Department of Dermatology and University of Colorado Cancer Center, University of Colorado Health Sciences Center, Denver, CO, USA

Correspondence to: I H Maxwell, Department of Dermatology B153, University of Colorado Health Sciences Center, 4200 East Ninth Avenue, Denver, CO 80262, USA

The autonomous parvoviruses are small, non-enveloped, single strand DNA viruses. They occur in many species and they have oncolytic properties. We are modifying the capsid of feline panleukopenia virus (FPV), a parvovirus which normally infects feline cells, with the goal of targeting human tumor cells for potential cancer therapy. Using recombinant viruses transducing a luciferase reporter, we show that insertion of a cyclically constrained, integrin-binding peptide at an exposed position on the FPV capsid enables transduction of an v integrin-expressing human rhabdomyosarcoma cell line (Rh18A). These cells were not transduced by virus with the unmodified FPV capsid. Transduction of Rh18A was specifically inhibited by an v integrin blocking antibody. However, other human tumor lines expressing v integrins were not transduced by virus with either the modified or unmodified capsid. We conclude that modification of the FPV capsid to bind v integrins can contribute to, but is not generally sufficient for, redirecting infection to human tumor cells. The permissiveness of Rh18A cells presumably involves additional factors unique to this line among various human cell lines tested. Gene Therapy (2001) 8, 324-331.

feline parvovirus; human tumor cells; integrin-binding peptide; rhabdomyosarcoma; retargeting

Introduction

Autonomous parvoviruses are small, icosahedral, single strand DNA viruses that replicate preferentially in tumor cells^1,2,3,4 and, unlike the adeno-associated parvovirus, AAV, do not require a helper virus for replication. Several members of the rodent group of autonomous parvoviruses replicate efficiently in human transformed cell lines.¹ However, feline panleukopenia virus (FPV) and canine parvovirus (CPV) use different receptors from the related rodent viruses and these viruses have not been found to infect human cells efficiently. We hypothesized that FPV or CPV might be directed to novel receptors on human tumor cells by inserting peptide ligands for specific cell surface molecules into their capsid proteins. By this means such a virus might be targeted specifically to tumor cells in a patient to achieve virus-mediated cytotoxicity or therapeutic gene delivery. Detailed structural data from X-ray crystallography of virions of both CPV and the closely related FPV^5,6 allow reasonable predictions as to positions where peptides could be inserted into the major capsid protein, VP2, and displayed on the virion surface. The crystallographic studies have shown that prominent spikes at the three-fold axes of symmetry of the virion are formed from polypeptide loops interwound among adjacent capsomers. These loops are interspersed between the -strand regions of VP2.⁵ Here, we made insertions at the extremity of loop 2 where there was evidence that small peptide substitutions or insertions could be tolerated by the CPV capsid.^7,8

We chose to target cell surface integrins in view of the fact that these molecules function as receptors for several viruses.^9,10,11,12 In particular, integrins of the v family can function as secondary receptors for both adenoviruses⁹ and adeno-associated virus (AAV).¹² Furthermore, v integrins are abundantly expressed not only by various tumor cells,^13,14 but also by endothelial cells in tumor-associated vasculature.¹⁵ Integrin v3 is frequently up-regulated on tumor cells during malignant progression.^13,14,16 Specific integrins participate in signal transduction by associating with components of the extra-cellular matrix (ECM - fibronectin, vitronectin, laminins, etc) via binding to well-characterized peptide motifs present in ECM proteins.^17,18,19 The corresponding synthetic peptides can compete for binding to the integrin. The most common binding motif (for various integrins, including v3) is the sequence RGD (arg.gly.asp), the affinity for particular integrins being influenced by the amino acids flanking this motif.¹⁹ Peptides containing these sequences bind integrins with greatly increased affinity if constrained in a cyclic conformation imposed by disulfide bridges, formed by flanking Cys residues.²⁰ Such peptides might be excellent candidates for retargeting a parvovirus if appropriately displayed on the capsid.

We previously showed that recombinant genomes derived from the rodent parvovirus, LuIII, could be packaged by capsids of FPV or CPV.²¹ Transduction of a luciferase reporter gene in a packaged LuIII genome can thus be used as a sensitive assay for function of targeted cell surface molecules as virus receptors. Here, we establish proof-of-principle for parvovirus retargeting by showing that insertion of a particular integrin-binding peptide (C4-RGD) into the FPV capsid allows transduction of a human rhabdomyosarcoma cell line (Rh18A).

