, but safety issues of these vectors have to be addressed before entering the clinic.50 We have shown that after infection with AAV particles encoding CD40L, the immune accessory molecule CD80 was expressed on infected CLL cells, but also on noninfected bystander leukemia B cells. Similar results were previously described for an adenoviral transfer system and also for AAV vectors being used after prestimulation of CLL cells by CD40L-expressing feeder cells.7,15 Using rAAV as a mainly nonimmunogenic vector transfer system together with engagement of the BCR, it might be possible to define CLL-specific T-cell responses in the context of a clinical vaccination protocol based on ex vivo CD40Ltransduced CLL cells.

1421

Gene Therapy


1422

Materials and methods Patients, cells, and cell culture After informed consent, peripheral blood was obtained from patients satisfying diagnostic criteria for B-CLL.51 Mononuclear cells were isolated on a Ficoll/Hypaque (Seromed, Berlin, Germany) density gradient by centrifugation and depleted from monocytes by adherence to plastic tissue culture flasks. More than 98% of isolated cells coexpressed CD5 and CD19, as assessed by flow cytometry. Patients were either untreated or had not received cytoreductive treatment for a period of at least 1 month before investigation. Human tonsillar lymphocytes were isolated from tonsils as described previously.52 HeLa cells were obtained from the American Type Culture Collection (ATCC, Rockville, MD, USA), 293 cells were a gift from M Lohse, University of Wu¨rzburg, Germany. Cells were cultured at 371C in 5% CO2 in air, in the culture medium consisting of DMEM (Biochrom, Berlin, Germany) supplemented with 10% fetal calf serum (FCS; Biochrom), 2 mM L-glutamine (Biochrom), 100 U/ml penicillin (Biochrom), and 100 mg/ml streptomycin (Biochrom). HeLa/SF cells transfected with human CD40L cDNA were produced and cultured as described previously.53 Before the addition of B-CLL cells, the feeder layers were washed twice with phosphate-buffered saline (PBS), and tumor cells were cultured at 2 106 cells/ml in Iscove’s medium (GibcoBRL) supplemented with 20% heatinactivated FCS, 2 mM L-glutamine, 100 U/ml penicillin, and 100 mg/ml streptomycin. For functional assays, CLL cells were harvested, purified by Ficoll density-gradient centrifugation, washed and analyzed by flow cytometry. Antibodies and reagents Immunophenotyping was performed with the following monoclonal antibodies (mAbs) conjugated with fluorescein isothiocyanate, phycoerythrin (PE), or PE cyanine 5 (PE-Cy5): CD5, CD19 (both from Beckman Coulter, Krefeld, Germany), CD38, CD80, anti-IgM, anti-IgG, antiIgA (BDPharMingen, Heidelberg, Germany). Fluorescein-conjugated mAbs specific for murine CD40L were purchased from BDPharMingen and expression was controlled by an isotype hamster IgG3 mAb (BDPharMingen). For the detection of HSPG, cells were stained with fluorescein-conjugated anti-heparan sulfate (10E4 epitope) mAb (Seikagaku America, MA, USA) or a murine IgM isotype control (BDPharMingen). As described previously, intracellular ZAP-70 expression in B-CLL cells was detected with an anti-ZAP-70-specific mAb (Upstate Biotechnology, Waltham, MA, USA) after cells were fixed and permeabilized with use of 4% formaldehyde (Sigma-Aldrich, St Louis, MO, USA) and 0.4% saponin (Sigma-Aldrich).54 Heparin (10 000 U/ml; Braun, Melsungen, Germany) was used for blocking experiments with infectious AAV. Rabbit anti-human Ig ImmunobeadR reagents with heavy-chain specificity for human IgG, IgA or IgM were used for BCR crosslinking and were purchased from Irvine Scientific, Santa Ana, CA, USA. ImmunobeadR reagents consist of purified antibodies covalently bound to micron-sized hydrophilic polyacrylamide beads. RNAse (Boehringer Mannheim, Mannheim, Germany) and propidium iodide (Bender MedSystems Diagnostic Gene Therapy

GmbH, Vienna, Austria) were used for cell cycle analysis. The fraction of apoptotic cells was determined by annexin binding assay using annexin V, annexin buffer, and propidium iodide (Bender MedSystems Diagnostic).

