WO2012047109A2

WO2012047109A2 - Mirna and mirna inhibitors that modulate adenovirus replication and/or adenovirus induced cell death

Info

Publication number: WO2012047109A2
Application number: PCT/NL2011/050688
Authority: WO
Inventors: Victor Willem Van Beusechem; Jan Willem Hendrik Van Ginkel
Original assignee: Orca Therapeutics B.V.
Priority date: 2010-10-08
Filing date: 2011-10-07
Publication date: 2012-04-12
Also published as: WO2012047109A3

Abstract

The invention relates to means and methods for identifying a mi RNA, a mi RNA mimic or a mi RNA inhibitor that modulates adenovirus replication in, and/or adenovirus induced cell death of a cell. Identified mi RNAs, mi RNA mimics or mi RNA inhibitors are used as anti-cancer agents in the context of adenovirus gene therapy.

Description

Title: MiRNA and MiRNA inhibitors that modulate adenovirus replication and/or adenovirus induced cell death.

The invention is related to the fields of genetic modification, biotechnology and medicine. In particular, the invention relates to microRNA (miRNA), miRNA mimics and miRNA inhibitors that modulate adenovirus replication and/or cell death of an adenovirus infected cell, preferably adenovirus induced cell death. The invention further relates to the use of miRNA, miRNA mimics and miRNA inhibitors to obtain more efficient replication of adenovirus or to obtain more efficient cell death of adenovirus infected cells, preferably adenovirus induced cell death in infected cells, particularly cancer cells.

Viruses are naturally evolved vehicles which efficiently transfer their genes into host cells. This ability makes them desirable for engineering virus vector systems for the delivery of genes to specific target cells. The viral vectors that are presently used in laboratory and clinical studies are based on RNA and DNA viruses possessing different genomic structures and host ranges. Particular viruses have been selected as vehicles because of their capacity to carry foreign genes and their ability to efficiently deliver these genes. At present, viral vectors derived from retroviruses, adenovirus, adeno- associated virus, herpesvirus and poxvirus are employed in more than 70% of clinical gene therapy trials worldwide. Of these viruses the adenoviruses have the advantage that they are easy to use, efficiently produce progeny virus, are able to very efficiently infect target cells and do not integrate into the host cell genome thereby avoiding any risk of insertional mutagenesis.

Adenoviruses are double-stranded DNA viruses whose major capsid components are hexon, penton and fiber. Adenoviral infection is mediated by binding of the knob region, located at the carboxy terminus of the fiber, to its corresponding receptor, which is the coxsackie-adenovirus receptor (CAR) for most serotypes. Binding is followed by interaction between cellular integrins and an arginine-glycine-aspartic acid motif (RGD-motif) located at the penton base. This binding leads to formation of endosomes and viral internalization. Subsequently, the adenoviral DNA is transported to the nucleus and

adenoviral protein synthesis, or in case of non-replicating adenoviruses (Ads), transgene expression begins. Adenoviral DNA is not integrated into the host genome, thereby resulting in a low risk of mutagenesis. The limited duration of gene expression is adequate for cancer gene therapy approaches, where the purpose typically is to kill the target cells. Infection is not dependent on cell cycle phase; therefore, both cycling and non-dividing cells are infected. A feature of Ad for therapy is its unparalleled capacity for gene transfer and expression in vivo.

Adenovirus replication involves roughly the following steps: (1) infection of the host cell by binding of the adenovirus particle to the cell surface, internalization and transport towards the cell nucleus, and import of the adenovirus DNA genome into the cell nucleus; (2) expression of adenovirus proteins encoded by the early regions in the adenovirus genome; (3) replication of the adenovirus genome, which marks the transition of the early replication phase to the late replication phase; (4) expression of adenovirus proteins encoded by the late regions in the adenovirus genome; (5) assembly of progeny adenovirus particles and inclusion of progeny adenovirus genomes into these particles; and (6) induction of cell death and release of adenovirus progeny from the cell.

During their life cycle, adenoviruses modulate cell death pathways. In different cell lines, p53 dependent as well as p53 independent apoptosis has been documented after adenovirus infection (Teodoro and Branton, J. Virol. 71(1997):1739-1746; and references therein). During the early replication phase, cell death is suppressed to prevent premature cell death, thereby allowing the adenovirus to complete its life cycle in the cell. In contrast, at late stages of infection cell death and lysis are promoted to enhance release of virus progeny from the cell. Replication competent viruses, in particular adenoviruses, are finding increasing utility for the treatment of cancer and other diseases involving inappropriate cell survival. In particular, conditionally replicating adenoviruses (CRAds) have been developed to selectively replicate in and kill cancer cells. Such cancer- specific CRAds represent a novel and very promising class of anticancer agents (reviewed by Heise and Kirn, J. Clin. Invest.

105(2000):847-851; Alemany et al., Nat. Biotech. 18(2000):723-727; Gomez- Navarro and Curiel, Lancet Oncol. 1(2000):148-158). The tumor-selective replication of this type of CRAds is achieved through, as yet, two alternative strategies. In the first strategy, the expression of an essential early adenovirus gene is controlled by a tumor- specific promoter (e.g., Rodriguez et al., Cancer Res. 57(1997):2559-2563; Hallenbeck et al., Hum. Gene Ther. 10(1999):1721- 1733; Tsukuda et al., Cancer Res. 62(2002):3438-3447; Huang et al., Gene Ther. 10(2003):1241-1247; Cuevas et al., Cancer Res. 63(2003):6877-6884). The second strategy involves the introduction of mutations in viral genes to abrogate the interaction of the encoded RNA or protein products with cellular proteins necessary to complete the viral life cycle in normal cells, but not in tumor cells (e.g., Bischoff et al., Science 274(1996):373-376; Fueyo et al., Oncogene 19(2000):2-12; Heise et al., Clin. Cancer Res. 6(2000):4908-4914; Shen et al., J. Virol. 75(2001:4297-4307); Cascallo et al., Cancer Res.

63(2003):5544-5550).

During their replication in tumor cells, CRAds destroy cancer cells, a process that is further referred to as "oncolysis". The release of viral progeny from infected cancer cells offers the potential to amplify CRAds in situ and to achieve lateral spread to neighbouring cells in a solid tumor, thus expanding the oncolytic effect. The restriction of CRAd replication to cancer or

hyperproliferative cells dictates the safety of the agent, by preventing lysis of normal tissue cells. Currently, CRAd-based cancer treatments are being evaluated in clinical trials (e.g., Nemunaitis et al., Cancer Res. 60(2000):6359- 6366; Khuri et al., Nature Med. 6(2000):879-885; Habib et al., Hum. Gene Ther. 12(2001):219-226).

However, despite very encouraging results from in vitro and animal studies, the anti-cancer efficacy of replicative adenovirus as a therapeutic agent in humans has been limited. The present invention provides means and methods to increase the potency of adenoviruses as oncolytic agents. This is achieved by enhancing their replication and/or cell death inducing capacities. In the present invention it has been found that miRNA, miRNA mimics and miRNA inhibitors can modulate adenovirus replication in and/or cell death of an adenovirus infected cell, preferably adenovirus induced cell death of a cell that is permissive for adenovirus replication.

The invention provides a method for determining whether a miRNA, a miRNA mimic or miRNA inhibitor modulates, preferably increases adenovirus replication in, and/or preferably cell death of an adenovirus infected cell, preferably adenovirus induced cell death of a cell, said method comprising providing a first cell that is preferably permissive for adenovirus replication, with an adenovirus and with a nucleic acid molecule comprising and/or encoding said miRNA, a miRNA mimic or miRNA inhibitor, said method further comprising culturing said first cell and determining whether replication of said adenovirus in said first cell and/or whether cell death, preferably adenovirus induced cell death of said first cell, is affected and preferably is increased. In one embodiment said miRNA, miRNA mimic or miRNA inhibitor increases replication in said first cell and increases cell death of said adenovirus infected first cell, preferably adenovirus induced cell death of said first cell. Said first cell is preferably permissive for infection by the adenovirus. In this way the cell can easily be provided with the adenovirus. For determining whether a miRNA, a miRNA mimic or miRNA inhibitor modulates adenovirus replication in said first cell, it is preferred that said first cell is a cell that is permissive for adenovirus replication. For determining cell death modulators it is preferred that said cell is permissive for adenovirus induced cell death. Preferably, modulators are detected that increase adenovirus replication in said first cell, or that increase cell death, preferably adenovirus induced cell death in said first cell. An advantage of such a method of the invention is that it can identify one or more miRNAs, miRNA mimics or miRNA inhibitors that, when present in the same cell as said adenovirus, increase replication of the adenovirus in said cell and/or that increase cell death of an adenovirus infected cell. The method can also identify one or more miRNAs, miRNA mimics or miRNA inhibitors that, when present in the same cell as said adenovirus, increase cell death induced by said adenovirus in said cell. When in the present invention there is a preference for increased adenovirus induced cell death, it is preferred that said adenovirus induced cell death is selectively increased.

MiRNAs are short non-coding RNAs and are among the most abundant class of small RNAs in animals. Other members of the class of small RNAs include small-interfering RNAs (siRNAs) and Piwi-interacting RNAs (piRNAs). While miRNAs and piRNAs are endogenously encoded by the cell, siRNAs originate from an extracellular source, i.e. viral RNAs. MiRNAs, siRNAs, and piRNAs are all regulators of post-transcriptional gene expression, and although their mature end-products are structurally almost identical, they are functionally quite different. piRNAs are complexed with PIWI proteins and are almost only active in the germline, mainly regulating gene expression during spermatogenesis. Although originating from a different source, the processing of miRNAs and siRNAs converge into the same molecular machinery within the cytosol, ultimately leading to the

incorporation of the mature processing product into the RNA Induced

Silencing Complex (RISC). While siRNAs regulate gene expression by inducing degradation of the target mRNA, miRNAs can also silence gene expression through translational repression. The differences between the various types of small RNAs are well-known in the art and are reviewed in (Chu and Rana, J. Cell. Physiol. 213: 412-419, 2007). A more detailed description of miRNA processing and function is given below.

Typically miRNAs comprise between 18 and 27 nucleotides, preferably between 20 to 23 nucleotides. MiRNAs often, though not always, have a uridine at their 5'-end. When reference is made herein to a range of nucleotides this range includes the numbers listed. For instance, a preferred size range for miRNA is between 20 to 23 nucleotides, this range includes the number 20, 21, 22 and 23. The principles of miRNAs and their post- transcriptional gene regulation are reviewed in e.g. Zeng, Oncogene 25, 6156- 6162, 2006; and Engels and Hutvagner, Oncogene 25, 6163-6169, 2006. The Sanger miRBase database (www.mirbase.org) is a searchable database of published miRNA sequences and annotation. Each entry in the miRBase Sequence database represents a predicted hairpin portion of a miRNA transcript, with information on the location and sequence of the mature miRNA.

MiRNAs are wide spread in living species. They are expressed in plants, in-vertebrates and in vertebrates. Even some viruses are reported to express miRNA. The numbers that are expressed in the different specifies vary with the different methods for estimating their number. Humans are reported to have over 1000 expressed miRNA. MiRNAs are typically differentially expressed in different cell types or tissues. Expression can be high, leading up to several thousands of copies per cell. Although many miRNAs have been identified sofar, it is often not completely known what targets are influenced by an individual miRNA. MiRNAs post-transcriptionally inhibit expression of certain target genes, however the mechanism by which this is achieved varies. In plants, most miRNA bind target mRNA sequences perfectly and seem to lead to mRNA degradation. In animals, both mRNA degradation and prevention of translation without mRNA degradation are observed. Currently, only a minor fraction of all known miRNAs have an experimentally determined function in vivo. These miRNAs include lin-4 and let-7 in Caenorhabditis elegans, bantam and miR-14 in Drosophila, miR-12b in zebrafish, and miR-23 in humans, playing vital roles in development and apoptosis. Moreover, the important role of miRNAs in oncogenesis, either acting as tumor suppressors or oncogenes is becoming more evident over the past few years (reviewed in Garzon et al. Annu Rev Med, 2009 60: 167-179).

It has been observed that toward the 5' end of the miRNA there is a perfect Watson-Crick base pair matching of at least seven consecutive nucleotides, this sequence is often referred to as the seed sequence. A miRNA often, but not necessarily always, binds to the 3' untranslated region (UTR) of a target mRNA. Recent experimental evidence has added more insights into 3' (UTR) -binding, but a complete understanding of miRNA-target interactions is not known. However, it is known that any given miRNA can have several target mRNAs.

In the cell, a miRNA is produced from a longer RNA transcript referred to as a pri-miRNA (primary miRNA). A single pri-miRNA may contain one or more miRNA precursors. The miRNA precursors (pre-miRNA) are hairpin loop structures composed of about 50-110 nucleotides each. A typical hairpin loop contains about 70 nucleotides. Each hairpin is flanked by sequences that facilitate efficient processing. The hairpin loop structures (pre- miRNAs) are liberated from pri-miRNAs in the nucleus by an enzyme complex comprising Drosha. The resulting pre-miRNA is transported to the cytoplasm where it associates with another enzyme complex that contains the enzyme Dicer. This Dicer complex releases the miRNA duplex from the pre-miRNA. The functional strand (or guide strand) of the mature miRNA is loaded together with Argonaute proteins into the RNA Induced Silencing Complex

(RISC), where it guides RISC to silence target mRNAs. Providing the cell with an expression cassette that codes for the mentioned nucleic acid(s) can also provide the nucleic acids to the cell. The disclosure therefore provides nucleic acid molecules, such as pri-miRNA and pre-miRNA, which encode miRNAs, miRNA mimics or miRNA inhibitors useful in the invention. The miRNAs follow a standard nomenclature system, as is known to the skilled person. An uncapitalized "mir-" refers to a pre-miRNA, while a capitalized "miR-" refers to a mature form. miRNAs with nearly identical sequences are annotated with an additional lower case letter. For example, miR- 123a is closely related to miR- 123b. miRNAs that are 100% identical but are encoded at different places in the genome are indicated with additional dash-number suffix: miR-123-1 and miR-123-2 are identical but are produced from different pre-miRNAs. Species of origin is designated with a three-letter prefix, e.g., hsa-miR-123 would be from human (Homo sapiens) and oar-miR- 123 would be a sheep (Ovis aries) miRNA. When relative expression levels are known, an asterisk following the name indicates an miRNA expressed at low levels relative to the miRNA in the opposite arm of a hairpin. For example, miR-123 and miR-123* would share a pre-miRNA hairpin, but relatively more miR-123 would be found in the cell. The suffices 5p and 3p indicate if a miRNA is derived from the 5'arm or the 3'arm of the pre-miRNA respectively. These suffices are used when both miRNAs are expressed at equal levels. For example miR423-5p and miR423-3p would share a pre-miRNA hairpin. The nomenclature of the miRNA should not be given absolute weight as it is still subject to change. The sequence for preferred miRNAs, miRNA mimics and/or miRNA inhibitors is given in figure 7. According to the present invention, miRNAs as mentioned herein are capable of counteracting expression of specific gene products. As used herein, the term "miRNA" encompasses any isoform of the said miRNA and all members of the said miRNA family are capable of counteracting expression of the specific gene products. The term "miRNA" thus includes star-sequences and family members.