Results

Generation of recombinant virions with an FPV capsid, modified by insertion of integrin-binding peptides

Recombinant parvoviruses are produced by cotransfection of a plasmid containing a partially deleted viral genome with one or more plasmids expressing helper functions, including the capsid proteins.^21,22 We used an expression plasmid encoding VP1 and VP2 of FPV to generate modified capsids with peptide insertions at the extremity of loop 2 on the virion surface.⁵ To facilitate characterization of the resulting viruses, the modified capsid proteins were used to package a recombinant parvovirus genome expressing a luciferase reporter gene,^21,22,23 as shown in Figure 1. In this way, virion production could be assayed using PCR to detect the packaged recombinant genome and successful infection (transduction) of target cells could be determined by luciferase expression. We focused on modifying the FPV, rather than the CPV, capsid since preliminary results (not shown) indicated that this gave better yields of transducing virus.

We inserted synthetic oligonucleotides encoding either of two, 10 amino acid, peptides into the SpeI restriction site in the VP1/VP2 common sequence. This resulted in insertion of the peptide sequences between amino acids 226 and 227 (VP2 numbering) in the capsid proteins, as indicated in Figure 2. Each peptide contained a cyclically constrained RGD integrin-binding sequence,^20,24 designated C2-RGD or C4-RGD, where C2 and C4 refer to the number of cysteine residues. These peptides were chosen on the basis of their previously reported properties.^20,24,25 Peptide C2-RGD showed high affinity for IIb3 integrin²⁴ and, when displayed on filamentous bacteriophages, conferred cell binding and penetration of the phages into human HEp-2 cells.²⁵ Peptide C4-RGD was identified by phage library screening as having high affinity for v3 and v5 integrins.²⁰

Plasmids expressing the above modified FPV capsid proteins were cotransfected with pGLuP38LUC2 into 324K cells and cell extracts plus culture medium were collected after 3 days (Figure 1). Quantitative PCR analysis using luciferase primers showed that the recombinant genome was packaged (protected from DNase) by the modified proteins (Figure 3). The yield of virions obtained from producer cells expressing FPV capsid proteins with the C4-RGD modification was decreased approximately 12-fold relative to those expressing wild-type capsid. A further two-fold decrease was seen with the C2-RGD modification (Figure 3).

Figure 4 shows the transducing activity observed in feline cells (CFK cell line; permissive host cells for FPV) after infection with recombinant virus with an unmodified FPV capsid, in comparison with virions generated using the RGD-modified capsids. Note that, in Figure 4a, where equal volumes of the virus preparations were used, the C4-RGD modified capsid gave three-fold increased activity over the wild-type, unmodified capsid, whereas the C2-RGD modified capsid gave decreased activity (approximately one twentieth the wild type). Allowing for the 12-fold decrease in yield (Figure 3) these results imply that the C4-RGD modification produced a 36-fold increase in specific infectivity for CFK cells. To substantiate this conclusion, we repeated the transduction of these cells using equivalent multiplicity of infection (expressed as packaged genomes per cell) of virus with the wild-type and C4-RGD capsid, as shown in Figure 4b. The results directly demonstrated a 35-40-fold increase in specific infectivity due to the capsid modification (Figure 4b), suggesting that a novel, perhaps integrin-mediated, interaction on the surface of these cells may have facilitated virus entry.

Transduction of human cells by virus with the C4-RGD modified FPV capsid

As we previously reported,²¹ virions generated using unmodified CPV or FPV capsids gave no detectable transducing activity in human cell lines. Upon initial screening for transduction of human tumor cell lines by the modified, recombinant viruses described above, we found that one line, the rhabdomyosarcoma Rh18A, showed efficient transduction by the C4-RGD virus. As shown in Figure 5a, Rh18A cells showed essentially no activity with virus having the wild-type FPV capsid, but expressed abundant activity after infection with virus having the C4-RGD modified FPV capsid. To confirm involvement of the inserted peptide sequence, we infected Rh18A cells in the presence of free C4-RGD peptide. As shown in Figure 5b, this peptide efficiently inhibited transduction by virus with the C4-RGD modified FPV capsid (presumably by competing for binding to cell surface integrins). In contrast, the peptide showed no specific inhibition of transduction by virus containing the same recombinant genome, packaged with the LuIII capsid (Figure 5b). (The latter allows transduction of various human cells,²² including Rh18A (unpublished observations).) We also found that transduction by the C4-RGD virus could be specifically competed by addition of a purified integrin (IIb3), known to bind RGD motifs (data not shown), further confirming involvement of the RGD sequence in Rh18A transduction. These results therefore demonstrate 'proof-of-principle' for retargeting the feline viral capsid to human cells by insertion of an appropriate peptide.

Interestingly, the C2-RGD modification allowed transduction of Rh18A cells only with very much lower efficiency than the C4-RGD (see legend to Figure 5a), further confirming that efficient transduction of Rh18A was a specific property of the C4-RGD peptide. The C2-RGD virus was not investigated further, in view of its low activity.