Plasmids The adenoviral pXX6 plasmid was a friendly gift from R Samulski and described previously.55 The constructs pAAV/EGFP coding for the EGFP and pAAV/mCD40L containing the murine CD40L encoding gene were described previously.7 The helper plasmid pXR1 containing the serotype-specific capsid coding domains (cap) of AAV-1 in the context of AAV-2-specific replication genes (rep) was a friendly gift from JE Rabinowitz, University of North Carolina, Chapel Hill, USA, and was used for crosspackaging of AAV-1-specific virions coding for EGFP.56 The plasmid for the AAV mutant RGD4C with a specific RGD motif, which allows AAV to enter a cell by a heparan sulfate-independent cell entry mechanism, was described previously.12 rAAV vector production and purification Packaging of rAAV vectors was performed by cotransfection of 293 cells by calcium phosphate with a total of 37.5 mg vector plasmid (pAAV/EGFP or pAAV/ mCD40L), packaging plasmid pRC, and adenoviral plasmid pXX6 (kindly provided by J Samulski) at a 1 : 1 : 1 molar ratio.55,57 Purification was achieved by ammonium sulfate precipitation followed by iodixanol gradient centrifugation.58,59 This resulted in an infectious AAV/EGFP and AAV/CD40L titer of 15 109/ml. AAV transduction A total of 5 105 primary CLL cells per well (96-well plate) were incubated in a total of 50 ml IMDM medium supplemented with 20% FCS and infectious AAV was added resulting in an MOI of 50. For indicated experiments ImmunobeadR reagent was added to leukemic cells together with rAAV. Cells were incubated for 2 h at 371C in 5% CO2 in air, followed by several washing steps. Infected cells were transferred on an g-irradiated feeder layer expressing CD40L (HeLa/SF) in some experiments, as indicated, and 150 ml Iscove’s medium was added. Flow cytometry At 48 h after AAV transduction, CLL cells were harvested, purified by Ficoll density-gradient centrifugation, and washed. Specific, directly conjugated antibodies were applied to cells for 30 min in PBS, 4% FCS, 0.1% sodium acide, 20 mM HEPES, and 5 mM EDTA pH 7.3 on ice and washed. Nonspecific binding was controlled by incubation with isotypic controls (murine isotype IgG1 mAb and hamster isotype IgG3 mAb, BDPharMingen). Fluorescence was measured with a Coulter Epics XL-MCL (Beckman Coulter, Krefeld, Germany). A minimum of 5000 cells were analyzed for each sample. The percentage of positive cells was defined as the fraction beyond the region of 99% of the control-stained cells. Data were analyzed with the use of WinMDI2.8 FACS software. The MFIR was calculated to compare the relative staining intensities of two or more cell populations.15


1423

Transactivation assay CD40L-transduced or mock-infected B-CLL cells were prelabeled with a green fluorescent dye (CellTrackert Green CMFDA, Molecular Probes) at a concentration of 1 mM for 15 min at 371C. After extensive washing, labeled stimulator cells were cocultured with noninfected, nonstained CLL cells from the same patient at 371C for another 48 h. Expression of CD80 on noninfected, nonlabeled naive CLL cells was assessed by PE-conjugated anti-CD80 mAb (Beckman Coulter). Statistics Statistical associations between dependent subgroups were analyzed by the t-test for paired samples; statistical associations between independent subgroups were carried out using the t-test for unpaired samples. In case of non-normality of variables, the Mann–Whitney U-test for unpaired samples or the Wilcoxon’s test for paired samples was applied. Pearson’s correlation was used for the assessment of the relationship between continuous variables. A statistical significance was accepted for Po0.05. The calculations were determined by the statistical software package SAS, version 8.2.