Cells can be provided with any miRNA by providing them with the miRNA duplex as it is released from the pre-miRNA, by providing them with a miRNA mimic, by providing them with a precursor molecule from which the miRNA duplex is released by Dicer. The miRNA precursor molecule can be the pri-miRNA, the pre-miRNA or a mimic thereof. The miRNA precursor can also be a precursor RNA that encodes a pri-miRNA comprising said miRNA stem- loop sequence flanked on each side by at least 30 single- stranded RNA nucleotides at the 5' and 3' sides of the miRNA stem-loop sequence. MiRNA mimics are typically functionally equivalent to the miRNA duplex that they mimic. It has been found that modification of the sugar backbone can be used to alter stability, hybridization, transport and other properties of the miRNA. For instance, LNA (locked nucleic acid) modifications of the miRNA backbone have been shown to increase the efficiency of silencing of the target mRNA. Similarly, changing the bases sequence by for instance changing an adenosine to an inosine, broadens the target specificity. Of some miRNA duplexes, both strands can be incorporated into RISC, providing two different mature miRNAs. In order to design miRNA mimics with the targeting repertoire of only one of the two alternative mature miRNAs, the sequence of the other strand can be modified such that it no longer acts as the alternative miRNA product. A miRNA precursor mimic provides a hairpin structure resembling a pre-miRNA hairpin structure as it occurs in nature, so that it serves as a template for the cellular pri-/pre-miRNA processing machinery to allow release of the miRNA duplex in the cell. The above nucleic acids can be provided to the cell as such. Providing the cell with an expression cassette that codes for the mentioned nucleic acid(s) can also provide the nucleic acids to the cell. The disclosure therefore provides nucleic acid molecules, such as pri-miRNA and pre-miRNA, which encode miRNAs, miRNA mimics or miRNA inhibitors useful in the invention.

A miRNA inhibitor is typically a nucleic acid molecule comprising and/or encoding an oligonucleotide with the reverse complement sequence of the miRNA it inhibits. The inhibitor hybridizes to the miRNA thereby negating its activity. The miRNA inhibitor typically comprises one or more modifications to enhance the hybridization of the inhibitor to the target miRNA. To this end often LNA modified nucleic acid is used. A preferred miRNA inhibitor of the invention is an antisense oligonucleotide comprising a sequence that is the reverse complement of the miRNA sequence it inhibits. Preferably said inhibitor further comprises a hairpin sequence. Preferably a hairpin sequence as depicted with the inhibitor sequences depicted in figure 7. In a preferred embodiment a miRNA inhibitor of the invention comprises the reverse complement of a sequence as depicted in figure 8. A miRNA, miRNA mimic or nucleic acid molecule precursor thereof preferably comprises an RNA backbone, preferably with one or more modification such as a 2-0'-methyl modification, a locked nucleic acid modification, a morpholino modification and/or a peptide nucleic acid (PNA) modification. The modifications can be at one or more sugar moieties of the backbone. The peptide nucleic acid modifications are typically suitable for the miRNA mimic or the miRNA inhibitor. Next to these chemical modifications, it has been shown that adding flanking sequences to the antisense-based miRNA inhibitors strongly enhances the inhibitory potency of these molecules. Preferably, these flanking sequences form stem-loop hairpin structures with a total of between 8 and 16 bases on both 5' and 3' end of the antisense oligomer. Most preferred are flanking sequences of 12 bases, with a 4 base pair (2x4 bases) stem region and a loop region of 4 bases. These added structural elements are thought to enhance and stabilize the interaction between the miRNA-RISC complex and the inhibitor, thus prolonging the effect of the inhibitor (Vermeulen et al. RNA (2007), 13: 723-730).

MiRNA mediated gene suppression is thought to occur by at least three different mechanisms involving the block of protein translation, inhibition of translation initiation, and destabilization of mRNA (reviewed in Chu and Rana, J. Cell. Physiol. 213: 412-419, 2007). The target mRNA contains a recognition site for the miRNA. This recognition site comprises a region that is the reverse complement of a part of the miRNA. The reverse complement in the target mRNA is typically, but not necessarily, 100% identical to the seed sequence in the miRNA. The adjacent part often contains one or more mismatches with the miRNA, leading to non-perfect fit of the hybrid. This non-perfect fit and the fact that mRNA secondary structure at a potential target site affects the binding of the miRNA makes it more complicated to decipher the target mRNAs for miRNA than for siRNA.

Algorithms for predicting miRNA targets are available and include e.g.

TargetScan (www.targetscan.org), miRanda (www.microrna.org), PicTar (pictar.mdc-berlin.de) and miRWalk (www.ma.uni- heidelberg.de/apps/zmf/mirwalk/index.html). The latter is a comprehensive database providing information on predicted as well as validated miRNA binding sites on their targets, based on the miRWalk algorithm and target prediction by eight established miRNA prediction programs, including

TargetScan, miRanda and PicTar.

Modulating the level of a miRNA in a cell is thus fundamentally different from, for instance, providing a cell with an siRNA molecule. An siRNA molecule generally requires perfect base pairing with its target mRNA in order to effect cleavage from the RISC complex. In this latter case, the target of the siRNA is identified. This allows directed manipulation of the level of a predetermined target mRNA in a cell, something that is not possible with the vast majority of miRNA for which the target mRNAs still remain to be identified. As miRNAs modulate the expression of a collection of genes, the effect of such modulation is often pleiotropic.

The present invention provides miRNAs, miRNA mimics and miRNA inhibitors that increase adenovirus replication in, and/or adenovirus induced cell death of a cell infected with said adenovirus. The cell is preferably an immortalized cell or more preferably a cancer cell. The miRNAs, miRNA mimics and miRNA inhibitors are useful in the methods and kits of parts described herein. The miRNAs and miRNA mimics of the invention and the miRNAs that are inhibited by the miRNA inhibitors of the invention

preferably block translation post-initiation of target mRNA, inhibit translation initiation of target mRNA, and/or destabilize target mRNA. As previously discussed, miRNAs differ from siRNAs in that perfect complementarity to the target sequence is not required for gene repression. miRNAs usually form bulge structures due to imperfect matching with the target sequence and the target specificity is determined by the seven nucleic acid seed sequence (Doensch and Sharp, Genes Dev 18:501-511 (2004)).

MiRNAs, miRNA mimics or miRNA inhibitors useful in the present invention can therefore be modified without negatively affecting their function. MiRNAs, miRNA mimics or miRNA inhibitors useful in the present invention can also be developed by optimizing miRNA molecules. In an exemplary embodiment, miRNAs, miRNA mimics, or miRNA inhibitors that have been identified to have a desired effect can be optimized to, for example, improve binding to target mRNAs or alter the repertoire of target mRNA. Optimization can be performed by, e.g., randomly mutagenizing the miRNA sequence or by selectively introducing nucleotide substitutions, deletions, or additions. In some embodiments, said optimization alters the seed sequence of the miRNA, miRNA mimic, or miRNA inhibitor. Example 5 of the present disclosure demonstrates that the deletion of 2 nucleotides from miR-324-3p shifts the seed sequence and results in an improved effect. The disclosure therefore provides miRNAs, miRNA mimics or miRNA inhibitors having a shifted seed sequence in comparison to an endogenous miRNA molecule. Molecules having nucleic acid sequences which have been modified from endogenous miRNA are for the purposes of this invention still referred to as miRNAs, miRNA mimics or miRNA inhibitors. These modified molecules share the same mechanism of gene regulation as endogenous miRNA molecules.

A miRNA, miRNA mimic or miRNA inhibitor of the invention is preferably a) a miRNA, miRNA mimic or mRNA inhibitor of category III, category lib, category, IV, category I, category Ila and/or category V of table 1 or a miRNA inhibitor that comprises a sequence that is the reverse

complement of a sequence according to figure 8 or b) a miRNA, miRNA mimic or miRNA inhibitor having 4 or less, 3 or less, or, preferably, 2 or less nucleotide additions, substitutions or deletions in respect to the miRNAs, miRNA mimics or miRNA inhibitors of category III, category lib, category IV, category I, category Ila and/or category V of table 1 or a miRNA inhibitor comprising the reverse complement of a miRNA sequence as depicted in figure 8 having 4 or less, 3 or less, or, preferably, 2 or less nucleotide additions, substitutions or deletions, and/or the miRNAs, miRNA mimics or miRNA inhibitors have at least 7, preferably at least 20, consecutive nucleotides which are identical to the miRNAs, miRNA mimics or the miRNA complementary part of the miRNA inhibitors of category III, category lib, category IV, category I, category Ila and/or category V of table 1. Preferably, said miRNA, miRNA mimic or miRNA inhibitor of the invention is a) a miRNA, miRNA mimic or mRNA inhibitor of table 1 or a miRNA inhibitor that comprises a sequence that is the reverse complement of a sequence according to figure 8 or b) a miRNA, miRNA mimic or miRNA inhibitor having 4 or less, 3 or less, or, preferably, 2 or less nucleotide additions, substitutions or deletions in respect to the miRNAs, miRNA mimics or miRNA inhibitors of table 1 or a miRNA inhibitor comprising the reverse complement of a miRNA sequence as depicted in figure 8 having 4 or less, 3 or less, or, preferably, 2 or less nucleotide additions, substitutions or deletions, and/or the miRNAs, miRNA mimics or miRNA inhibitors have at least 7, preferably at least 20, consecutive nucleotides which are identical to the miRNAs, miRNA mimics or, the miRNA complementary part of the miRNA inhibitors of table 1. Preferably said miRNAs, miRNA mimics or miRNA inhibitors have a seed sequence shifted by 4 or less, 3 or less, or, preferably, 2 or less nucleotide in comparison with the miRNAs, miRNA mimics or miRNA inhibitors of table 1. Preferably said miRNAs, miRNA mimics or miRNA inhibitors have a seed sequence shifted by no more than 1 nucleotide in comparison with the miRNAs, miRNA mimics or miRNA inhibitors of table 1. More preferably, said miRNAs, miRNA mimics or miRNA inhibitors share an identical seed sequence with the miRNAs, miRNA mimics or miRNA inhibitors of table 1. Preferably, said miRNA, miRNA mimic or miRNA inhibitor of the invention is a) a miRNA, miRNA mimic or mRNA inhibitor of category II-IV of table 1 or a miRNA inhibitor of category II-IV that comprises a sequence that is the reverse complement of a sequence according to figure 8 or b) a miRNA, miRNA mimic or miRNA inhibitor having 4 or less, 3 or less, or, preferably, 2 or less nucleotide additions, substitutions or deletions in respect to the miRNAs, miRNA mimics or miRNA inhibitors of category II-IV of table 1 or a miRNA inhibitor of category II-IV of table 1 comprising the reverse

complement of a miRNA sequence as depicted in figure 8 having 4 or less, 3 or less, or, preferably, 2 or less nucleotide additions, substitutions or deletions, and/or the miRNAs, miRNA mimics or miRNA inhibitors have at least 7, preferably at least 20, consecutive nucleotides which are identical to the miRNAs, miRNA mimics or miRNA inhibitors of category II-IV of table 1. Preferably said miRNAs, miRNA mimics or miRNA inhibitors have a seed sequence shifted by 4 or less, 3 or less, or, preferably, 2 or less nucleotide in comparison with the miRNAs, miRNA mimics or miRNA inhibitors of category II-IV of table 1. Preferably said miRNAs, miRNA mimics or miRNA inhibitors have a seed sequence shifted by no more than 1 nucleotide in comparison with the miRNAs, miRNA mimics or miRNA inhibitors of category II-IV of table 1. More preferably, said miRNAs, miRNA mimics or miRNA inhibitors share an identical seed sequence with the miRNAs, miRNA mimics or miRNA inhibitors of category II-IV of table 1.

Preferably, said miRNA or miRNA mimic of the invention is a) a miRNA or miRNA mimic of category II-III of table 1 or a miRNA inhibitor of category II-III that comprises a sequence that is the reverse complement of a sequence according to figure 8 or b) a miRNA or miRNA mimic having 4 or less, 3 or less, or, preferably, 2 or less nucleotide additions, substitutions or deletions in respect to the miRNA or miRNA mimic of category II-III of table 1 or a miRNA inhibitor of category II-III of table 1 comprising the reverse complement of a miRNA sequence as depicted in figure 8 having 4 or less, 3 or less, or, preferably, 2 or less nucleotide additions, substitutions or deletions and/or the miRNA or miRNA mimic have at least 7, preferably at least 20, consecutive nucleotides which are identical to the miRNAs or miRNA mimics of category II-III of table 1. Preferably said miRNAs or miRNA mimics have a seed sequence shifted by 4 or less, 3 or less, or, preferably, 2 or less nucleotide in comparison with the miRNAs or miRNA mimics of category II-III of table 1. Preferably said miRNAs or miRNA mimics have a seed sequence shifted by no more than 1 nucleotide in comparison with the miRNAs or miRNA mimics of category II-III of table 1. More preferably, said miRNAs or miRNA mimics share an identical seed sequence with the miRNAs or miRNA mimics of category II-III of table 1.

Preferably, said miRNA or miRNA mimic of the invention is a) a miRNA or miRNA mimic of category lib or III of table 1 or b) a miRNA or miRNA mimic having 4 or less, 3 or less, or, preferably, 2 or less nucleotide additions, substitutions or deletions in respect to the miRNAs or miRNA mimics of category lib or III of table 1 and/or the miRNAs or miRNA mimics have at least 7, preferably at least 20, consecutive nucleotides which are identical to the miRNAs or miRNA mimics of category lib or III of table 1. Preferably said miRNAs or miRNA mimics have a seed sequence shifted by 4 or less, 3 or less, or, preferably, 2 or less nucleotide in comparison with the miRNAs or miRNA mimics of category lib or III of table 1. Preferably said miRNAs or miRNA mimics have a seed sequence shifted by no more than 1 nucleotide in comparison with the miRNAs or miRNA mimics of category lib or III of table 1. More preferably, miRNAs or miRNA mimics share an identical seed sequence with the miRNAs or miRNA mimics of category lib or III of table 1.

Preferably, said miRNA or miRNA mimic of the invention is a) a miRNA or miRNA mimic of category III of table 1 or b) a miRNA or miRNA mimic having 4 or less, 3 or less, or, preferably, 2 or less nucleotide additions, substitutions or deletions in respect to miRNAs or miRNA mimics of category III of table 1 and/or the miRNAs or miRNA mimics have at least 7, preferably at least 20, consecutive nucleotides which are identical to miRNAs or miRNA mimics of category III of table 1. Preferably miRNAs or miRNA mimics have a seed sequence shifted by 4 or less, 3 or less, or, preferably, 2 or less nucleotide in comparison with the miRNAs or miRNA mimics of category III of table 1. Preferably miRNAs or miRNA mimics have a seed sequence shifted by no more than 1 nucleotide in comparison with miRNAs or miRNA mimics of category III of table 1. More preferably, miRNAs or miRNA mimics share an identical seed sequence with the miRNAs or miRNA mimics of category III of table 1.