The C4-RGD modified virus was then tested for transducing activity in a panel of human cell lines, including melanomas, carcinomas and several other rhabdomyosarcomas, with the results shown in Table 1. These cells mostly displayed little or no transduction, although certain lines variably showed a few percent of the activity seen in Rh18A cells. We had obtained Rh18A as a subline derived by passaging in other laboratories of Rh18 rhabdomyosarcoma cells.²⁶ Surprisingly, when we subsequently tested Rh18 cells, obtained from the laboratory where this line was derived from a rhabdomyosarcoma tumor,²⁶ we detected no transduction by the C4-RGD virus.

To authenticate their human origin, Rh18A cells were subjected to karyotypic analysis (Cytogenetics Core Facility, University of Colorado Cancer Center). This analysis indicated an aneuploid complement of chromosomes which showed specific hybridization with a human centromeric probe by fluorescence in situ hybridization. To exclude the possibility that transduction of a minor contaminant consisting of some other permissive cell type was being observed, we isolated single cell clones from Rh18A and tested these for transduction by the C4-RGD virus. Seven of eight clones were transduced with efficiency comparable with the parental R18A cell population (results not shown), indicating that this property was not restricted to a small subfraction of the cells.

Involvement of v integrins in transduction of Rh18A cells

The C4-RGD peptide binds integrins v3 and v5 with high affinity.²⁰ Expression of v integrins by Rh18A cells was confirmed by subjecting these cells to flow cytometry after reaction with monoclonal antibodies recognizing v3 (LM609), v5 (P1F6) or pan-v integrins (L230) and fluorescein-conjugated secondary antibodies. Abundant fluorescence was observed with each antibody, confirming cell surface expression of these integrins (results not shown). We then examined whether the function-blocking²⁷ pan-v, L230 antibody would prevent transduction. As shown in Figure 6, this antibody indeed strongly inhibited transduction of Rh18A by virus with the C4-RGD modified FPV capsid, but showed only minimal effect on transduction of these cells by the same recombinant genome when packaged by the LuIII capsid. These results confirm the involvement of v integrins in transduction of Rh18A cells by virus with the modified capsid.

v integrins are expressed by cells of many human tumor types and v3 is considered a marker of tumor progression.^13,14,16 Several of the cell lines negative for transduction in Table 1 have been reported to express v3. We confimed by flow cytometry that the melanoma line, Hs294T, and the rhabdomyosarcoma line, Rh18, expressed v integrins including v3. Rh18 was also positive for v5. In these assays, both these lines displayed somewhat higher levels of fluorescence than did Rh18A cells with each of the anti-v monoclonal antibodies tested (results not shown). Since neither Hs294T nor Rh18 cells were transduced by the C4-RGD modified virus (Table 1), these results taken together indicate that v integrin expression is necessary but not sufficient for transduction of human cells by this virus.

Transduction of Rh18A cells by virus with C4-RGD inserted in either of the capsid proteins, VP1 or VP2

The FPV minor and major capsid proteins, VP1 and VP2, are normally translated from alternatively spliced mRNAs.^23,28 However, in our recombinant system these proteins can be supplied from separate expression plasmids,^23,29 enabling these proteins to be independently modified by peptide insertion. We generated recombinant virions with C4-RGD separately inserted into VP1 or VP2 at the same SpeI site in their shared coding sequence, with the other protein retaining the wild-type sequence. As shown in Figure 7, equivalent transduction of Rh18A cells was observed regardless of whether the C4-RGD sequence was present only in VP1 or VP2, or in both proteins. We also generated a virus with insertion into VP1 of an alternative peptide that was identical to C4-RGD except that the central RGD motif (Figure 2), conferring integrin binding, was replaced by a different trimer of amino acids (NGR). This control virus showed no detectable transducing activity for Rh18A cells, although it was able to transduce the feline, CFK, cell line efficiently (data not shown). This result again confirmed the specificity of the C4-RGD, integrin-binding peptide in enabling transduction of the Rh18A, human rhabdomyosarcoma line.

Discussion

Autonomous parvoviruses of the rodent group preferentially replicate in, and kill, malignantly transformed cells.³ They can also be used as vectors for gene delivery.^4,22,30 These viruses therefore have potential as therapeutic agents against cancer, either by exploiting their natural oncolytic properties or by using them for delivery of therapeutic genes. Prospects for in vivo use would be improved by viral modification conferring specific targeting of infection to human tumor cells. As an initial step in this direction we have shown that the capsid of FPV can be modified to allow infection (transduction) of a human rhabdomyosarcoma cell line that is normally refractory to this virus.

We inserted either of two integrin-binding peptides into the extremity of loop 2 of the FPV capsid. Packaging of recombinant parvoviral DNA was observed with either modification (Figure 3), indicating that these insertions did not disrupt capsid assembly. However, the yields of packaged DNA were lower than with the wild-type capsid (Figure 3). Since we observed similar stability of transducing activity during prolonged storage of virus preparations with wild-type or modified capsid it is likely that the decreased yields were due to less efficient production, rather than lower stability, of the modified virions.