Acknowledgements We gratefully acknowledge the support of many colleagues who enabled the preparation of this report: Dr J Samulski and Dr J Rabinowitz, University of North Carolina, Chapel Hill, USA, for providing the pXX6 and the pXR1 plasmid, respectively. We thank Dr J Enssle, Genzentrum, Munich for providing the plasmid for the AAV mutant RGD4C. We are grateful to Kristin Leike for excellent technical assistance. We thank our colleagues and the nursing staff form the Medical Clinic III at the KGMC who took care of the patients on the wards and in the outpatient clinic. CMW and MH were supported by grants from the Deutsche Forschungsgemeinschaft (SFB 455), Wilhelm-Sander-Stiftung (1995.056.2) and Else Kro¨ner-Fresenius-Stiftung, Germany.

References 1 Caligaris-Cappio F, Hamblin TJ. B-cell chronic lymphocytic leukemia: a bird of a different feather. J Clin Oncol 1999; 17: 399–408. 2 Lagneaux L et al. Chronic lymphocytic leukemic B cells but not normal B cells are rescued from apoptosis by contact with normal bone marrow stromal cells. Blood 1998; 91: 2387–2396. 3 Kitada S, Zapata JM, Andreeff M, Reed JC. Bryostatin and CD40ligand enhance apoptosis resistance and induce expression of cell survival genes in B-cell chronic lymphocytic leukaemia. Br J Haematol 1999; 106: 995–1004. 4 Meinhardt G, Wendtner C-M, Hallek M. Molecular pathogenesis of chronic lymphocytic leukemia: factors and signaling pathways regulating cell growth and survival. J Mol Med 1999; 77: 282–293. 5 Burger JA et al. Blood-derived nurse-like cells protect chronic lymphocytic leukemia B cells from spontaneous apoptosis through stromal cell-derived factor-1. Blood 2000; 96: 2655–2663. 6 Bernal A et al. Survival of leukemic B cells promoted by engagement of the antigen receptor. Blood 2001; 98: 3050–3057.

7 Wendtner C-M et al. Efficient gene transfer of CD40 ligand into primary B-CLL cells using recombinant adeno-associated virus (rAAV) vectors. Blood 2002; 100: 1655–1661. 8 Fais F et al. Chronic lymphocytic leukemia B cells express restricted sets of mutated and unmutated antigen receptors. J Clin Invest 1998; 102: 1515–1525. 9 Crespo M et al. ZAP-70 expression as a surrogate for immunoglobulin-variable-region mutations in chronic lymphocytic leukemia. N Engl J Med 2003; 348: 1764–1775. 10 Zupo S et al. CD38 expression distinguishes two groups of B-cell chronic lymphocytic leukemias with different responses to antiIgM antibodies and propensity to apoptosis. Blood 1996; 88: 1365–1374. 11 Summerford C, Bartlett JS, Samulski RJ. AlphaVbeta5 integrin: a co-receptor for adeno-associated virus type 2 infection. Nat Med 1999; 5: 78–82. 12 Shi W, Bartlett JS. RGD inclusion in VP3 provides adenoassociated virus type 2 (AAV2)-based vectors with a heparan sulfate-independent cell entry mechanism. Mol Ther 2003; 7: 515–525. 13 Chen L et al. Expression of ZAP-70 is associated with increased B-cell receptor signalling in chronic lymphocytic leukaemia. Blood 2002; 100: 4609–4614. 14 Wiestner A et al. ZAP-70 expression identifies a chronic lymphocytic leukaemia subtype with unmutated immunoglobulin genes, inferior clinical outcome, and distinct gene expression profile. Blood 2003; 101: 4944–4951. 15 Kato K, Cantwell MJ, Sharma S, Kipps TJ. Gene transfer of CD40-ligand induces autologous immune recognition of chronic lymphocytic leukemia B cells. J Clin Invest 1998; 101: 1131–1141. 16 Cantwell MJ, Hua T, Pappas J, Kipps TJ. Acquired CD40-ligand deficiency in chronic lymphocytic leukemia. Nat Med 1997; 3: 984–989. 17 Wierda G et al. CD40-ligand (CD154) gene therapy for chronic lymphocytic leukemia. Blood 2000; 96: 2917–2924. 18 Theiss HD et al. Enhancement of gene transfer with recombinant adeno-associated virus vectors into primary B-CLL cells by CpG-oligodeoxynucleotides. Exp Hematol 2003; 31: 1223–1229. 19 Van der Voort R et al. Regulation of cytokine signaling by B cell antigen receptor and CD40-controlled expression of heparan sulfate proteoglycans. J Exp Med 2000; 192: 1115–1124. 20 Maass G et al. Recombinant adeno-associated virus for the generation of autologous, gene-modified tumor vaccines: evidence for a high transduction efficiency into primary epithelial cancer cells. Hum Gene Ther 1998; 9: 1049–1059. 21 Zhang L et al. Efficient expression of CFTR function with adenoassociated virus vectors that carry shortened CFTR genes. Proc Natl Acad Sci USA 1998; 95: 10158–10163. 22 Wang B, Li J, Xiao X. Adeno-associated virus vector carrying human minidystrophin genes effectively ameliorates muscular dystrophy in mdx mouse model. Proc Natl Acad Sci USA 2000; 97: 13714–13719. 23 Takahashi H et al. Long-term systemic therapy of Fabry disease in a knockout mouse by adeno-associated virus-mediated muscle-directed gene transfer. Proc Natl Acad Sci USA 2002; 99: 13777–13782. 24 Manno CS et al. AAV-mediated factor IX gene transfer to skeletal muscle in patients with severe hemophilia B. Blood 2003; 101: 2963–2972. 25 Haxhinasto SA, Hostager BS, Bishop GA. Cutting Edge: Molecular mechanism of synergy between CD40 and the B cell antigen receptor: role for TNF receptor-associated factor 2 in receptor interaction. J Immunol 2002; 169: 1145–1149. 26 Tanaka Y et al. Heparan sulfate proteoglycan on leukemic cells is primarily involved in integrin triggering and its mediated adhesion to endothelial cells. J Exp Med 1996; 184: 1987–1997. Gene Therapy