In a particularly preferred embodiment said miRNA or miRNA mimic is a) miR-383, miR-324-3p, miR-432, miR-517c, or miR-520e or b) a miRNA or mRNA mimic having 4 or less, 3 or less, or, preferably, 2 or less nucleotide additions, substitutions or deletions in respect to miR-383, miR-324-3p, miR- 432, miR-517c, or miR-520e and/or the miRNAs or miRNA mimics have at least 7, preferably at least 20, consecutive nucleotides which are identical to the miR-383, miR-324-3p, miR-432, miR-517c, or miR-520e. Preferably said miRNAs or miRNA mimics have a seed sequence shifted by 4 or less, 3 or less, or, preferably, 2 or less nucleotide in comparison with miR-383, miR-324-3p, miR-432, miR-517c, or miR-520e. Preferably said miRNAs or miRNA mimics have a seed sequence shifted by no more than 1 nucleotide in comparison with miR-383, miR-324-3p, miR-432, miR-517c, or miR-520e. More preferably, said miRNAs or miRNA mimics share an identical seed sequence with miR-383, miR-324-3p, miR-432, miR-517c, or miR-520e.

In a particularly preferred embodiment said miRNA or miRNA mimic is a) miR-383, miR-324-3p, miR-lOa, miR-26b, miR-199a-3p, miR-520e or miR- 517c, or b) a miRNA or mRNA mimic having 4 or less, 3 or less, or, preferably, 2 or less nucleotide additions, substitutions or deletions in respect to miR-383, miR-324-3p, miR-lOa, miR-26b, miR-199a-3p, miR-520e or miR-517c and/or the miRNAs or miRNA mimics have at least 7, preferably at least 20, consecutive nucleotides which are identical to the miR-383, miR-324-3p, miR- 10a, miR-26b, miR-199a-3p, miR-520e or miR-517c. Preferably said miRNAs or miRNA mimics have a seed sequence shifted by 4 or less, 3 or less, or, preferably, 2 or less nucleotide in comparison with miR-383, miR-324-3p, miR- 10a, miR-26b, miR-199a-3p, miR-520e or miR-517c. Preferably said miRNAs or miRNA mimics have a seed sequence shifted by no more than 1 nucleotide in comparison with miR-383, miR-324-3p, miR-lOa, miR-26b, miR-199a-3p, miR- 520e or miR-517c. More preferably, said miRNAs or miRNA mimics share an identical seed sequence with miR-383, miR-324-3p, miR-lOa, miR-26b, miR- 199a-3p, miR-520e or miR-517c.

In the present invention it has been found that some miRNA, miRNA mimics and/or miRNA inhibitors increase adenovirus replication in a cell that is permissive for adenovirus replication. It has also been found that some miRNA, miRNA mimics and/or miRNA inhibitors increase cell death of an adenovirus infected cell, preferably adenovirus induced cell death of a cell that is permissive for adenovirus infection, preferably said cell is also permissive for adenovirus replication. MiRNAs, miRNA mimics and/or miRNA inhibitors can be provided to the cell as such or by providing the cell with a nucleic acid molecule comprising and/or encoding said miRNA, a miRNA mimic or miRNA inhibitor. Increasing adenovirus replication is useful in, for instance, a virus producer cell. Decreasing the replication of an adenovirus can be useful for the production of adeno-associated virus (AAV) vectors. By decreasing competition of a replicating adenovirus in the same cell, AAV replication is favoured.

Increasing replication of an adenovirus can be useful for producing an adenovirus vector. Modulating cell death is another preferred embodiment of the invention. Increasing cell death is useful for instance, when adenoviruses are used to kill undesired cells, such as cancer cells. Replication of the adenovirus can be determined for example by determining adenovirus DNA content in host cells, or functional infectious adenovirus progeny produced in host cells, or expression of an endogenous gene or transgene expressed from the adenovirus genome, preferably from the adenovirus major late

transcription unit, in host cells. Cell viability or cell death is preferably determined by means of counting viable cells or intact nuclei, or by measuring metabolic activity using a method known in the art, such as e.g. CellTiter-Blue (Promega) or MTS (Promega) or WST-1 (Roche) tetrazolium-based assay, or ATP content, such as e.g. CellTiter-Glow (Promega). Alternatively, release of cellular protein, e.g. lactate dehydrogenase (CytoTox-ONE, Promega), into the culture medium can be measured, or induction of apoptosis can be determined using a method known in the art, such as e.g., a caspase activity assay like Apo-ONE (Promega).

A method of the invention preferably compares adenovirus replication of said adenovirus in said first cell and/or, preferably adenovirus induced, cell death of said first cell with a reference. Similarly, adenovirus replication of said adenovirus in said second cell and/or, preferably adenovirus induced, cell death of said second cell is preferably compared with a reference This facilitates quantitation of the adenovirus replication and or cell death, preferably adenovirus induced cell death. The reference is preferably the same test using the same type of cell and the same adenovirus but without the indicated miRNA, miRNA mimic and/or miRNA inhibitor. The reference may contain no miRNA, miRNA mimic and/or miRNA inhibitor, or, alternatively, the reference may contain an irrelevant control miRNA, miRNA mimic and/or miRNA inhibitor without expected activity in said cell. When the method of the invention is done in a high-throughput screening setting, the median result of the miRNAs, miRNA mimics and/or miRNA inhibitors included in the screen can be considered as reference, assuming that the majority of included miRNAs, miRNA mimics and/or miRNA inhibitors represent irrelevant miRNAs, miRNA mimics and/or miRNA inhibitors. In cases where the adenovirus comprises and/or encodes the miRNA, miRNA mimic and/or miRNA inhibitor, the comparison is preferably made with the same test but wherein the adenovirus is a reference adenovirus. The reference adenovirus preferably has the same structure as the test adenovirus but for the absence therein of the miRNA, miRNA mimic and/or miRNA inhibitor or a nucleic acid encoding the miRNA, miRNA mimic and/or miRNA inhibitor.

Increasing cell death is typically measured by comparing the cell death in populations of cells provided with the respective molecules with control populations of cells. Death of adenovirus infected cells is typically quantified by comparing populations of cells that have the same miRNA, a miRNA mimic or miRNA inhibitor but differ in the presence or absence of an adenovirus (populations A and B respectively). Preferably there is a third and a fourth population of the same cells but that do not have the miRNA, a miRNA mimic or miRNA inhibitor either without the adenovirus (population C) or with the adenovirus (population D).

A = + Ad + miRNA, a miRNA mimic or miRNA inhibitor

B = - Ad + miRNA, a miRNA mimic or miRNA inhibitor

C = - Ad - miRNA, a miRNA mimic or miRNA inhibitor

D = + Ad - miRNA, a miRNA mimic or miRNA inhibitor

Adenovirus induced cell death is said to be increased when cell death in population A is statistically significantly higher than the cell death in population D and wherein the difference in cell death between populations A and D is statistically significantly higher than the difference in cell death between populations B and C.

Adenovirus induced cell death is said to be selectively increased when cell death in population A is statistically significantly higher than the cell death in population D and cell death is not significantly different between populations B and C.

Cell death of an adenovirus infected cell is said to be increased when cell death in population A is statistically significantly higher than the cell death in population D. A method of the invention preferably further comprises timing the provision of said adenovirus to said first cell such that the adenovirus is provided before, at the same time, or within 24 hours of providing said first cell with said nucleic acid molecule comprising and/or encoding said miRNA, a miRNA mimic or miRNA inhibitor. In this way the method of the invention more effectively resembles the situation wherein the adenovirus comprises a nucleic acid molecule encoding said miRNA, miRNA mimic and/or miRNA inhibitor. Molecules that affect entry of the adenovirus into the cell are thereby disregarded in this setup of the experiment.

In the present invention it was found that some miRNA, miRNA mimics and/or miRNA inhibitors affect replication of the adenovirus, and/or cell death, in cells of different origins. Other miRNA, miRNA mimics and/or miRNA inhibitors were more cell origin specific. Both cell specific and promiscuous effects are useful. Specificity is desired when only a specific cell type is to be targeted, whereas more promiscuous molecules have the advantage that they are effective in a wider range of cells. If the adenovirus is for instance used in the treatment of cancer, it is an advantage if the adenovirus can replicate in and/or kill more types of cancer. This would reduce the number of different adenovirus vectors, and thus medicines that need to be produced. On the other hand, an adenovirus with a cell type specific miRNA, miRNA mimic and/or miRNA inhibitor according to the invention could be less toxic to other cell types than the target cell type. Thus in a preferred

embodiment a method of the invention further comprises determining whether said miRNA, miRNA mimic and/or miRNA inhibitor modulates adenovirus replication in, and/or cell death, preferably adenovirus induced cell death, of a second cell that is preferably permissive for adenovirus replication and wherein said second cell is of a different tissue origin than said first cell. In a preferred embodiment a method of the invention comprises determining whether said miRNA, miRNA mimic and/or miRNA inhibitor increases adenovirus replication in, and/or cell death, preferably adenovirus induced cell death, of said second cell. In one embodiment said miRNA, miRNA mimic or miRNA inhibitor increases replication in said second cell and increases cell death of said adenovirus infected second cell, preferably adenovirus induced cell death of said second cell. Said first and said second cell are preferably mammalian cells, preferably human cells. In a preferred embodiment of a method of the invention said first cell and/or said second cell is a cancer cell. Preferably said first cell is a prostate cancer cell. Alternatively, said first cell is a cancer cell of another type of cancer considered particularly amenable for treatment with an oncolytic adenovirus, such as e.g. a liver cancer cell, a head and neck cancer cell, a bladder cancer cell, a pancreas cancer cell, an ovarian cancer cell or a glioblastoma cell. Preferably said second cell is a non- small cell lung cancer cell, a breast cancer cell and/or an osteosarcoma cell. As a second cell these cancer cells are preferred because they allow the identification of miRNA or miRNA inhibitors with a broad- spectrum activity in various cancer cells, also in cancer cells that are not in the panel.

The adenovirus can be a replication defective adenovirus, i.e. an adenovirus that can replicate and produce functional progeny only in cell that supplies at least one essential adenovirus function in trans. Typically, such adenoviruses lack the capacity to code for a functional E1A, E1B, E2A and/or E4orf6 gene, although other deficiencies are also envisioned, such as for instance, a deficiency in an adenovirus late gene. When the adenovirus is a replication defective adenovirus it is preferred that the first and/or the second cell provides at least the adenovirus early genes that are essential for replication in trans. This can be done, for example, by separately transfecting DNA encoding the missing genes.

In a preferred embodiment, the adenovirus is a replication competent adenovirus. A replication competent adenovirus, as defined herein, is a virus that can replicate in an otherwise unmodified cell that is permissive for adenovirus replication. The replication competent virus contains in its genome all the cis and trans sequences that are necessary for replication in a permissive cell that is otherwise unmodified. A prototype for the replication competent adenovirus is wild type adenovirus. A cell is a cell permissive for adenovirus replication when a wild type adenovirus can replicate in said cell. Typical cells that are permissive for adenovirus replication are replicating cells, immortalized cell lines and cancer cells. Replication competent viruses can be deficient in one or more adenovirus genes that are not essential for producing progeny in an infected cell. Examples of such genes are the E3 region genes. These genes are dispensable in a replication competent virus, although the efficiency of replication might be slightly altered. A gene that performs a similar function in the virus can replace an essential gene in the virus to produce a replication competent virus. Non-limiting examples of such genes are similar genes from another serotype of adenovirus.

In a particularly preferred embodiment the virus is an oncolytic adenovirus. In a preferred embodiment, said adenovirus is a replication- competent, oncolytic adenovirus. In a particularly preferred embodiment said adenovirus is a conditionally replicating adenovirus (CRAd). Preferably said CRAd comprises an E1A gene with a 24-bp deletion corresponding to amino acids 122-129 in the CR2 domain of E1A. Preferably said adenovirus is a replication competent oncolytic adenovirus comprising an E1A gene with a 24- bp deletion corresponding to amino acids 122-129 in the CR2 domain of E1A.

An adenovirus of the invention is preferably an adenovirus serotype 5, an adenovirus serotype 24, an adenovirus serotype 35, or an adenovirus serotype 51 based virus, or a chimaeric adenovirus, for example based on a serotype 5 with serotype 35 tropism by replacing a part of the serotype 5 fiber with a part of the fiber of serotype 35. Said adenovirus may further comprise one or more modifications to provide it with an expanded or selective infection tropism known in the art. Non-limiting examples include mutations in the fiber knob that abrogate binding to the high affinity receptor CAR, such as e.g. deletion of fiber amino acids 489-492 TAYT (Roelvink et al., Science 19, 1568- 1571, 1999), replacing a large part of the fiber protein with an alternative trimerization domain such as e.g. the T(II) domain of reovirus sigma-1 protein and a binding ligand (WO 2007/094663), and insertion of a binding peptide in the fiber protein, preferably in the fiber knob HI-loop, such as e.g. an integrin binding cyclic RGD motif (Dmitriev et al., J.Virol. 72, 9706-9713, 1998) or a prostate specific membrane antigen (PSMA) binding WQPDTAHHWATL motif (US 2009/0274625). Said adenovirus may further provide expression of a therapeutic transgene such as, preferably, p53, a vaccine, or a secreted protein such as a growth factor. Said adenovirus is replication competent, in particular oncolytic. Said adenovirus is a human or primate adenovirus for infection of, and replication in, human and primate cells. Said adenovirus is another mammalian adenovirus, for example a canine or equine adenovirus for infection of, and replication in, dog cells or horse cells, respectively.

In another aspect the invention provides a method for providing a cancer cell with an anti-cancer treatment said method comprising providing said cancer cell with an adenovirus and with a nucleic acid molecule

comprising and/or encoding a miRNA, a miRNA mimic or miRNA inhibitor that modulates and preferably increases replication of said adenovirus in said cancer cell, and/or cell death of said cancer cell, preferably induced by said adenovirus. Replication of said adenovirus is said to be increased by said nucleic acid molecule comprising and/or encoding a miRNA, a miRNA mimic or miRNA inhibitor when the replication of said adenovirus is higher in the presence of said nucleic acid molecule comprising and/or encoding a miRNA, a miRNA mimic or miRNA inhibitor than in the same method but in absence of said nucleic acid molecule. With the same method is here preferably a method performed in parallel wherein essentially all parameters are the same but for the presence of said nucleic acid molecule comprising and/or encoding a miRNA, a miRNA mimic or miRNA inhibitor. In a preferred embodiment said nucleic acid molecule comprising and/or encoding a miRNA, a miRNA mimic or miRNA inhibitor is identified with a method for identifying functional miRNA, miRNA mimics or miRNA inhibitors as indicated herein above. It is preferred that said adenovirus encodes said miRNA, a miRNA mimic or miRNA inhibitor. Preferably the nucleic acid molecule comprises a miRNA, a miRNA mimic or miRNA inhibitor identified by a method of the invention. The invention further provides a method for sensitizing a cancer cell for adenovirus replication and/or cell death, preferably adenovirus induced cell death, said method comprising providing said cancer cell with a nucleic acid molecule comprising and/or encoding a miRNA, a miRNA mimic or miRNA inhibitor. Preferably said miRNA, miRNA mimic or miRNA inhibitor is identified with a method for identifying functional miRNA, miRNA mimics or miRNA inhibitors as indicated herein above. Preferably the nucleic acid molecule comprises a miRNA, a miRNA mimic or miRNA inhibitor identified by a method of the invention. In a preferred embodiment said method further comprises providing said (sensitized) cell with an adenovirus. In a preferred embodiment said adenovirus is an oncolytic adenovirus. Preferably, the method comprises administering to a patient in need of an anti-cancer treatment an adenovirus and a nucleic acid molecule comprising and/or encoding a miRNA, a miRNA mimic or miRNA inhibitor that increases replication of said adenovirus in, and/or cell death of an adenovirus infected cell preferably induced by said adenovirus.