Using a luciferase transduction assay we showed that the modified viruses retained the ability to infect their normal host cells, ie the feline cell line, CFK. In fact, the C4-RGD insertion significantly increased transducing activity for these cells (Figure 4). More importantly, the C4-RGD virus was able to transduce the human rhabdomyosarcoma cell line, Rh18A, although these cells did not support transduction by virus with the unmodified, wild-type FPV capsid (Figure 5a), nor by a control virus with an inserted peptide identical to C4-RGD except for a substitution of the central, integrin-binding RGD trimer (see last section of Results). Transduction of Rh18A by the C4-RGD virus was inhibited either by free C4-RGD peptide or by a pan-v integrin antibody (Figures 5b and 6), confirming dependence on interaction of the inserted motif with v-integrin(s) on the cell surface. These results therefore establish proof-of-principle for targeting of the FPV capsid to human tumor cells.

We chose to use the extremity of loop 2 for peptide display both because of its location on the virion surface⁵ and because studies in a baculovirus system suggested that modifications close to this position were well tolerated for CPV capsid production.^7,8 We also showed that equivalent targeting of Rh18A cells could be achieved by insertion of C4-RGD into either the major capsid protein, VP2, or into the minor protein, VP1, at this same position in their common sequence. There are a total of 60 molecules of the capsid proteins per virion, approximately 10 of which are VP1. VP1 differs from VP2 only by the presence of an N-terminal extension of 143 amino acids. These results imply that the surface disposition of VP1 is similar to that of VP2, at least for the loop 2 region.

To our knowledge, this is the first report of the genetic modification of an autonomous parvovirus capsid to confer an altered host range via peptide insertion. However, insertion of essentially the same C4-RGD peptide as used here into the fiber protein of the adenovirus type 5 capsid conferred an expanded host range among human cells, by circumventing the requirement for interaction of adenovirus vectors with their normal primary receptor on the cell surface.^31,32 There have been several recent reports of experiments to modify the host range of the dependent parvovirus, AAV.^33,34,35,36 Girod et al³⁵ made peptide insertions in the AAV capsid at putatively exposed positions (predicted by computer modeling of structural homology with CPV). Insertion of an integrin-binding peptide (derived from laminin) into one of six such positions enabled infection of cells normally refractory to AAV. Although this insertion did not interfere with virion production (as determined from the amount of packaged DNA), infectivity for cells permissive for AAV was reduced by at least 100-fold, while remaining 10-fold higher than that for the targeted non-permissive cells. It was not determined whether this was due to a general impairment of infectivity of the modified virus at some post-receptor level, or to impairment of binding to the normal AAV receptor or co-receptors.^12,37,38 Our results with modification of the FPV capsid differ in that we observed an increase in transduction of the normal CFK host cells with the C4-RGD peptide insertion, as well as targeting of transduction to the Rh18A human cell line with an efficiency approaching that seen in the CFK cells.

Our results indicate that interaction of the peptide displayed by the C4-RGD virus with cell surface v integrin(s) was necessary for transduction of human Rh18A cells but that this interaction was not sufficient for targeting various other human tumor lines expressing v-integrins (Table 1). The basis for the unique permissiveness of Rh18A for the modified FPV capsid is unknown. A working hypothesis, that these cells may express a product which the modified virus can use as an ancillary receptor in conjunction with an v-integrin, is being tested in gene expression studies in progress. With improvement in understanding of the mechanisms by which targeting can be achieved, capsid-modified parvoviruses should prove valuable as local delivery vectors in basic studies of gene function and as therapeutic agents for metastatic disease.

Materials and methods

Plasmids

Plasmid pF-VP, described previously,²¹ expresses both the major and minor capsid proteins (VP2 and VP1, respectively) of FPV by alternative splicing of a common transcript from the viral P38 promoter. (These proteins share most of their amino acid sequence but VP1 has an N-terminal extension of 143 amino acids.)^28,39 Plasmids pF.VP1 and pF.VP2, expressing either VP1 or VP2 alone, have also been described.²³ For most experiments, virions were produced containing the GLuP38LUC2 LuIII-luciferase genome (see Figure 1), by cotransfection of capsid protein expression plasmids with pGLuP38LUC2²¹ which expresses LuIII non-structural (NS) proteins. For some experiments, virions were produced containing the alternative GLuP4LUC1 genome^22,40 by supplying NS helper functions from the additionally cotransfected plasmid, pRSVhel.²¹

Insertion of sequences encoding C4-RGD or C2-RGD into FPV capsid protein expression plasmids