1424 27 Kopper L, Sebestyen A. Syndecans and the lymphoid system. Leukemia Lymphoma 2000; 38: 271–281. 28 Burger JA, Burger M, Kipps TJ. Chronic lymphocytic leukemia B cells express functional CXCR4 chemokine receptors that mediate spontaneous migration beneath bone marrow stromal cells. Blood 1999; 94: 3658–3667. 29 Netelenbos T et al. Proteoglycans guide SDF-1-induced migration of hematopoietic progenitor cells. J Leukocyte Biol 2002; 72: 353–362. 30 Makatsori E, Tsegenidis T, Karamanos NK. Effects of mitogens on synthesis and distribution of heparan and chondroitin sulphates by human leukemic B, T cells and monocytes studied by high-performance liquid chromatography and radiochemical detection. Biomed Chromatogr 2001; 15: 534–538. 31 Bone H, Williams NA. Antigen-receptor cross-linking and lipopolysaccharide trigger distinct phosphoinositide 3-kinasedependent pathways to NF-kappa B activation in primary B cells. Int Immunol 2001; 13: 807–816. 32 Teoh G et al. CD40 activation mediates p53-dependent cell cycle regulation in human multiple myeloma cell lines. Blood 2000; 95: 1039–1046. 33 Granziero L et al. Survivin is expressed on CD40 stimulation and interfaces proliferation and apoptosis in B-cell chronic lymphocytic leukemia. Blood 2001; 97: 2777–2783. 34 McConkey DJ et al. Induction of DNA fragmentation in chronic B-lymphocytic leukemia cells. J Immunol 1991; 146: 1072–1076. 35 Zupo S et al. Apoptosis or plasma cell differentiation of CD38positive B-chronic lymphocytic leukaemia cells induced by cross-linking of surface IgM or IgD. Blood 2000; 95: 1199–1206. 36 Osorio LM et al. CD6 ligation modulates the Bcl-2/Bax ratio and protects chronic lymphocytic leukemia B cells from apoptosis induced by anti-IgM. Blood 1997; 89: 2833–2841. 37 Lankester AC et al. CD5 is associated with the human B cell antigen receptor complex. Eur J Immunol 1994; 24: 812–816. 38 Pers JO et al. CD5-induced apoptosis of B cells in some patients with chronic lymphocytic leukemia. Leukemia 2002; 16: 44–52. 39 Cantor GH et al. CD5 is dissociated from the B-cell receptor in B cells from bovine virus-infected, persistently lymphocytotic cattle: consequences to B-cell receptor-mediated apoptosis. J Virol 2001; 75: 1689–1696. 40 Thompson AA et al. Aberrations of the B-cell receptor B29 (CD79b) gene in chronic lymphocytic leukaemia. Blood 1997; 90: 1387–1394. 41 Alfarano A et al. An alternativley spliced form of CD79b gene may account for altered B-cell receptor expression in B-chronic lymphocytic leukemia. Blood 1999; 93: 2327–2335. 42 Cragg MS et al. The alternative transcript of CD79b is overexpressed in B-CLL and inhibits signaling in apoptosis. Blood 2002; 100: 3068–3076. 43 Pogue SL, Kurosaki T, Bolen J, Herbst R. B cell antigen receptorinduced activation of Akt promotes B cell survival and is dependent on Syk kinase. J Immunol 2000; 165: 1300–1306. 44 Barragan M et al. Involvement of protein kinase C and phosphatidylinositol 3-kinase pathways in the survival of B-cell chronic lymphocytic leukemia cells. Blood 2002; 99: 2969–2976.