The invention further provides a kit of parts for the treatment of cancer, said kit comprising an adenovirus and a nucleic acid molecule comprising and/or encoding a miRNA, a miRNA mimic or miRNA inhibitor that increases replication of said adenovirus in a cancer cell, and/or cell death of an adenovirus infected cancer cell, preferably cancer cell death induced by said adenovirus. In a preferred embodiment said miRNA, miRNA mimic or miRNA inhibitor is identified with a method for identifying functional miRNA, miRNA mimics or miRNA inhibitors as indicated herein above. Preferably said miRNA, miRNA mimic and/or miRNA inhibitor is a miRNA, miRNA mimic and/or miRNA inhibitor of the invention as defined herein above. Preferably said miRNA, miRNA mimic or miRNA inhibitor is a miRNA, a miRNA mimic or mRNA inhibitor of category III, category lib, category IV, category I, category Ila and/or category V of table 1. Preferably said miRNA, miRNA mimic or miRNA inhibitor is a miRNA, a miRNA mimic or mRNA inhibitor of table 1. Preferably said miRNA, miRNA mimic or miRNA inhibitor is a miRNA, a miRNA mimic or mRNA inhibitor of category I-IV of table 1.

Preferably said miRNA, miRNA mimic or miRNA inhibitor is a miRNA, a miRNA mimic or mRNA inhibitor of category II-IV of table 1. Preferably said miRNA or miRNA mimic is a miRNA of category II-III of table 1. In a particularly preferred embodiment said miRNA or miRNA mimic is a miRNA of category lib or III of table 1. In a particularly preferred embodiment said miRNA is a miRNA of category III of table 1.

The invention further provides a viral vector comprising an expression cassette comprising an expressible nucleic acid molecule that codes for a miRNA, a miRNA mimic or miRNA inhibitor that modulates and preferably increases adenovirus replication in a cell, and/or cell death of an adenovirus infected cell, preferably adenovirus induced cell death of a cell infected with said adenovirus. Preferably said miRNA, miRNA mimic and/or miRNA inhibitor is a miRNA, miRNA mimic and/or miRNA inhibitor of the invention as defined herein above. Preferably said miRNA, miRNA mimic or miRNA inhibitor is a miRNA, a miRNA mimic or mRNA inhibitor of category III, category lib, category IV, category I, category Ila and/or category V of table 1. Preferably said miRNA, miRNA mimic or miRNA inhibitor is a miRNA, miRNA mimic or mRNA inhibitor of table 1. The invention further provides a viral vector comprising an expression cassette comprising an expressible nucleic acid molecule that codes for a miRNA, a miRNA mimic or miRNA inhibitor of the invention as defined herein above. Preferably said miRNA, miRNA mimic or miRNA inhibitor is a miRNA, a miRNA mimic or mRNA inhibitor of category III, category lib, category IV, category I, category Ila and/or category V of table 1. The invention further provides an expression cassette comprising an expressible nucleic acid molecule that codes for a miRNA, a miRNA mimic or miRNA inhibitor that increases adenovirus replication in a cell, and/or cell death of an adenovirus infected cell, preferably adenovirus induced cell death of a cell infected with said adenovirus. Preferably said miRNA, miRNA mimic and/or miRNA inhibitor is a miRNA, miRNA mimic and/or miRNA inhibitor of the invention as defined herein above. Preferably said miRNA, miRNA mimic or miRNA inhibitor is a miRNA, a miRNA mimic or mRNA inhibitor of category III, category lib, category IV, category I, category Ila and/or category V of table 1. Preferably said miRNA, miRNA mimic or miRNA inhibitor is a miRNA, miRNA mimic or mRNA inhibitor of table 1. The invention further provides an expression cassette comprising an expressible nucleic acid molecule that codes for a miRNA, a miRNA mimic or miRNA inhibitor of category III, category lib, category IV, category I, category Ila and/or category V of table 1. Said expression cassette can advantageously be used in a kit of parts of the invention. Thus the invention further provides a kit of part comprising an adenovirus and an expression cassette comprising an

expressible nucleic acid molecule that codes for a miRNA, a miRNA mimic or miRNA inhibitor that increases adenovirus replication in, and/or cell death of an adenovirus infected cell, preferably adenovirus induced cell death of a cell infected with said adenovirus. Preferably said miRNA, miRNA mimic and/or miRNA inhibitor is a miRNA, miRNA mimic and/or miRNA inhibitor of the invention as defined herein above. Preferably said miRNA, miRNA mimic or miRNA inhibitor is a miRNA, a miRNA mimic or mRNA inhibitor of category III, category lib, category IV, category I, category Ila and/or category V of table 1. Preferably said miRNA, miRNA mimic or miRNA inhibitor is a miRNA, miRNA mimic or mRNA inhibitor of table 1.

The promoter driving the expression of the nucleic acid molecule that codes for the miRNA, miRNA mimic or miRNA inhibitor can be adapted to specific needs. For instance, a short hairpin miRNA (shmiRNA) precursor molecule is preferably expressed by means of a pol-III promoter. Thus the expression cassette preferably contains pol-III promoter in linkage with nucleic acid molecule coding for said miRNA, miRNA mimic and/or miRNA inhibitor. A shmiRNA typically comprises a 50-100 nucleotide long RNA molecule comprising two stretches of nucleotides that are essentially complementary and can base-pair, whereby the two stretches are

interconnected through a hairpin turn. The shmiRNA hairpin structure is cleaved by the cellular machinery into 18-23 (typically 19) nucleotide-long double stranded RNA molecules with 2 nucleotide-long 3' overhangs with one of the strands exhibiting extensive reverse complement homology to a part of a mRNA transcript from a target gene. The miRNA, miRNA mimic, or miRNA inhibitor can exhibit its function upon release of the miRNA, miRNA mimic or miRNA inhibitor from the short hairpin. Expression of the shmiRNA can be driven by a polymerase II or polymerase III enhancer/promoter. The adenovirus can also code for an RNA molecule that is longer than 100 nucleotides. Such RNA molecules can, for instance contain two or more miRNA, miRNA mimics and/or miRNA inhibitors. Such longer transcripts are preferably transcribed by means of a suitable pol-II promoter. Thus preferably said expression cassette encodes a precursor RNA comprising said miRNA, a miRNA mimic or miRNA inhibitor. Preferably, said precursor RNA is a pri- miRNA or pre-miRNA encoding said miRNA. In a particularly preferred embodiment said precursor RNA encodes a pre-miRNA comprising said miRNA as part of a pre-miRNA stem-loop sequence, wherein said pre-miRNA is flanked by at least 30 single- stranded RNA nucleotides at the 5' and 3' sides of the miRNA stem-loop sequence. In another embodiment said precursor RNA molecule essentially comprises part of the endogenous pri-miRNA sequence of said miRNA, wherein said part comprises the miRNA stem-loop sequence flanked by at least 30 single-stranded RNA nucleotides at the 5' and 3' sides of the miRNA stem-loop sequence, preferably at least 60 single-stranded RNA nucleotides at the 5' and 3' sides of the miRNA stem-loop sequence, more preferably at least 100 single-stranded RNA nucleotides at the 5' and 3' sides of the miRNA stem-loop sequence.

Various viral vectors are used for delivering nucleic acid to cells in vitro or in vivo. Non-limiting examples are vectors based on Herpes Viruses, Pox- viruses, Adeno-associated virus, Lentivirus, and others. In principle all of them are suited to deliver the expression cassette comprising an expressible nucleic acid molecule that codes for a miRNA, a miRNA mimic or miRNA inhibitor that modulates adenovirus replication in, and/or adenovirus induced cell death of a cell infected with said adenovirus. In a preferred embodiment said viral vector is an adenoviral vector, preferably a replication competent adenovirus. In this way only one gene transfer vector needs to be produced.

In a further preferred embodiment, a replication competent adenovirus according to the invention is a conditionally replicating adenovirus (CRAd). A CRAd will only replicate in cells in which the particular conditions exist that are required for replication of the CRAd. CRAds are designed to meet the specific requirements for replication in a chosen type of cell and not in other types of cells. This property makes CRAds particularly useful for several embodiments of the present invention where the intention is to treat a disease by specific lytic replication of the recombinant adenovirus according to the invention in diseased cells in an animal or human body resulting in specific removal of said diseased cells from said body.

A CRAd comprises an adenoviral genome from which one or more parts that are necessary for efficiently completing at least one step of the adenovirus infectious life cycle under certain physiological conditions (herein also "first conditions") but not under certain other physiological conditions (herein also "second conditions") have been modified, removed or have been otherwise engineered to be not expressed under the first conditions. Said first and second conditions could, e.g., be dictated by the physiological conditions that exist in a particular type of cells (herein also "first cells"), but not in another type of cells (herein also "second cells"). Such a first type of cell is e.g. a cell derived from a particular type of tissue, where said cell contains a protein that is not or much less present in cells from other tissues (second type of cells). An example of a second type of cell is a cell that has lost proper cell growth control, such as e.g. a cancer cell, where said cell either lacks a protein that is present in cells that have not lost proper cell growth control or where said cell has gained expression (or over-expression) of a protein that is not or much less present in cells that have not lost proper cell growth control.

Another example of a second condition is a condition that exist in a particular stage of the cell cycle or in a particular developmental stage of the cell, where a certain protein is expressed specifically. Thus, CRAds can be designed such, that replication thereof is enabled in particular cells, such as cancer cells or a particular type of cancer cells, whereas in normal cells, replication of CRAds is not possible, or strongly reduced.

A preferred CRAd is provided by an adenovirus according to the invention, wherein said adenovirus comprises at least one mutation in one or more genes from the group consisting of E1A, E1B, E4, and VA-RNAs, to achieve selective replication in tumors. An adenovirus according to the invention preferably carries a mutation in the E1A region encompassing at least a part of the CR2 domain of E1A, preferably a deletion encompassing amino acids 122 to 129 (LTCHEAGF) of E1A. The term gene, as used herein, comprises the complete genomic region that is required for expression of a gene including, for example, the enhancer/promoter region and intronic en exonic sequences.

An adenovirus according to the invention may further comprise modifications that increase its replication potential, such as e.g.

overexpression of the E3-11.6K ADP gene (Doronin et al., Virology, 305, 378- 387, 2003) or deletion of the E1B-19K gene (Sauthoff et al. Hum. Gene Ther. ll(2000):379-388), or that increase the replication selectivity for a certain type of cells, including but not limited to the modifications to make CRAds (supra), or that reduce the immunogenicity (i.e., their potency to induce an immune response when introduced into an animal body), such as e.g. retention of the E3B region (Wang et al., Nature Biotechnol. 21(2003):1328-1335).

An adenovirus according to the invention may further be modified to express one or more transgenes, such as e.g. a gene encoding a cytokine, a pro- apoptotic protein, an anti- angiogenic protein, a membrane fusogenic protein or a prodrug converting enzyme.

Expression control sequences for expression of a miRNA, miRNA mimic or miRNA inhibitor in a target cell preferably comprise a polymerase II or polymerase III enhancer/promoter. A preferred polymerase II promoter for expression of a pri-miRNA is a selective RNA polymerase II promoter, such as a tissue- specific or a cell-specific promoter that directs expression of the miRNA, miRNA mimic or miRNA inhibitor specifically or exclusively in the target cell. Expression control sequences for expression of a miRNA, miRNA mimic or miRNA inhibitor preferably also comprise transcriptional stop sequences such as a poly(A) signal for polymerase Il-mediated expression, and a termination signal such as a stretch of at least 4 consecutive thymidine nucleotides for polymerase Ill-mediated expression.

A preferred polymerase II promoter is selected from a CMV promoter, the immediate early gene of human cytomegalovirus, the SV40 promoter, and the long terminal repeat of Rous sarcoma virus. Another preferred promoter comprises regulatable elements, such as tetracycline, radiation or hormone regulated elements allowing control of the timing and level of transcription driven by the promoter. Preferred expression control sequences according to the invention comprise a selective RNA polymerase II promoter.

In another embodiment, the one or more expression control sequences in an adenovirus according to the invention comprise an RNA polymerase III promoter. Preferred polymerase III promoter sequences are selected from the group consisting of 5S rRNA, tRNAs, VA RNAs, Alu RNAs, HI, and U6 small nuclear RNA promoter sequences. A preferred host cell for an adenovirus according to the invention, is a cancer cell. The term cancer refers to malignant primary and/or metastasized cancers. Examples of a cancer include, but are not limited to, a carcinoma; a sarcoma, a lymphoma, a leukemia, or a myeloma. A cancer can be present in any tissue or part of a body, including but not limited to bone, brain, eye, breast, skin, bladder, lung, ureter, urethra, thyroid, parathyroid, salivary gland, kidney, prostate, genital system including ovary and testis,

endometrium, blood/hematologic system, or in a gastrointestinal tissue. In a preferred embodiment said cancer is a prostate cancer.

The invention further provides an adenovirus according to the invention for use as a medicament. The invention further provides an adenovirus according to the invention for use as a medicament for the treatment of cancer, preferably prostate cancer. Adenoviruses are propagated according to standard methods in the fields of adenovirology and adenoviral vectors. The preferred method of propagation is by infecting a suitable cell line that allows replication of adenoviruses. An example of a method for generating adenoviruses may further comprise the steps of collecting the cells when they show cytopathic effect, indicative of virus production and freeze-thawing of the cells to generate a cellular extract. The virus is purified from the cellular extract using standard techniques, e.g. banding on a cesium chloride gradient and dialysis, for example against Phosphate-Buffered Saline- 10% glycerol. The dialyzed virus may be aliquoted and stored at -80 oC. The quantification of the number of plaque -forming adenovirus particles and units is performed according to standard protocol. A saline phosphate buffer with 10% glycerol is a standard formulation for the storage of adenovirus.

The adenovirus can be administered to an animal or human body to infect cells in vivo. Administration can be done via several routes including, but not limited to, locoregional injection into the tumor or into a body cavity where the tumor is located, injection into the blood circulation, inhalation and application to the surface of a certain body area. Following infection, the replication competent adenovirus can replicate and spread to other cells, provided that the infected cells support replication of said recombinant adenovirus. The replication competent adenovirus can thus be used to re-infect new cells to further propagate and expand said replication competent adenovirus.

The adenovirus preferably is formulated into an aqueous or solution medium for the preservation of viral particles which can directly be

administered to an organism. The formulation preferably comprises

pharmaceutical acceptable salts and excipients such as, for example, human serum albumin, sugars such as sucrose and mannitol, and/or a surfactant such as, for example, a difunctional block copolymer surfactant terminating in primary hydroxyl groups (Pluronic F68™).