We chose for insertions a position at the extremity of loop 2⁵ where an SpeI site is present in the shared coding sequence of VP1 and VP2 (at VP2 nt 673).^21,28,39 Oligonucleotides encoding the C4-RGD amino acid sequence shown in Figure 2 were ligated into this restriction site in pFVP. The nucleotide sequences were chosen to incorporate the most frequently used codon for each amino acid in the FPV VP2 sequence.^28,39 Oligonucleotides used were: 5'-CT AGT GGT GCA TGT GAT TGT AGA GGT GAT TGT TTT TGT GGT G (plus strand 42-mer; triplets indicate the codons) and 5'-CT AGC ACC ACA AAA ACA ATC ACC TCT ACA ATC ACA TGC ACC A (minus strand 42-mer; both from Gibco BRL, Life Technologies, Gaithersburg, MD, USA). These complementary oligonucleotides were designed with protruding ends (5'-CTAG) capable of annealing with the sticky ends of the cleaved SpeI site. The fifth nucleotide of the plus strand was T, and that of the minus strand was C, so that an SpeI site would be regenerated only at the left end of the insertion (with respect to the coding sequence). The mixed oligonucleotides (non-phosphorylated; 10 pmole each) were ligated with SpeI-cleaved pFVP (0.05 pmole) by a linker tailing method.⁴¹ The plasmid was purified from excess oligonucleotides using a Wizard PCR column (Promega, Madison, WI, USA) and the tailed ends were then annealed in buffer containing 50 mM NaCl and 10% v/v DMSO, by slowly cooling from 70°C to room temperature. The product was transfected into E. coli, strain SURE (Stratagene, La Jolla, CA, USA) by electroporation and clones were screened for the correctly oriented insert by restriction mapping, and confirmed by automated DNA sequencing (Sequencing Core, University of Colorado Cancer Center).

Oligonucleotides encoding the C2-RGD sequence (Figure 2) were: 5'-CT AGT GGT GGT TGT AGA GGT GAT ATG TTT GGT TGT GGT G (plus strand 39-mer) and 5'-CT AGC ACC ACA ACC AAA CAT ATC ACC TCT ACA ACC ACC A (minus strand 39-mer). These were inserted in a similar way except that, instead of linker tailing, the phosphorylated oligonucleotides (0.15 pmole each) were ligated with phosphatase-treated, SpeI-cleaved pFVP (0.05 pmole). The correct insert was again confirmed by DNA sequencing.

Derivatives of pFVP1 and pFVP2 with the C4-RGD-encoding oligonucleotide insertion were constructed by exchange of the relevant restriction fragment from pFVP with this insertion (details available on request).

Cell culture and virus production

Cell lines were grown in RPMI1640 with 10% NuSerum IV (Collaborative Biomedical Products; Becton Dickinson, Bedford, MA, USA) at 37°C in an atmosphere of 5% CO₂. Virus production was as described^21,22 using electroporation to cotransfect the 324K cell line (or, in some experiments, COS cells) with a plasmid containing a LuIII-luciferase recombinant genome and one or more plasmids encoding wild-type or modified FPV capsid proteins. Virus with the peptide insert in either VP1 or VP2 (Figure 7) was obtained by cotransfection with combinations of pF-VP1 or pF-VP2²³ with their derivative plasmids containing the C4-RGD insert. Virus stocks, consisting of combined culture medium and freeze- thawed extracts of the electroporated producer cells were stored at 4°C.

Assay for transduction of recipient cells

Recipient cells were seeded in 12-well plates (5 ´ 10⁴ cells per well) one day before infection with virus samples diluted in medium (0.25 ml total volume per well). After 1.5-2 h at 37°C, with frequent swirling, more medium (0.75 ml) was added and incubation was continued for 2 days before luciferase assay (Promega kit with a TD-20/20 luminometer (Turner Designs, Sunnyvale, CA, USA)).

For peptide competition (Figure 5b), cells were infected as above for 2.5 h, in presence or absence of peptide, washed with PBS and then incubated with fresh medium for 2 days before luciferase assay. Since the C4-RGD peptide caused some cell rounding and detachment during the infection period, detached cells were recovered by centrifugation at the end of this period, washed, and returned to the well containing the fresh medium.

The hybridoma M-23 producing the pan-v, function-blocking monoclonal antibody L230 was obtained from ATCC (Rockville, MD, USA; catalog number HB-8448). For competition by this antibody (Figure 6), cells were infected as above in a total volume of 0.25 ml, containing up to 60 l culture supernatant from the hybridoma, for 2 h. The cells were washed with PBS and were then incubated in fresh medium for 2 days before luciferase assay.