Gene Therapy

45 Lanham S et al. Differential signaling via surface IgM is associated with VH gene mutational status and CD38 expression in chronic lymphocytic leukemia. Blood 2003; 101: 1087–1093. 46 Rieger R, Kipps TJ. CpG oligodeoxynucleotides enhance the capacity of adenovirus-mediated CD154 gene transfer to generate effective B-cell lymphoma vaccines. Cancer Res 2003; 63: 4128–4135. 47 Bernasconi NL, Onai N, Lanzavecchia A. A role for Toll-like receptors in acquired immunity: upregulation of TLR9 by BCR triggering in naive B cells and constitutive expression in memory B cells. Blood 2003; 11: 4500–4504. 48 Bourke E et al. The toll-like receptor repertoire of human B lymphocytes: inducible and selective expression of TLR9 and TLR10 in normal and transformed cells. Blood 2003; 102: 956–963. 49 Wendtner C-M et al. High level of transgene expression in primary chronic lymphocytic leukemia cells using helper-virusfree recombinant Epstein-Barr virus vectors. Exp Hematol 2003; 31: 99–108. 50 Perabo L et al. In vitro selection of viral vectors with modified tropism: the adeno-associated virus display. Mol Ther 2003; 8: 151–157. 51 Cheson B et al. National Cancer Institute – Sponsored Working Group guidelines for chronic lymphocytic leukemia: revised guidelines for diagnosis and treatment. Blood 1996; 87: 4990–4997. 52 Koopman G et al. Adhesion through the LFA-1 (CD11a/CD18)ICAM-1 (CD54) and the VLA-4 (CD49d)-VCAM-1 (CD106) pathways prevents apoptosis of germinal center B cells. J Immunol 1994; 152: 3760–3767. 53 Buhmann R et al. CD40-activated B-cell chronic lymphocytic leukemia cells for tumor immunotherapy: stimulation of allogeneic versus autologous T cells generates different types of effector cells. Blood 1999; 93: 1992–2002. 54 Emoto M et al. Critical role of NK cells rather than V14+NKT cells in lipopolysaccharide-induced lethal shock in mice. J Immunol 2002; 169: 1426–1432. 55 Nathwani AC et al. Efficient gene transfer into human cord blood CD34+ cells and the CD34+CD38 subset using highly purified recombinant adeno-associated viral vector preparations that are free of helper virus and wild-type AAV. Gene Therapy 2000; 7: 183–195. 56 Rabinowitz JE et al. Cross-packaging of a single adenoassociated virus (AAV) type 2 vector genome into multiple AAV serotypes enables transduction with broad specificity. J Virol 2002; 76: 791–801. 57 Girod A et al. Genetic capsid modifications allow efficient retargeting of adeno-associated virus type 2. Nat Med 1999; 5: 1052–1056. 58 Summerford C, Samulski RJ. Membrane-associated heparan sulfate proteoglycan is a receptor for adeno-associated virus type 2 virions. J Virol 1998; 72: 1438–1445. 59 Zolotukhin S et al. Recombinant adeno-associated virus purification using novel methods improves infectious titer and yield. Gene Therapy 1999; 6: 973–985.