Cancer cells and cell lines are the result of neoplastic

transformation. The genetic events underlying neoplastic transformation include activation of proto-oncogenes and inactivation of tumor-suppressor genes. A major player in this respect is the gene encoding the tumor- suppressor protein p53. The p53 protein is the central coordinator of damage- induced cell-cycle checkpoint control. In a perturbed cell, p53 can induce growth arrest and cell death. p53 exerts these effects by functioning as a specific transcription factor that controls the expression of a large panel of genes involved in growth control, DNA repair, cell-cycle arrest, apoptosis promotion, redox regulation, nitric oxide production, and protein degradation (Polyak et al., Nature 389(1997):237-238; El-Deiry, Sem. Cancer. Biol.

8(1998):345-357; Yu et al., Proc. Natl. Acad. Sci. USA 96(1999):14517-14522; Hupp et al., Biochem. J. 352(2000):1-17; and references therein). The induction of cell death by p53 is mediated at least in part by activation of pro-apoptotic death genes of the bcl-2 family, such as bax, bak, bad, bid, bik, bim, bok, blk, hrk, puma, noxa and bcl-xS (Miyashita and Reed, Cell 80(1995):293-299; Han et al., Genes Dev. 10(1996):461-477; Zoernig et al., Biochim. Biophys. Acta 1551(2001):F1-F37). On the other hand, anti-apoptotic members of the bcl-2 family, such as bcl-2 itself and bcl-xL, bcl-w, bfl-1, brag-1 and mcl-1 inhibit p53-dependent cell death (Zoernig et al., supra). The anti-apoptotic protein Bax Inhibitor- 1 (BI-1) suppresses apoptosis through interacting with bcl-2 and bcl-xL (Xu and Reed, Mol. Cell l(1998):337-346). The immediate effector proteins of p53 as well as p53 itself target mitochondria, thereby releasing cytochrome c into the cytosol to activate the caspase cascade via the initiator caspase-9/Apaf-l complex (Juergensmeier et al., Proc. Natl. Acad. Sci. USA 95(1998):4997-5002; Fearnhead et al., Proc. Natl. Acad. Sci. USA

95(1998):13664-13669; Soengas et al., Science 284(1999):156-159; Marchenko et al., J. Biol. Chem. 275(2000):16202-16212). Negative regulators of the caspase cascade include but are not limited to members of the Inhibitor of Apoptosis Protein (IAP) family of proteins, such as cIAPl, cIAP2, cIAP3, XIAP and survivin (Zoernig et al., supra).

The inventors found that oncolysis and release of adenovirus progeny from infected cancer cells can be accelerated by restoring p53 functions in said cancer cells (van Beusechem et al., Cancer Res.

62(2002):6165-6171; WO 03/057892, incorporated by reference herein). Said restoring of p53 functions is done by expressing in said cancer cells a restoring factor, i.e. a functional factor of the p53-dependent apoptosis pathway, the function whereof is not or insufficiently expressed in said cancer cells, wherein said restoring factor preferably comprises a protein (WO 03/057892). Hence, said restoring factor is an essential positive component of the p53-dependent apoptosis pathway.

In a further embodiment, a replication competent adenovirus according to the invention further comprises a DNA sequence that encodes at least one restoring factor functional in restoring the p53 dependent apoptosis pathway in the host cells, operably linked to one or more expression control elements, functional in the host cells. Said restoring factor preferably is selected from the pro-apoptotic death genes of the bcl-2 family, such as bax, bak, bad, bid, bik, bim, bok, blk, hrk, puma, noxa and bcl-xS (Miyashita and Reed, Cell 80(1995):293-299; Han et al., Genes Dev. 10(1996):461-477; Zoernig et al., Biochim. Biophys. Acta 1551(2001):F1-F37), and/or p53, or a functional part or derivative thereof. A preferred restoring factor functional in restoring the p53 dependent apoptosis pathway is p53.

In a preferred embodiment of the invention said adenovirus, and preferably an oncolytic CRAd of the invention further comprises a nucleic acid encoding p53, or a functional equivalent thereof. A functional equivalent of p53 is a mutant p53 with essentially the same apoptosis restoring capacity in kind not necessarily in amount as wild type p53.

A loss of normal function of p53 is associated with resistance to programmed cell death, cell transformation in vitro and development of cancers in vivo. In approximately 50% of human cancers the gene encoding p53 is non-functional through deletion or mutation (Levine et al, Nature

351(1991):453-456; Hollstein et al, Science 253(1991):49-53; Chang et al, J. Clin. Oncol. 13(1995):1009-1022). In many of the other 50% cancer cells that do express wild-type p53 protein, p53 function is still hampered by the action of a p53 antagonist. An example of a p53 antagonist is MDM2. Loss of the tumor- suppressor protein pl4ARF or overexpression of MDM2 protein can lead to functional inactivation of p53 by binding to the MDM2 protein and subsequent degradation. In addition, even if p53 function itself is intact, p53-dependent cell death can be hampered due to overexpression of anti-apoptotic proteins acting on the p53 pathway down- stream from p53, such as the anti-apoptotic bcl-2 and IAP family members and BI-1. Another example is p73DeltaN, which binds to p53-responsive promoters competing with p53, thereby antagonizing p53-dependent cell death (Kartasheva et al, Oncogene 21(2002):4715-4727).

Therefore, in one embodiment, a replication competent adenovirus according to the invention further comprises an expression cassette that mediates expression of one or more RNAi-mediating molecules that are specific for one or more p53 antagonists and/or inhibitors of the p53 pathway or a combination thereof in a cell, as described and defined in WO 2005/100576 included herein by reference. The expression of one or more RNAi- mediating molecules that are specific for one or more p53 antagonists and/or inhibitors of the p53 pathway or a combination thereof in a cell will enhance the lysogenic activity in a target cell that comprises functional p53.

Said antagonists and/or inhibitors of the p53 pathway are preferably selected from synoviolin, MDM2, Pirh2, COP1, Bruce, HPV-E6, herpesvirus-8 LANA, Pare, Mortalin, Plk-1, BI-1, p73DeltaN, bcl-2, bcl-xL, bcl-w, bfl-1, brag- 1, mcl-1, cIAPl, cIAP2, cIAP3, XIAP and survivin. The expression cassette further comprises one or more expression control sequences, functional in the said host cells such as an enhancer/promoter and a terminator that are operably linked to the one or more RNAi-mediating molecules. It has been found that a cancer cell can contain more than one p53 antagonists and/or inhibitors of the p53 pathway. Such cells are more effectively lysed when they are provided by RNAi against at least two of those p53 antagonists and/or inhibitors of the p53 pathway.

The invention further provides a method of lysing a cancer cell comprising the step of providing the cancer cell with a virus according to invention, or a kit of parts according the invention, thereby inducing lysis of the cancer cell. The cancer cell is preferably present in an animal body, preferably a human body. A preferred cancer cell is a prostate cancer cell.

The invention further provides a method for treatment of a subject suffering from a cancer, preferably a prostate cancer, the method comprising the step of administering to the said subject an effective amount of the replication competent virus according to the invention or a kit of parts according to the invention. Brief description of the figures:

Figure 1.

PC-3 cell viability after miRNA mimic or inhibitor transfection, with or without Ad5 infection.

MicroRNA mimics and inhibitors selected for cell death modulating activity in primary cell viability screens in the presence of Ad5 were transfected into PC-3 cells on six replicate plates. Three plates were infected with Ad5 and three plates were not infected. Four days after infection, cell viability was

determined using CTB reagent. Data shown are the mean results for each miRNA molecule from two independent experiments. The x-axis depicts the mean robust Z-scores of miRNA molecules alone. A Z-score below -3, which correlated with less than 77% cell viability compared to negative controls, was considered to indicate direct cytotoxicity of the miRNA molecule. The y-axis depicts the mean log-transformed relative cell viability in Ad5-infected cultures compared to the corresponding cultures without Ad5. In mock transfected controls, Ad5 reduced cell viability on average to 57% (²Log ratio with/without Ad5 = -0.82). The mean ²Log ratio minus 3 standard deviations was -1.13. MicroRNA molecules causing a ²Log ratio below -1.13 were considered to sensitize PC-3 cells for Ad5-induced cell death. Data from the 43 negative control transfections included in the experiments are not shown. Classification of hits into Categories I, II, and III is indicated.

Figure 2.

Correlation between effects of miRNA mimics (Figure 2A) and miRNA inhibitors (Figure 2B) on PC-3 cell viability (x-axes) and adenovirus replication (y-axes).

Plotted are the mean robust B- scores as determined using CellHTS2 software of three independent cell viability screens and three independent adenovirus replication screens using miRIDIAN libraries. Scores below -3 or above 3 were considered to represent hits, i.e., miRNA molecules changing cell viability and/or adenovirus replication. Note that not all cell viability hits are shown in this figure, because some were identified on the basis of robust Z-scores. Figure 3.

Functional adenovirus progeny production and release by PC-3 cells transfected with miRNA mimics and infected with replication-competent adenovirus expressing firefly lucif erase.

PC-3 cells were transfected with miRNA mimics and infected with AdEl+Luc virus. Progeny virus was harvested from culture medium and cell lysate separately. Virus titers were determined on the basis of luciferase activity one day after infecting PC-3 cells with diluted culture supernatant or cell lysate. Data shown are mean relative virus titers + standard deviations, in

percentages compared to irrelevant control miRNA-transfected control cultures. Each miRNA mimic was tested in 3-6 independent experiments. Bars are grouped according to the miRNA mimic Categories (see Table 1). From left to right; Category I (3 miRNA mimics), Category Ila (4 miRNA mimics), Category lib (9 miRNA mimics) and Category III (5 miRNA mimics). (A) Total progeny virus production, i.e., in cell lysate plus in culture medium, (B) progeny virus released into the culture medium.

Figure 4.

Cell viability of four different cancer cell lines after miRNA mimic transfection and MOI 100 Ad ^'5 infection.

Results are of two individual experiments each performed in triplicate. Cell viability was analyzed at the day post infection indicated between parentheses for the two experiments in each panel. Depicted are mean cell viability percentages relative to the "no miRNA" controls, where cells were only infected with Ad5. Irr miRNA = irrelevant miRNA. Figure 5.

Sigmoidal dose-response curves, showing relative cancer cell viability after combined treatment with Ad5 in a dose range and miRNA mimics at a fixed dose, compared to the effect of the miRNA mimics alone.

Examples of miRNA mimics showing clear effects compared to Ad5 alone are given. The dose response curves for Ad5 alone (dashed lines) are shown for comparison.

Figure 6.

Relative oncolytic potency of Ad 5 on four human cancer cell lines transfected with miRNA mimics.

Depicted are the Relative l/EC-50 values, normalized to the "no miRNA" condition (i.e., Ad5 infection only), of two or three independent experiments. Values higher than 1 indicate increased cell death after transfection with the given miRNA mimic and Ad5 infection, compared to Ad5 infection alone. Note: In experiment 1, miR-26b and miR-383 mimics were not tested on U20S cells. Irr miRNA = irrelevant miRNA.

Figure 7. Oligo sequences in miRIDIAN miRNA mimic and inhibitor libraries used. The inhibitor sequences depicted comprise the reverse complement of the miRNA that they inhibit plus flanking sequences that render the inhibitor more effective.

Figure 8. Mature sequence of the miRNA to which the miRIDIAN inhibitors depicted in figure 7 bind. The sequence of the inhibitor in figure 7 comprises the reverse complement of the miRNA depicted in this figure. The sequence of the inhibitors in figure 7 further comprises other nucleotides, flanking each side that render the inhibitor more effective. Figure 9.

Relative expression levels as compared to uninfected cells of mature miR-1 (A) and PTK-9 mRNA (B), 24 and 48 hours after infection with viruses expressing the indicated miR-1 precursor formats. Expression levels of miR-1 or PTK-9 as compared to uninfected cells were determined by standardizing their Ct values against the internal controls, RNU48 for miR-1 and GAPDH for PTK-9, using the following formula (also known as the AACt method):

Expression = 2^A-((Ct[gene of interest] - Ct[control gene])infected - (Ct[gene of interest] - Ct[control gene])uninfected)

Figure 10.

Acumen eX3 scans of PC-3 cells in 96-well cultures transfected with the indicated miRNA mimics and infected with AdDelta24-CMV-GFP oncolytic adenovirus. Whole-well scans were made 5 days post infection at 1 by 4 micrometer resolution using the 488 nm laser and the first detection channel. Fluorescence intensity and position readings were stored and analyzed using Acumen Explorer software. Three-dimensional intensity profiles were used to define GFP-positive cells (fluorescent objects with 20-1,000 micrometer width and depth). Images shown are virtual well views generated using Acumen Explorer software and stored in TIFF format. The value given in the lower right corner of each image represents the number of GFP-positive cells identified in the image field.

EXAMPLES Example 1.

Identification of miRNAs that modulate cell death of adenovirus-infected PC-3 prostate cancer cells.

In this example, we identified miRNAs that influence death of adenovirus-infected cancer cells. For proof-of-concept, we performed miRNA screens on a single cancer cell line, PC-3 prostate cancer cells. To allow identification of relevant miRNAs irrespective of the endogenous miRNA expression profile of the test cell line, screens were performed with miRNA mimic and inhibitor libraries. Synthetic miRNA mimics are designed to enter the miRNA pathway and act as mature miRNA species; synthetic miRNA inhibitors are designed to specifically target and irreversibly bind endogenous miRNAs. The expectation is that introduction of miRNA mimics may identify relevant miRNAs that are not or insufficiently expressed in PC-3 cells and introduction of miRNA inhibitors may identify relevant miRNAs that are highly expressed in PC-3 cells. Three independent replicate screens were done on PC-3 cells with Dharmacon miRIDIAN human miRNA mimic (Cat. No. CS- 001000, Lot 060417) and inhibitor (Cat. No. IS-001000, Lot 060417) libraries, comprising 426 human miRNAs (miRBase v.8). PC-3 cells were grown at 37°C and 5% C02 in a humidified incubator in Dulbecco's modified Eagle's medium F12 (DMEM-F12) supplemented with 10% fetal bovine serum and antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin). Cells were seeded 5,000 cells per well in 96-well microtiter plates in 80 microliter culture medium without antibiotics; and transfected 2 days later with 100 nM miRNA mimic or inhibitor with 0.2 microliter DharmaFect-2 transfection reagent (Dharmacon). Transfection mixes were prepared by mixing 10 microliter miRNA diluted in IX siRNA buffer (Dharmacon) with 10 microliter DharmaFect-2 diluted in serum-free, antibiotics-free culture medium and incubation at room temperature for 20-40 minutes. Mixtures were then added to the cells to reach a total volume of 100 microliter per well. One day later, cells were infected with wild type adenovirus serotype 5 (Ad5). Wild type Ad5 was chosen for these experiments in order to identify relevant miRNAs independent of specific oncolytic adenovirus modifications. Adenovirus was added in 20 microliter culture medium with FBS and antibiotics per well and infection was done at high MOI (100 IU/cell), such as to infect all cells in the culture. Pilot experiments had shown that under the experimental conditions used, complete infection of PC-3 cells was reached at MOI 30. Adenovirus infection was done only one day after miRNA transfection, when miRNA effects are expected not to be fully developed. This was done to avoid that miRNAs involved in the adenovirus uptake process would be identified. Such miRNAs are not useful for the ultimate goal, i.e., to empower oncolytic adenoviruses by expressing the miRNA from the adenovirus genome. Cell viability was measured using Cell- Titer Blue reagent (CTB; Promega), 3-4 days after adenovirus infection, when cell viability in control cultures declined by 30-50%. To this end, 10 microliter CTB was added to all wells and mixed by shaking. After 2 hours incubation at 37°C and 5% C02 in a humidified incubator, reactions were stopped by adding 50 microliter 3% SDS. After at least 30 minutes incubation at room

temperature, cell viability was determined by measuring fluorescence at 540 (25) nm excitation and 590 (20) nm emission wave lengths using a Tecan Infinite F200 reader. Data analysis was done using CellHTS2 software and primary hits were selected on the basis of robust Z and B-scoring. The primary hit selection threshold was set at a mean Z-score above 2 or below -2 or a mean B-score above 3 or below -3. This hit selection did not select any of the negative controls included in the screens. This identified 36 miRNA mimics (26 that increased cell death and 10 that decreased cell death) and 21 miRNA inhibitors (16 that increased cell death and 5 that decreased cell death).