Assay for DNA packaged in virions by quantitative real-time PCR

To degrade any residual plasmid DNA, samples of virus stocks were first incubated with DNase I (Sigma, St Louis, MO, USA; crude; 200 g/ ml for approximately 4 h at 37°C followed by 15 min at 64°C to inactivate the DNase).⁴² Control experiments involving addition of luciferase plasmid DNA showed that the DNase treatment was effective in eliminating any PCR signal from this source (data not shown). Serially diluted samples were then subjected to PCR using primers for luciferase (5'-TGGCCCTTCCGCATAGAA and 5'-TGATTGCC AAAAATAGGATCTCTG) together with a quenched-fluorescence-tagged Taqman probe⁴³ hybridizing to the luciferase sequence between the primers (5'-FAM.TGCCTGCGTCAGATTCTCGCATG.TAMRA). PCR was performed (ABI Prism 7700 Sequence Detector, PE Biosystems, Foster City, CA, USA) and quantification was obtained from cycle numbers for threshold appearance of fluorescence, relative to a standard curve for dilutions of linearized luciferase plasmid DNA amplified in parallel (PCR Core facility, UCHSC Cancer Center). Under PCR conditions viral DNA becomes available as template for amplification due to disruption of the virions during the initial incubation at 95°C.^21,42

Acknowledgements

This work was supported by seed grants from the Melanoma Research Foundation and the University of Colorado Cancer Center. JAC was the recipient of a postdoctoral fellowship from the Susan G Komen Foundation. We thank Dr W Arap for generous donation of peptides. Service support was provided by the Tissue Culture, DNA Sequencing, Flow Cytometry, Cytogenetics, and PCR Core facilities of the Cancer Center (NIH grant number CA46934).

1 Siegl G. The Parvoviruses. Virology Monographs, volume 15. Springer-Verlag: New York, 1976,

2 Cotmore S, Tattersall, P. The autonomously replicating parvoviruses of vertebrates. Adv Virus Res 1987; 33: 91-174, MEDLINE

3 Rommelaere J, Tattersall P. Oncosuppression by parvoviruses. In: Tijssen P (ed). Handbook of Parvoviruses. Volume 2, Chapter 3: CRC Press: Boca Raton, 1990, pp 41-57.

4 Corsini J, Afanasiev B, Maxwell IH, Carlson JO. Autonomous parvovirus and densovirus gene vectors. Adv Virus Res 1996; 47: 303-351, MEDLINE

5 Tsao J et al. The three-dimensional structure of canine parvovirus and its functional implications. Science 1991; 251: 1456-1464, MEDLINE

6 Agbandje M et al. Structure determination of feline panleukopenia virus empty particles. Proteins 1993; 16: 155-171, MEDLINE

7 Hurtado A et al. Identification of domains in canine parvovirus VP2 essential for the assembly of virus-like particles. J Virol 1996; 70: 5422-5429, MEDLINE

8 Rueda P et al. Minor displacements in the insertion site provoke major differences in the induction of antibody responses by chimeric parvovirus-like particles. Virology 1999; 263: 89-99, MEDLINE

9 Wickham TJ, Mathias P, Cheresh DA, Nemerow GR. Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell 1993; 73: 309-319, MEDLINE

10 Berinstein A et al. Antibodies to the vitronectin receptor (integrin alpha V beta 3) inhibit binding and infection of foot-and-mouth disease virus to cultured cells. J Virol 1995; 69: 2664-2666, MEDLINE

11 Evander M et al. Identification of the alpha6 integrin as a candidate receptor for papillomaviruses. J Virol 1997; 71: 2449-2456, MEDLINE

12 Summerford C, Samulski RJ. V5 integrin: a co-receptor for adeno-associated virus type 2 infection. Nat Med 1999; 5: 78-82, Article MEDLINE

13 Montgomery AM, Reisfeld RA, Cheresh DA. Integrin alpha v beta 3 rescues melanoma cells from apoptosis in three-dimensional dermal collagen. Proc Natl Acad Sci USA 1994; 91: 8856-8860, MEDLINE

14 Varner JA, Cheresh DA. Integrins and cancer. Curr Opin Cell Biol 1996; 8: 724-730, MEDLINE

15 Pasqualini R, Koivunen E, Ruoslahti E. Alpha v integrins as receptors for tumor targeting by circulating ligands. Nat Biotechnol 1997; 15: 542-546, MEDLINE

16 Nip J et al. Human melanoma cells derived from lymphatic metastases use integrin alpha v beta 3 to adhere to lymph node vitronectin. J Clin Invest 1992; 90: 1406-1413, MEDLINE

17 Mason MD, Allman R, Quibell M. Adhesion molecules in melanoma - more than just superglue? J Roy Soc Med 1996; 89: 393-395,

18 Ruoslahti E, Obrink B. Common principles in cell adhesion. Exp Cell Res 1996; 227: 1-11, MEDLINE

19 Ruoslahti E. RGD and other recognition sequences for integrins. Annu Rev Cell Dev Biol 1996; 12: 697-715, MEDLINE

20 Koivunen E, Wang B, Ruoslahti E. Phage libraries displaying cyclic peptides with different ring sizes: ligand specificities of the RGD-directed integrins. Biotechnology 1995; 13: 265-270, MEDLINE

21 Spitzer AL, Maxwell F, Corsini J, Maxwell IH. Species specificity for transduction of cultured cells by a recombinant LuIII rodent parvovirus genome encapsidated by canine parvovirus or feline panleukopenia virus. J Gen Virol 1996; 77: 1787-1792, MEDLINE