Next, all primary hit miRNA mimics and inhibitors were re- screened in two independent screens, each consisting of six plate sets, i.e. three that were subsequently subjected to Ad5 infection as in the primary screens and three that were subjected to control medium without Ad5. Test miRNA mimics and inhibitors were randomly distributed over plates; at different positions in the two secondary screens and at different positions than in the primary screens. Each plate set furthermore included 43 negative control mock transfections. Experimental setup was the same as in the primary screens, except for the miRNA mimic or inhibitor concentration, which was reduced to 50 nM. In these experiments, Ad-infected controls caused approximately 40% cell kill at the time of analysis. The data of the screens were analyzed in two ways. First, robust Z-score analysis was done on plates without virus, to identify direct miRNA-mediated effects on cell viability. Second, the mean ²Log (ratio with/without Ad5) was calculated, to identify miRNA-mediated sensitization to or protection against Ad5-induced cell death. One miRNA mimic was excluded from further analysis, because it showed inconsistent results, i.e., protection against cell death in primary screens and sensitization to Ad5-induced cell death in secondary screens. Figure 1 shows the data from secondary screens on the other miRNA mimics and inhibitors, plotting direct miRNA-mediated effects on cell viability (x-axis) against miRNA-mediated effects on Ad5-induced cell death (y-axis). Hit selection thresholds were set as follows: mean robust Z-score below -2 or above 2 for direct miRNA-mediated effects on cell viability; and mean ²Log (ratio with/without Ad5) higher or lower than the mean +/- 3 standard deviations of the 43 negative controls included in the screens. The secondary screens did not confirm any of the 15 miRNA mimics or inhibitors protecting against PC- 3 cell death found in the primary screens. Apparently, the detection window for these events was too small for the stringent selection criteria set in the secondary screens. In contrast, the secondary screens did confirm 24 primary hits that increased PC- 3 cell death, only one of which was an inhibitor. The confirmed hits were classified into three categories, i.e. Category I: five that were directly cytotoxic to PC-3 cells, but that did not significantly change Ad5-induced cell death; Category II: 14 that were directly cytotoxic to PC-3 cells and that on top of that significantly increased Ad5-induced cell death; and Category III: five that alone did not significantly influence PC-3 cell viability, but that did

significantly increase Ad5-induced cell death. Category II miRNAs were later divided into two sub-classes (see Example 2). A list of the identified miRNAs with confirmed effect on PC-3 cell viability and their classification is given in Table I.

Example 2.

Identification of miRNAs that modulate Ad5 replication in PC-3 prostate cancer cells.

In this example, we identified miRNAs that influence replication of human adenovirus in cancer cells. We performed three independent replicate screens with the miRIDIAN mimic and inhibitor libraries and PC-3 cells as in Example 1, but now infecting the cells with AdDelta24.SA-GFP.

AdDelta24.SA-GFP was made according to the method described by Carette et al. (J. Gene Med. 7, 1053-1062, 2005). Briefly, the green fluorescent protein (GFP) cDNA gene was obtained by polymerase chain reaction amplification using primers with overhanging Xbal and Sacl restriction sites and pAdTrack (He et al., Proc. Natl. Acad. Sci. USA 95, 2509-2514, 1998) as template; and inserted into Xbal/Sacl- digested pABS.4-SA-MSC (Carette et al., J. Gene Med. 7, 1053-1062, 2005). This plasmid was digested with Pad and inserted into Pad-digested pBHGll (Microbix Biosystems). The kanamycin resistance gene was removed by digestion with Swal followed by self-ligation. The resulting plasmid was recombined with pXCl-Delta24 (Fueyo et al., Oncogene 19, 2-12, 2000) following co-transfection into HEK293 cells. Viruses were plaque purified and propagated on A549 NSCLC cells. AdDelta24.SA-GFP is a recombinant oncolytic adenovirus that expresses GFP driven by the

endogenous adenovirus major late promoter. Late gene expression driven by this promoter is replication- dependent. GFP expression in cells infected with this virus is thus an indirect measure for adenovirus DNA replication in the host cell. Also in these screens, adenovirus infection was done only one day after miRNA transfection, to avoid that miRNAs involved in the adenovirus uptake process would be identified. Based on pilot experiments, cells were infected at a lower, non- saturating MOI (10 IU/cell); and GFP expression was measured two days after infection. To this end, cells were lysed by adding 50 microliter 3% SDS and GFP expression was determined by measuring fluorescence at 488 (10) nm excitation and 530 (25) nm emission wave lengths using a Tecan Infinite F200 reader. Screen results were analyzed by B- score calculation using cellHTS2 software and the hit selection threshold was set at a mean B- score above 3 or below -3. This identified 9 miRNA mimics (3 that increased Ad5 replication and 6 that decreased Ad5 replication) and 3 miRNA inhibitors (2 that increased Ad5 replication and 1 that decreased Ad5 replication). Figure 2 shows the robust B scores of CTB cell viability screens plotted against the robust B scores of GFP adenovirus replication screens (mimics in Figure 2A and inhibitors in Figure 2B). As can be seen, most miRNA mimics and inhibitors that influenced cell viability did not

significantly alter Ad5 replication. However, four miRNA mimics scored double-positive, decreasing cell viability and Ad5 replication. These double- positive hits were all from Category II, i.e. showing direct cytotoxicity and augmenting Ad5-induced cell kill (see Example 1). The fact that they also decreased GFP expression in AdDelta24.SA-GFP infected cells suggested that they killed PC-3 cells independently from Ad5 in a way that interfered with Ad5 replication. Therefore, Category II hits were subdivided into Category Ila (interfering with Ad5 replication) and Category lib (not affecting Ad5 replication). Two categories were added, i.e., Category IV: molecules that stimulated Ad5 replication, without affecting cell viability; and Category V: molecules that inhibited Ad5 replication, without affecting cell viability. In Category IV, five hits were found. However, one was a miRNA mimic that was present on three locations in the miRIDIAN library. Only one of these was selected as a hit. It was therefore considered to be a false positive result. In Category V, three hits were found. Table 1 lists the identified miRNAs and their classification. Note that an observed effect of a miRNA inhibitor in PC-3 cells suggests that the corresponding mimic could yield the opposite effect in certain cells with less expression of that miRNA than in PC-3 cells; and that an observed effect of a miRNA mimic in PC-3 cells suggests that the corresponding inhibitor could yield the opposite effect in certain cells with more expression of that miRNA than in PC-3 cells.

Example 3.

Effects of miRNA mimics on functional Ad5 progeny production in PC-3 prostate cancer cells. Selected miRNA mimics from different Categories identified in

Examples 1 and 2 were investigated for their effect on functional infectious virus production in PC-3 cells and progeny virus release from these cells. To this end, PC-3 cells were seeded 10,000 cells/well in 96-well plates and transfected with 50 nM miRNA mimics and 0.2 microliter Dharmafect-2 in a total volume of 100 microliter per well. The next day, the cells were infected at high MOI (500 IU/cell) with recombinant adenovirus AdEl+Luc expressing firefly luciferase (Grill et al., Mol. Ther. 6, 609-614, 2002). AdEl+Luc has a wild type Ad5 genome, with the E3-gpl9k open reading frame replaced by a firefly luciferase open reading frame. Excess virus was washed away after a few hours and intracellular virus was allowed to replicate. When cytopathic effects became apparent in control cultures, cell and supernatant fractions were collected separately and virus was released by multiple freeze-thaw cycles. The supernatant fraction was collected by harvesting the upper 50 microliters from the wells. This fraction was considered to contain 50% of the total amount of released virus. The remaining 50 microliters culture medium with adherent, semi-adherent and detached cells were collected as cell fraction. The amount of progeny virus in the cells was calculated by subtracting the virus titer in the supernatant fraction from the titer in the cell fraction. To determine the infectious virus titers, samples were diluted and 10 microliter was used to infect fresh PC-3 cells seeded 10,000 cells/well in 100 microliter DMEM-F12 supplemented with 10% fetal bovine serum and antibiotics in 96- well plates the day before. A dilution titration of AdEl+Luc with known IU titer (determined using Adeno-X Rapid Titer Kit; BD Biosciences) was taken along in triplicate. One day after infection, the cells were lysed and luciferase expression was measured using the Promega Luciferase Chemiluminescent Assay System and an EG&G Berthold Lumat LB 9507 luminometer. Assay development experiments had shown that IU virus titer correlated with firefly luciferase expression (relative light units, RLU) in the range of 100 to 100,000 IU per 10 microliter. In addition, pre- diluting samples with high IU titer up to more than 1,000-fold caused a linear decrease in RLU values. Together, this allowed titrating AdEl+Luc over a range of 100 to 10(E+8) IU per 10

microliter, i.e., 1,000 to 10(E+9) IU per 100 microliter culture or 0.1 to 100,000 IU per cell. However, titers determined this way in individual wells differed considerably (approximately 4-fold difference between the highest and lowest value in triplicates). Therefore, the effects of each miRNA were determined in at least three independent experiments and differences were only considered meaningful if the titer was increased or decreased more than 2-fold.

Figure 3A shows the relative total infectious virus output of PC-3 cells transfected with miRNA mimics from Categories I, Ila, lib and III, as compared to PC-3 cells transfected with an irrelevant miRNA mimic control, i.e. a C. elegans miRNA with minimal sequence identity to human miRNAs (cel-miR-67). As can be seen, Category I and III miRNAs did not significantly change adenovirus progeny production in PC-3 cells, despite their direct cytotoxic and selective sensitization to adenovirus effects, respectively. In contrast, all Category Ila miRNAs decreased adenovirus progeny production (by on average 65-80%). This correlated with the inhibition of adenovirus replication observed for these miRNA mimics in Example 2. Among Category lib mimics, there were two molecules that inhibited adenovirus progeny production by more than 50%. Possibly, these miRNAs inhibit a step in the adenovirus life cycle after late gene expression. The majority of Category lib mimics, however, did not significantly change adenovirus progeny production in PC-3 cells. None of the mimics in these Categories increased adenovirus progeny production, which is not expected for molecules selected to increase death of host cells. In contrast, this could be expected for Category IV molecules. However, none of the four molecules in this Category showed a consistent increase in virus progeny production (not shown). We also studied the release of adenovirus progeny from host PC-3 cells into the culture medium. Figure 3B shows the relative amount released infectious virus compared to irrelevant miRNA mimic transfection controls. Results were highly variable between experiments, yielding high standard deviations.

Nevertheless, several trends were observed. Category I miRNA mimics increased release of progeny virus, consistent with their cell death promoting activity. Category III miRNA mimics did not change the amount of released progeny, despite their selective augmentation of adenovirus-induced cell death. Category Ila miRNA mimics, which clearly inhibited total progeny virus production, affected progeny virus release to a lesser extend. Category lib miRNA mimics exhibited variable effects on virus release, with four mimics increasing release and four mimics inhibiting release. Example 4.

Selection of miRNA mimics that are most useful to increase the anti-cancer effect of oncolytic adenoviruses. To select miRNA molecules that are most useful to increase the anticancer potency of oncolytic adenoviruses, the following criteria were used. First, miRNAs were selected that increase PC-3 cell kill (Example 1) without severely compromising adenovirus replication (Example 2) and progeny virus production (Example 3). This excluded Category Ila miRNAs and miRNA-155 from Category lib. Second, miRNAs that increased cell kill (Example 1) and appeared to cause a faster or more effective release of adenovirus progeny (Example 3) were considered of particular interest. A fast or effective release of progeny virus is a favorable attribute, because it could allow fast spread of a replicating adenovirus in a population of cancer cells, such as for example a solid tumor, resulting in a fast destruction of the population of cancer cells and, therefore, an effective anti-cancer treatment. All Category I miRNAs, most Category III miRNAs and several Category lib miRNAs (miR-9, miR-26b, miR-181a and miR-199a*) showed this characteristic. Finally, Category III miRNAs were considered of exceptional interest, because they did not exert any direct toxicity to the PC-3 cells (Example 1). They are thus expected to cause less toxicity to cells in which the oncolytic adenovirus is not replicating.

Interestingly, several members of the same miRNA family were shown to have very similar effects. miR-lOa and miR-lOb both selectively sensitized PC-3 cells for Ad5-induced cell death (Category III) without strong effects on virus replication, virus progeny production or virus progeny release. miR-517a and miR-517c both exhibited strong cytotoxicity to PC-3 cells and inhibited Ad5 replication and progeny virus production. miR-520d, e, and g also showed quite comparable effects, although miR-520g was classified in a different category because it modestly sensitized PC-3 cells for Ad5-induced cell death. The observation of similar effects of miRNA family members confirmed the validity of the assays and strengthened the conclusions drawn from the screens. In conclusion, we identified several miRNAs that appear useful to increase the anticancer potency of oncolytic adenoviruses, at least against PC- 3 prostate cancer cells and probably against prostate cancer cells in patients with prostate cancer and possibly against cancer cells in patients with other types of cancer.

Example 5.

Analysis of the combined anti-cancer effect of Ad 5 and selected miRNA mimics on four different human cancer cell lines.

To validate the results described in Examples 1-3 and the selection made in Example 4, selected miRNA mimics were tested for their combined effect with Ad5 infection on PC- 3 prostate cancer cells as in Example 1 and on three additional cancer cell lines representing cancers from different tissue origins, i.e., A549 non-small cell lung cancer cells, MDA-MB-231 breast cancer cells and U20S osteosarcoma cells. The cells were grown at 37°C and 5% C02 in a humidified incubator in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS and antibiotics. Cells were seeded in a

concentration of 5,000 cells per well in 96-well plates in 80 microliter DMEM with 10% FBS, but without antibiotics. Twenty-four hours (A549, MDA-MB- 231, and U20S) or 48 hours (PC-3) after seeding, the cells were transfected with 50 nM (first experiment on PC-3 cells only) or 25 nM (all other

experiments) miRNA mimics using DharmaFect (DF) transfection reagent (Dharmacon Thermo Scientific), essentially as described in example 1. The different DF transfection reagents used were: 0.1 microliter DF-1 per well for A549 and MDA-MB-231, 0.2 microliter DF-1 for U20S and 0.2 microliter DF-2 for PC-3. The following Dharmacon miRIDIAN miRNA mimics were used: hsa- miR-lOa (Cat no C-300549-03-0005), hsa-miR-26b (Cat no C-300501-07-0005), hsa-miR-30a-3p (Cat no C-300506-03-0005; miR-30a*), hsa-miR-150 (Cat no C- 300632-03-0005), hsa-miR-181a (Cat no C-300552-05-0005), hsa-miR-199a-3p (Cat no C-300535-05-0005), hsa-miR-324-3p (Cat no C-300705-05-0005), hsa- miR-383 (Cat no C-300692-03-0005), hsa-miR-432 (Cat no C-300759-03-0005), hsa-miR-517c (Cat no C-300832-03-0005), hsa-miR-520e (Cat no C-300772-03- 0005), and the irrelevant negative controls miRNA cel-miR-67 (negative ctrl #1, Cat no CN-001000-01-05) in the first experiment and cel-miR-239b

(negative ctrl #2, Cat no CN-002000-01-05) in the second experiment. It should be noted that the miR-324-3p mimic with Cat no C-300705-05-0005 is a different product than the one present in the miRIDIAN mimic library with Cat. No. CS-001000 used for the screens described in Examples 1 and 2. The mature sequence of the mimic used in the screens was

CCACUGCCCCAGGUGCUGCUGG according to Sanger miRBase v.8, whereas the new mimic has a mature sequence of ACUGCCCCAGGUGCUGCUGG according to miRBase v.10. The updated product has two (CC) bases removed from the 5' end, which shifts the seed sequence. To check if the update changed the effects of the mimic on cell viability and sensitivity to adenovirus infection, the old and new mimic were compared side-by-side on PC-3 and A549 cells. The new miR-324-3p mimic showed a much stronger effect than the old version, specifically on PC-3 cells (data not shown). Apart from sensitizing cells to cell death in the presence of Ad5, it also exerted moderate cytotoxicity on its own. The new product would therefore be classified as a Category II miRNA mimic. Also hsa-miR-199a-3p (new nomenclature for hsa-miR-199a*) had a modified mature sequence, i.e., ACAGUAGUCUGCACAUUGGUUA for the new product versus UACAGUAGUCUGCACAUUGGUU for the mimic in the library (U removed at the 5' side and A added at the 3'side). We did not check the effect of this modification on cell viability and sensitivity to adenovirus infection in a side-by- side comparison.

Cells were infected with Ad5 at 100 IU/cell, 24 hours after miRNA mimic transfection. Cell viability was determined using CTB reagent on at least two subsequent days between 3 and 6 days post infection (depending on the cell line). Thirty microliter CTB reagent was added to the culture medium and plates were placed at 37°C in a 5% C02 incubator for three hours. After three hours, 50 microliter 3% SDS was added and cell viability was determined by measuring fluorescence at 540 nm excitation and 590 nm emission wave lengths using a Tecan Infinite F200 microplate reader. All cell lines and miRNA mimics were tested in two independent experiments in triplicate, and the post infection time points yielding the most discriminative results were analysed with GraphPad Prism 5 for Windows (Version 5.03, December 10, 2009). In this analysis, cell viability after miRNA mimic transfection and Ad5 infection was normalized to the "No miRNA" condition (Ad5 infection only). The results are given in Figure 4. Overall, the most dramatic effects were observed in PC- 3 cells, which was to be expected since the screens had been done in this cell line. On these cells, clear anti-cancer effects were observed with miR-324-3p, miR-383, miR-432, miR-517c, and miR-520e. However, several miRNAs showed effects on other cell lines as well, suggesting that their combined anti- cancer effect together with Ad5 replication is not cell line or cancer type specific. For example, miR-199a-3p and miR-383 increased cell kill in MDA-MB-231 cells as well; and miR-517c increased cell kill in U20S cells as well. The data do not discriminate between direct effects of miRNA mimics on cell viability and effects of the miRNA mimics on Ad5-induced cell death. In stead, they show the combined anti-cancer efficacy of miRNA mimic expression and oncolytic adenovirus treatment. In general, however, miRNA mimics showing the most profound effects in the presence of Ad5 also exhibited direct toxicity. Example 6.

Selective sensitization of cancer cells to adenovirus oncolytic propagation by miRNA mimic transfection. To test whether (apart from their own toxicity), the tested miRNA mimics specifically enhanced adenovirus oncolytic propagation in the four cancer cell lines, a next set of experiments was done. Here, cells were infected with a serial dilution of Ad5, ranging from 0.001 - 1000 IU/cell, in steps of 3 and 3.33 (0.001 - 0.003 - 0.01 - 0.03 - 0.1 - etc), 24 hours after mimic

transfection essentially as described in Example 1. Six days post infection, cell viability was determined, using CTB reagent as described in Example 1. All cell lines and miRNA mimics were tested in three independent experiments. In the first experiment, mimics were transfected at 50nM concentration and cel- miR-67 was included as irrelevant miRNA control; in the second experiment, the mimic concentration was decreased to 25nM; and in the third experiment, which was also done with 25nM mimics, the irrelevant miRNA control was replaced by cel-miR-239b. In some experiments, the irrelevant miRNA, serving as a negative control, also had considerable effect on cell viability in

combination with Ad5 (i.e., experiment 1 on A549 cells, experiment 2 on MDA- MB-231 cells, and experiment 3 on U20S cells; data not shown). These data were therefore excluded from further analysis and interpretation. Data analysis was done with GraphPad Prism 5, where the highest value within each data set (i.e., each tested mimic on each cell line) was set at 100% and the lowest at 0%, to exclude the toxic effects of the miRNA mimic alone.

Subsequently, a sigmoidal dose-response curve (variable slope) was fitted to the data points in order to obtain cell viability curves and determine the EC- 50 value (i.e., the dose Ad5 causing a 50% cell killing effect).

Figure 5 shows examples of dose-response curves obtained on the four cell lines. Reciprocal l/EC-50 values, which are quantitative measures for oncolytic potency, were normalized to the "No miRNA" condition, to yield relative oncolytic potency factors. Figure 6 shows the normalized reciprocal l/EC-50 values calculated from all curves, where the values indicate the fold increase in oncolytic potency of Ad5 due to the addition of the specific miRNA mimic. As can be appreciated from Figures 5 and 6, several miRNA mimics clearly increased the oncolytic potency of Ad5, irrespective of their direct effects on cell viability. A shift of the dose-response curve to the left compared to the no miRNA control in Figure 5 indicates enhanced oncolytic replication. Whereas Ad5 alone killed 50% of the cells after infection at an MOI usually around 10, Ad5 killed 50% of cells transfected with some miRNA mimics that survived the mimic transfection itself, at an MOI as low as 0.1. Note that because miRNA mimics often exerted direct cell killing effect as well, the actual cell survival was less than 50% at MOI causing EC-50. A quantification of the enhancement in individual experiments can be read from Figure 6. The desired potency- enhancing effect on Ad5 replication in cancer cells was observed most clearly for miR-lOa, miR-26b, miR-199a-3p, miR-324-3p, miR- 383, miR-517c, and miR-520e. The miRNA-517c mimic exhibited this desired effect on all four tested cancer cell lines; the other miRNA mimics augmented Ad5 oncolytic propagation in only one or two cancer cell lines. For most of these miRNAs, a beneficial effect on oncolytic adenovirus treatment was predicted based on their classification shown in Table 1 and their effect on viable virus production and release shown in Figure 3. The exception was miR- 517c, which despite its inhibiting effect on adenovirus replication and progeny production caused a broad sensitization to adenovirus oncolytic propagation in cancer cell lines of various tissue origins. Apparently, its cell death promoting activity compensated for its adenovirus inhibitory activity. It can be

appreciated from Figure 3 that, although the total adenovirus progeny production was reduced by miR-517c, the amount of released progeny virus recovered from the culture medium was not reduced. Thus, although less virus was produced, the effective induction of cell death made a considerably larger fraction of virus progeny available for spread and further propagation in the population of cancer cells.

Example 7.

Construction of oncolytic adenoviruses expressing miRNA mimic precursor molecules.

Endogenous expression of miRNAs occurs via a sequence of processing steps, ultimately generating a 21-23 nucleotide mature miRNA duplex, of which the strand containing the mature sequence is incorporated into the RISC silencing complex, leading to post translational gene silencing. First, a pri-miRNA is transcribed from the genome, comprising of the miRNA hairpin, flanked by stretches of single stranded RNA. This pri-miRNA is recognized and processed by the class 3 RNase III Drosha, leading to cleavage of the single stranded RNA stretches. This results in the pre-miRNA hairpin, which is transported out of the nucleus by Exportin-5. In the cytosol, Dicer cleaves off the loop of the hairpin, resulting in the mature miRNA duplex. The different miRNA precursor molecules, or precursor molecules for miRNA mimics, can be expressed from the genome of a replication competent adenovirus. Expression of pri-miRNA molecules can be driven by a Pol II promoter; Pol III promoters can be used to drive expression of pri-miRNA and pre-miRNA molecules and short hairpin molecules encoding miRNA mimics (shMimics). Several ways to construct these recombinant replication competent adenoviruses are described in WO2005/100576, which is

incorporated by reference herein. Depending on the choice of promoter to drive expression of the miRNA precursor molecule, the proper transcription initiation and termination signals should be used. One or more expression cassettes, each encoding a single miRNA precursor molecule can be inserted, or alternatively, multiple miRNAs can be expressed on a single polycistronic transcript, with the miRNA hairpin structures interspaced with a stretch of preferably 10-100 nucleotides can be used. Here, we follow the teachings of WO2005/100576 to construct conditionally replication competent adenoviruses expressing different miRNA (mimic) precursor molecules. By way of example, a single miRNA (mimic) precursor molecule expression cassette is inserted between the adenovirus E4 region and the right-hand ITR in a replication competent adenovirus with the Ε1Α-Δ24 mutation that confers tumor-selective replication.

First, a shuttle vector was made carrying a full length adenovirus genome flanked with Pad sites, comprising the ElAA24-mutation (Fueyo et al., Oncogene 19(2000):2-12) and the Gateway recombination destination cassette between the adenovirus E4 region and the right-hand ITR. To this end, the constructs pEndK/DEST-R and pAdA24E3 were used. pEndK/DEST-R was derived from the construct pEndK/Spel (generously provided by Dr. R. Alemany, Institut Catala d'Oncologia, Barcelona, Spain). pEndK/Spel was made by first digesting pTG3602 (Chartier et al., J. Virol, 70(1996):4805-4810) with Kpnl and religating the vector fragment comprising Ad5 map units 0-7 and 93-100 to create pEndK. Next, a unique Spel site was introduced into pEndK by changing Ad5 nucleotide 35813 from A to T by site directed mutagenesis to create pEndK/Spel. PEndK/Spel carries Pad restriction sites flanking the two Ad5 ITRs. pEndK/Spel was made compatible with the Gateway system by ligating the Gateway destination cassette rfa (Gateway Vector Conversion System; Invitrogen, Carlsbad, CA) as a blunt fragment into the Spel site (filled in with Klenow polymerase). A plasmid was selected that contained the Gateway destination cassette with the coding sequence of the ccdB gene on the adenovirus R strand and was designated pEndK/DEST-R. To obtain pAdA24E3, linear dsDNA was isolated from Ad5-A24E3 (Suzuki et al., Clin. Cancer Res. 8(2002) :3348-3359) virions and recombined with Kpnl- linearized pEndK/Spe (supra) in BJ5183 bacteria. Full length AdA24E3 DNA was released from pAdA24E3 by Pad digestion. This DNA was recombined in BJ5183 bacteria with Kpnl- digested pEndK/DEST-R to obtain pAdA24E3- DEST-R. pAdA24E3-DEST-R is propagated in the E.coli STBL2-DB3.1 strain, which contains a gyrase mutation that renders it resistant to the lethal effects of the CcdB protein thereby allowing propagation of plasmids carrying the ccdB gene in the DEST cassette.

Second, miRNA (mimic) precursor-encoding sequences are introduced into GATEWAY system (Invitrogen, Carlsbad, CA) entry clone pSHAG-1 (Paddison et al., Genes Dev. 16(2002)948-958; generously provided by Dr. G.J. Hannon, Cold Spring Harbor Laboratory, NY). pSHAG-1 contains a U6 promoter- driven expression cassette flanked by the Gateway attLl and attL2 recombination sites such that the expression cassette can be transported into destination plasmid vectors including pAdA24E3-DEST-R using the Gateway system. MicroRNA (mimic) precursor-encoding sequences are introduced by ligation of pSHAG-1 digested with BseRI and BamHI with two annealed synthetic oligonucleotides with compatible overhanging DNA sequences.

To construct a pSHAG-shMimic plasmid, the first of the two oligonucleotides should be designed to contain in the 5' to 3' order: a first stretch of nucleotides comprising of the mature sequence of the miRNA, as taken from miRBase (www.mirbase.org), a loop sequence, a second stretch of nucleotides of the same length and of reverse complementary sequence to the first stretch of nucleotides, and a stretch of at least 4 thymidines (termination sequence for Pol III). To preserve the correct seed sequence of the mature miRNA (nucleotides 2-8), the first 5' nucleotide of the mature sequence is replaced by a G, which is the +1 start nucleotide of the U6 promoter driven Pol III transcript and the terminating nucleotide of pSHAG-1 when digested with BseRI. This G will wobble base-pair with the first U of the Pol III termination sequence, thereby favoring incorporation of the mature miRNA oligonucleotide into RISC due to strand biasing. Alternatively, the location of the mature sequence can be placed on the 3' strand of the stem, which omits the necessity to remove the first 5' nucleotide of the mature sequence. However, since Dicer might process the shMimic at different positions, leading to cleavage into the seed sequence, placement of the mature sequence at the 5' strand of the stem is preferred. The second oligonucleotide should be reverse complementary to the first oligonucleotide. Furthermore, when annealed the then double- stranded oligonucleotides should form overhanging sites compatible with BseRI and BamHI restriction sites.

To construct a pSHAG-pre-miRNA plasmid, the first of the two oligonucleotides should be designed to contain in the 5' to 3' order: the 5' to 3' stem-loop-stem sequence of the miRNA of choice as given by miRBase, comprising the mature sequence, the endogenous miRNA loop and the complementary stem strand, including the unpaired residues; and a stretch of at least 4 thymidines. The second oligonucleotide should be reverse

complementary to the first oligonucleotide; and, when annealed the then double-stranded oligonucleotides should form overhanging sites compatible with BseRI and BamHI restriction sites.