22 Maxwell IH et al. Recombinant LuIII autonomous parvovirus as a transient transducing vector for human cells. Hum Gene Ther 1993; 4: 441-445, MEDLINE

23 Spitzer AL, Parrish CR, Maxwell IH. Tropic determinant for canine parvovirus and feline panleukopenia virus functions through the capsid protein VP2. J Gen Virol 1997; 78: 925-928, MEDLINE

24 O'Neil KT et al. Identification of novel peptide antagonists for GPIIb/IIIa from a conformationally constrained phage peptide library. Proteins 1992; 14: 509-515, MEDLINE

25 Hart SL et al. Cell binding and internalization by filamentous phage displaying a cyclic Arg- Gly-Asp-containing peptide. J Biol Chem 1994; 269: 12468-12474, MEDLINE

26 Hazelton BJ et al. Characterization of cell lines derived from xenografts of childhood rhabdomyosarcoma. Cancer Res 1987; 47: 4501-4507, MEDLINE

27 Wickham TJ et al. Targeted adenovirus-mediated gene delivery to T-cells via CD3. J Virol 1997; 71: 7663-7669, MEDLINE

28 Carlson J et al. Cloning and sequence of DNA encoding structural proteins of the autonomous parvovirus feline panleukopenia virus. J Virol 1985; 55: 574-582, MEDLINE

29 Maxwell IH, Spitzer AL, Maxwell F, Pintel DJ. The capsid determinant of fibrotropism for the MVMp strain of minute virus of mice functions via VP2 and not VP1. J Virol 1995; 69: 5829-5832, MEDLINE

30 Dupont F et al. Use of an autonomous parvovirus vector for selective transfer of a foreign gene into transformed human cells of different tissue origins and its expression therein. J Virol 1994; 68: 1397-1406, MEDLINE

31 Wickham TJ et al. Increased in vitro and in vivo gene transfer by adenovirus vectors containing chimeric fiber proteins. J Virol 1997; 71: 8221-8229, MEDLINE

32 Dmitriev I et al. An adenovirus vector with genetically modified fibers demonstrates expanded tropism via utilization of a coxsackievirus and adenovirus receptor-independent cell entry mechanism. J Virol 1998; 72: 9706-9713, MEDLINE

33 Bartlett JS, Kleinschmidt J, Boucher RC, Samulski RJ. Targeted adeno-associated virus vector transduction of nonpermissive cells mediated by a bispecific F(ab'gamma)2 antibody. Nat Biotechnol 1999; 17: 181-186, Article MEDLINE

34 Yang Q et al. Development of novel cell surface CD34-targeted recombinant adenoassociated virus vectors for gene therapy. Hum Gene Ther 1998; 9: 1929-1937, MEDLINE

35 Girod A et al. Genetic capsid modifications allow efficient re-targeting of adeno-associated virus type 2. Nat Med 1999; 5: 1052-1056, Article MEDLINE

36 Wu P et al. Mutational analysis of the adeno-associated virus type 2 (AAV2) capsid gene and construction of AAV2 vectors with altered tropism. J Virol 2000; 74: 8635-8647, MEDLINE

37 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, MEDLINE

38 Qing K et al. Human fibroblast growth factor receptor 1 is a co-receptor for infection by adeno-associated virus type 2. Nat Med 1999; 5: 71-77, Article MEDLINE

39 Martyn JC, Davidson BE, Studdert MJ. Nucleotide sequence of feline panleukopenia virus: comparison with canine parvovirus identifies host-specific differences. J Gen Virol 1990; 71: 2747-2753, MEDLINE

40 Corsini J, Carlson JO, Maxwell F, Maxwell IH. Symmetric-strand packaging of recombinant parvovirus LuIII genomes that retain only the terminal regions. J Virol 1995; 69: 2692-2696, MEDLINE

41 Lathe R, Kieny MP, Skory S, Lecocq JP. Linker-tailing: unphosphorylated linker oligonucleotides for joining DNA termini. DNA 1984; 3: 173-182, MEDLINE

42 Maxwell IH, Maxwell F, Schaack J. An adenovirus type 5 mutant with the preterminal protein gene deleted efficiently provides helper functions for the production of recombinant adeno-associated virus. J Virol 1998; 72: 8371-8373, MEDLINE

43 Sanburn N, Cornetta K. Rapid titer determination using quantitative real-time PCR. Gene Therapy 1999; 6: 1340-1345, MEDLINE

Figure 1 Transient cotransfection system for production of recombinant parvovirus. The recombinant LuIII genome, GLuP38LUC2, containing the luciferase reporter in place of most of the capsid coding region, is present within a pUC plasmid.²¹ An oligonucleotide encoding the displayed peptide is inserted in the SpeI site in the sequence encoding VP1 and VP2 capsid proteins of FPV in the plasmid pF-VP.²¹ The supercoiled plasmids are cotransfected by electroporation into producer cells and recombinant virions are harvested after 3 days (see Materials and methods).

Figure 2 RGD peptides that were inserted into the FPV capsid proteins at the extremity of loop 2. The upper line shows the wild type coding sequence flanking VP2 residues 226 and 227. Insertions were made between these residues by ligation of synthetic oligonucleotides into the SpeI site (see Materials and methods). The sequences of the integrin-binding C2-RGD and C4-RGD motifs are underlined within the box. The flanking amino acid residues shown were included with the intention of providing flexible linkers.

Figure 3 Comparison of yields of recombinant virions containing the wild type (WT) or C2-RGD or C4-RGD modified FPV capsid by PCR assay of packaged DNA. Amounts of packaged GLuP38LUC2 DNA in DNase I-treated virions were determined by quantitative 'real-time' PCR, using luciferase primers and Taqman probe, by reference to linearized plasmid DNA standards (see Materials and methods). The bars show mean ± S.D. of six measurements (triplicate preparations of each virus, assayed in duplicate).

Figure 4 Transducing activity in feline cells (CFK line) of recombinant virus containing the wild-type (WT), C2-RGD or C4-RGD modified FPV capsid. (a) CFK cells were infected with the viruses indicated, containing the GLuP38LUC2 genome, and luciferase assays were performed after 2 days (see Materials and methods). The bars show mean ± S.D. of nine measurements (triplicate infections with samples (0.1 ml) of triplicate preparations of each virus). The ordinate baseline represents approximately the luminometer background (also equal to values observed with control extracts of non-transduced cells). (b) CFK cells were infected with viruses having the wild-type or C4-RGD modified FPV capsid, as indicated, using equivalent multiplicity of infection, estimated from the amounts of packaged GLuP38LUC2 DNA (Figure 3). Luciferase assays were performed after 2 days; values shown are mean ± S.D. of triplicate infections.

Figure 5 Transduction of the human rhabdomyosarcoma cell line, Rh18A, by recombinant virus with the C4-RGD modified FPV capsid, and specific competition by the corresponding peptide. (a) Rh18A or CFK cells were infected with the viruses indicated, containing the GLuP38LUC2 genome, and luciferase assays were performed after 2 days. Note that CFK cells were efficiently transduced by virus with either wild-type or C4-RGD modified FPV capsid, whereas Rh18A cells displayed activity only with the latter. The bars show mean ± S.D. of seven to nine measurements (replicate infections with samples (0.1 ml) of replicate preparations of each virus). The values for CFK transduction are from the same experiment as in Figure 4. The value for activity of virus with the wild-type capsid in Rh18A cells did not differ significantly from the luminometer background. Virus with the C2-RGD capsid gave very low activity in Rh18A cells (up to 10 times the luminometer background; data not shown). (b) Competition by peptide ACDCRGDCFCG. Cells were infected with virions containing the GLuP4LUC1 genome, packaged either with the C4-RGD modified FPV capsid or, as a control, with the LuIII capsid, in the presence of the indicated concentrations of competing peptide (see Materials and methods). Luciferase assays were performed after 2 days. Results shown are mean ± S.D. of triplicate infections, expressed relative to activity measured without peptide. The data shown are from a representative experiment; in additional experiments we have observed similar inhibition by the peptide of transduction by virus containing the GLuP38LUC2 genome in the C4-RGD modified FPV capsid, with essentially complete inhibition by higher concentrations of peptide (in the range 3-10 M) (results not shown).

Figure 6 Specific inhibition by the v integrin-blocking antibody, L230, of transduction of Rh18A cells by recombinant virus with the C4-RGD modified FPV capsid. Cells were infected with virions containing the GLuP4LUC1 genome, packaged either with the C4-RGD modified FPV capsid or with control LuIII capsid, in the presence of the indicated amounts of hybridoma supernatant containing L230 antibody (see Materials and methods). Luciferase assays were performed after 2 days. Results shown are mean ± S.D. of triplicate infections, expressed relative to activity measured without antibody.

Figure 7 Insertion of C4-RGD into either the major (VP2) or minor (VP1) FPV capsid protein allows transduction of Rh18A cells with comparable efficiency. Rh18A cells were infected with recombinant virus containing the indicated combinations of wild-type (WT) or C4-RGD modified VP1 and VP2. The viruses were generated by transfection of producer cells with pGLuP38LUC2, together with pF-VP1, pF-VP2²³ or their derivatives encoding the C4-RGD modified proteins (see Materials and methods). Luciferase assays were performed after 2 days. The activities are shown relative to the activity seen with C4-RGD present in both VP1 and VP2. The bars show mean ± S.D. of four measurements (duplicate infections with samples of duplicate preparations of each virus).

Table 1 Transduction activity of C4-RGD virus in various human tumor cell lines