To construct a pSHAG-pri-miRNA plasmid, the sequence of the desired miRNA (according to miRBase), plus 100-150 5'and 3' flanking sequences (according to the human Genome Reference Consortium, build 37, GRCh37) is synthesized containing the following modifications. A Hhal restriction site is introduced at the 5' end of the sequence, creating BseRI compatible ends after digestion; and a stretch of at least 4 thymidines and a BamHI restriction site are introduced at the 3' end of the sequence. After digestion with Hhal and BamHI, the pri-miRNA sequence is released, containing compatible DNA ends for the BseRI and BamHI restricted pSHAG- 1. Alternatively, the desired miRNA and flanking sequences can be obtained by a PCR reaction on genomic DNA using a forward primer with an added 5'- Hhal sequence and a reverse primer with an added 5'-BamHI-AAAA sequence. For efficient digestion of the PCR product with the given restriction enzymes, preferably at least 3 additional nucleotides should be added adjacent to the restriction sites, or the PCR product should be ligated into a PCR cloning vector (e.g. pDrive, Qiagen). Digestion of a pri-miRNA containing plasmid or PCR product with Hhal and BamHI, followed by ligation into pSHAG-1 digested with BseRI and BamHI results in the formation of pSHAG-pri- miRNA.

Third, the miRNA (mimic) precursor molecule expression cassettes obtained in the second step are transferred from pSHAG constructs into pAdA24E3-DEST-R via an LR GATEWAY in vitro recombination reaction using the GATEWAY LR Clonase enzyme mix (Invitrogen) according to manufacturer's protocol, to create pAdA24E3-pri-miRNA, pAdA24E3-pre- miRNA, or pAdA24E3-shmimic oncolytic adenovirus constructs.

Finally, full-length AdA24 oncolytic adenovirus genomes with inserted miRNA (mimic) precursor molecule expression cassettes are released from pAdA24E3-pri-miRNA, pAdA24E3-pre-miRNA, or pAdA24E3-shmimic constructs by Pad digestion and transfected using lipofectamine reagent in 911 cells or A549 cells to obtain the different AdA24E3-derived miRNA (mimic) precursor molecule-expressing replication competent adenoviruses, which are further propagated on A549 cells according to standard cell culture and virology methods known in the art. The Ε1Δ24 deletion and the U6-shRNA insertion and orientation are confirmed by PCR on the final products, miRNA (mimic) precursor molecule sequences are confirmed by sequencing and functional virus titers are determined by limiting- dilution titration according to standard techniques. Example 8.

Construction of oncolytic adenoviruses expressing miRNA inhibitor precursor molecules. Similar to the methods described in Example 7 to construct miRNA mimic precursor molecule expression cassettes, which are transferred to the oncolytic adenovirus genome, hairpins can be designed to function as miRNA inhibitors, where the end product of processing is a single stranded RNA that is fully complementary to a specific miRNA target, thus sequestering this miRNA away from its target mRNA. The utility of this miRNA inhibitor design has already been confirmed in other vector systems. It is, e.g., already exploited commercially under the name "miRZip anti-microRNAs", which are stably incorporated into the target cell genome through a lentiviral vector (System Biosciences; http://www.systembio.com/microrna-research/microrna- knockdown/mirzip/). Furthermore, the company GeneCopoeia offers a system where a hairpin is expressed, containing two antisense stretches, allowing the sequestration of two mature target miRNAs, under the name "miArrest" (http://www.genecopoeia.com/product/mirna/inhibitor/).

An miRNA inhibitor sequence can be synthesized as double- stranded oligonucleotide that forms overhanging sites compatible with BseRI and

BamHI restriction sites. This is then cloned into pSHAG-1 and subsequently transported into a destination plasmid vector such as pAdA24E3-DEST-R using the Gateway system as decribed in Example 7.

Alternatively, methods have been described where miRNAs are blocked by so-called miRNA sponges (Ebert et al., Nat. Meth. 4, 721-726, 2007). Using this methodology, Pol II- or Pol III- driven transcripts can be designed which contain multiple copies of a miRNA binding site of interest, thereby serving as decoy RNAs for the desired miRNA. This system has already been shown to work in vivo using lentiviral vectors (Gentner et al., Nat. Meth. 6, 63-66, 2009). Pol-III driven miRNA inhibitor expression cassettes can be cloned into the adenovirus genome essentially as described in Example 7; Pol-II driven miRNA inhibitor expression cassettes can be cloned into the adenovirus genome using methods known in the art. Example 9.

Functional comparison of miRNA mimics expressed from the genome of an oncolytic adenovirus in different miRNA mimic precursor formats.

To demonstrate that miRNA expression from the genome of an oncolytic adenovirus leads to functional silencing of a target mRNA, expression cassettes of several different miRNA mimic precursor formats of miR-1 were generated and transferred to the virus genome. Mir-1 was chosen since it is reported not to be expressed in A549 cells and one of its targets is known to be PTK-9 (Nasser et al., J. Biol. Chem. 48, 33394-405, 208; Lim et al., Nature 7027, 769-73, 2005). Previously, it was shown that when a short hairpin siRNA precursor against firefly luciferase was expressed from an oncolytic adenovirus genome, significant silencing of a luciferase gene expressed in cancer cells was observed when these cells were infected with this virus (Carette et al., Cancer Res. 64, 2663-7, 2004). Therefore, the first miR-1 format generated was a short hairpin miR-1 (sh- miR-1) containing the mature miR-1 sequence; an 8 nucleotide loop identical to the one reported by Carette et al.; and a sequence completely complementary to the mature miR-1 sequence to form the stem. In the second format, miR-1 was expressed as a premature miRNA (pre-miR-1), containing the endogenous loop sequence and stem mismatches of human miR- 1 (as presented in miRBase v.16, Sept. 2010). In the third format, miR-1 was expressed as a primary miRNA (pri-miR-1), where 120 nucleotides of flanking sequence (according to the human Genome Reference Consortium build 37 (GRCh37), Feb 2009) were added up- and downstream of the human miR-1 hairpin. The three different miR-1 formats were synthesized with BamHI and BseRI restriction sites at the 5' and 3' end, respectively, and cloned in a plasmid at Geneart (Regensburg, Germany). The miRNA fragments were released from the plasmid using BamHI and BseRI; and subsequently ligated into pSHAG-1 plasmid (Paddison et al., Genes Dev. 16, 948-58, 2002), digested with the same restriction enzymes. This generated U6 promoter driven miR-1 expression cassettes with flanking Gateway attL sites, through which the expression cassettes could be transferred to a Gateway attR sites-containing oncolytic adenovirus genome (pAdA24E3-DEST-R; Example 7), using the Gateway LR reaction (Invitrogen, Carlsbad, CA). miR-1 expressing viruses, as well as a control virus containing an empty expression cassette, were generated by transfection of linearized virus genomes into A549 cells. To determine mature miR-1 expression and PTK-9 mRNA silencing, A549 cells were seeded 60,000 cells per well in 24-well plates and infected the next day with 100 IU/cell of the four different viruses. Six to eight hours after infection, residual virus was removed by medium refreshment and exactly 24 and 48 hours after infection, total RNA was isolated from the cells using TRIzol, after which mature miR-1 and PTK-9 mRNA levels were determined with RT-qPCR using TaqMan MicroRNA assays (Applied Biosystems) and the 1st strand cDNA synthesis kit for RT-PCR + SYBR Green (Roche), respectively. Figure 9 shows the relative expression as compared to uninfected cells of mature miR-1 (Figure 9A) and PTK-9 (Figure 9B) 24 and 48 hours after infection with viruses expressing the indicated miR-1 formats. As can be seen in Figure 9A, miR-1 is expressed at least 1,000 fold higher compared to endogenous expression levels when the cells are infected with any of the three different precursor format-expressing adenoviruses. Infection with the control virus containing an empty cassette did not increase miR-1 expression, indicating that the elevated miR-1 expression was induced by the exogenous expression cassettes and not by induction of endogenous expression. The highest miR-1 expression was achieved using the pri-miR-1 format, and this level was already reached within 24 hours. Figure 9B shows the relative PTK-9 mRNA levels as compared to uninfected cells. As can be seen, significant silencing of PTK-9 mRNA could be obtained only after expression of miR-1 in the pri-miR- 1 format. Twenty-four hours after infection PTK-9 expression levels were decreased by ~40%, while after 48 hours this was ~70%. From these data it can be concluded that expression of a miRNA from the adenovirus genome in a pri- miRNA format yields the highest mature miRNA expression levels and the most efficient silencing of a target mRNA. This is thus the most preferred format for expressing miRNAs from the adenovirus genome.

Example 10.

Identification of miRNAs that augment the propagation of oncolytic adenovirus in PC-3 prostate cancer cells.

Several miRNA mimics from Categories lib and III were analyzed for their capacity to support propagation (i.e., infection, replication, release and re- infection) of oncolytic adenovirus in PC-3 prostate cancer cells. These experiments were done using the oncolytic adenovirus AdDelta24-CMV-GFP (Idema et al., J. Gene Med. 12, 564-571, 2010). This virus expresses GFP in infected cells driven by the constitutive CMV promoter, thus allowing monitoring of viral spread in a population of cells by detecting GFP

fluorescence. PC-3 cells were seeded 5,000 cells per well in 96-well plates and transfected 2 days later with 25 nM miRNA mimic using 0.2 microliter

DharmaFect-2 transfection reagent in a total volume of 100 microliter per well as above, or mock treated with culture medium without miRNA and

transfection reagent. The following Dharmacon miRIDIAN miRNA mimics were used: hsa-miR-lOa (Cat no C-300549-03-0005), hsa-miR-26b (Cat no C- 300501-07-0005), hsa-miR-150 (Cat no C-300632-03-0005), hsa-miR-181a (Cat no C-300552-05-0005), hsa-miR-324-3p (Custom product no C-300705-03-000 with same sequence as the mimic present in the miRIDIAN mimic library with Cat. No. CS-001000 used for the screens), hsa-miR-383 (Cat no C-300692-03- 0005), hsa-miR-432 (Cat no C-300759-03-0005), hsa-miR-454-3p (Cat no C- 301004-03-0005), and the irrelevant negative control miRNA cel-miR-239b (negative ctrl #2, Cat no CN-002000-01-05). The next day, 5,000 infectious units AdDelta24-CMV-GFP were added in a volume of 5 microliter. One day post infection, 100 microliter fresh culture medium was added to all wells. On days 1, 4, 5 and 7 post infection, the number of GFP-expressing and thus AdDelta24-CMV-GFP infected cells was counted using an Acumen eX3 microplate laser scanning cytometer (TTP Labtech, Melbourn, UK). One day after infection, 63+/-36 SD green fluorescent cells were detected per well. The following days, the number of adenovirus-infected and GFP-expressing cells had increased significantly, reaching a maximum after 5 days. Figure 10 shows pseudo-images of the cultures and the number of GFP-positive cells detected in the cultures on day 5 post infection. As can be seen, upon transfection of miR-lOa, miR-181a and miR-454-3p, similar numbers of GFP- positive cells were found as in control cultures without miRNA or transfected with irrelevant control miRNA, indicating that these miRNAs had little or no effect on oncolytic adenovirus propagation. MiRNA-432 had a slight inhibitory effect on adenovirus propagation. In contrast, transfection of miR-26b, miR- 150, miR-324-3p and miR-383 increased the number of GFP-positive cells 5 days post infection, indicating that these miRNAs stimulated oncolytic adenovirus propagation.

Table 1. Summary of miRNA mimic and inhibitor effects on Ad5 replication and cell death in PC- 3 prostate cancer cells (see description in Examples 1 and 2). When in this application a reference is made to a miRNA, a miRNA mimic and/or a miRNA inhibitor of this table, the reference includes a reference to the sequence of the miRNA as depicted in figure 7, or a reference to the reverse complement of a sequence as depicted in figure 8.

Claims

1. A method for determining whether a miRNA, a miRNA mimic or a miRNA inhibitor increases adenovirus replication in a cell, and/or cell death of an adenovirus infected cell, preferably adenovirus induced cell death of a cell, said method comprising providing a first cell that is permissive for adenovirus replication with an adenovirus and with a nucleic acid molecule comprising and/or encoding said miRNA, miRNA mimic or miRNA inhibitor, said method further comprising culturing said first cell and determining whether replication of said adenovirus in said first cell and/or whether cell death, preferably adenovirus induced cell death of said first cell is increased.

2. A method according to claim 1, further comprising determining whether said nucleic acid molecule comprising and/or encoding said miRNA, miRNA mimic or miRNA inhibitor increases adenovirus replication in a second cell, and/or cell death of an adenovirus infected second cell, preferably adenovirus induced cell death of a second cell that is permissive for adenovirus replication and wherein said second cell is of a different tissue origin than said first cell.

3. A method for providing a cancer cell with an anti-cancer treatment said method comprising providing said cancer cell with an adenovirus and with a nucleic acid molecule comprising and/or encoding a miRNA, a miRNA mimic or a miRNA inhibitor that increases replication of said adenovirus in said cancer cell, and/or death of said cancer cell, preferably cell death induced by said adenovirus.

4. A method according to claim 3, wherein said miRNA, miRNA mimic or miRNA inhibitor was identified with a method according to claim 1 or claim 2.

5. A method according to claim 3 or claim 4, wherein said adenovirus encodes said miRNA, miRNA mimic or miRNA inhibitor.

6. A method for sensitizing a cancer cell for adenovirus replication and/or cell death, preferably adenovirus induced cell death, said method comprising providing said cancer cell with a nucleic acid molecule comprising and/or encoding a miRNA, a miRNA mimic or a miRNA inhibitor identified with a method according to claim 1 or claim 2.

7. A kit of parts for the treatment of cancer, said kit comprising an adenovirus and a nucleic acid molecule comprising and/or encoding a miRNA, a miRNA mimic or a miRNA inhibitor that increases replication of said adenovirus in a cancer cell, and/or cell death of an adenovirus infected cancer cell, preferably cancer cell death induced by said adenovirus.

8. A kit of parts according to claim 7, wherein said miRNA, miRNA mimic or miRNA inhibitor is identified with a method according to claim 1 or claim 2.

9. A kit of parts according to claim 7 or claim 8, wherein said miRNA, miRNA mimic or miRNA inhibitor is a miRNA, a miRNA mimic or mRNA inhibitor of category III, category lib, category IV, category I, category Ila and/or category V of table 1.

10. A kit of parts according to claim 9, wherein said miRNA or miRNA mimic is a miRNA or miRNA mimic of category III.

11. A viral vector comprising an expression cassette comprising an expressible nucleic acid molecule that codes for a miRNA, a miRNA mimic or a miRNA inhibitor that increases adenovirus replication in a cell, and/or cell death of an adenovirus infected cell, preferably adenovirus induced cell death of a cell infected with said adenovirus, preferably wherein said viral vector is an adenovirus vector.

12. A viral vector according to claim 11, wherein said miRNA, miRNA mimic or miRNA inhibitor is a miRNA, a miRNA mimic or mRNA inhibitor of category III, category lib, category IV, category I, category Ila and/or category V of table I.

13. A viral vector according to claim 11 or claim 12, wherein said expression cassette comprises a polymerase III promoter.

14. A viral vector according to any one of claims 11-13, wherein said expression cassette encodes a precursor RNA comprising said miRNA, miRNA mimic or mRNA inhibitor.

15. A viral vector according to claim 14, wherein said precursor RNA encodes a pre-miRNA comprising said miRNA and a miRNA stem-loop sequence, wherein said pre-miRNA is flanked by at least 30 single-stranded RNA nucleotides at the 5' and 3' sides of the miRNA stem-loop sequence.

16. A viral vector according to any one of claims 11-15, for use in the treatment of cancer.