Compositions and methods for immunostimulatory rna oligonucleotides   

Abstract: The present invention provides 4-nucleotide (4mer) RNA motifs that confer immunostimulatory activity, in particular, IFN-α-inducing activity to a RNA oligonucleotide. The present invention also provides RNA oligonucleotides, including siRNA, with high or low immunostimulatory activity. The present invention further provides the use of the RNA oligonucleotides of the invention for therapeutic purposes.

FIELD OF THE INVENTION

The present invention relates to the field of immunotherapy and drug discovery. The present invention provides a method for determining the immunostimulatory activity of a RNA oligonucleotide. The present invention also provides a method for predicting the immunostimulatory activity of a RNA oligonucleotide. The present invention further provides a method for preparing RNA oligonucleotides with high or low immunostimulatory activity. Moreover, the present invention provides RNA oligonucleotides with immunostimulatory activity and the therapeutic uses thereof. In addition, the present invention provides RNA oligonucleotides with gene silencing activity and with either high or low immunostimulatory activity, the methods of their preparation, and their therapeutic uses.

BACKGROUND OF THE INVENTION

The vertebrate immune system established different ways to detect invading pathogens based on certain characteristics of their microbial nucleic acids. Detection of microbial nucleic acids alerts the immune system to mount the appropriate type of immune response that is required for the defense against the respective type of pathogen detected. Detection of viral nucleic acids leads to the production of type I interferon (IFN), the key cytokine for anti-viral defence. An understanding of how nucleic acids interact with the vertebrate immune system is important for developing different nucleic acid-based therapeutic strategies for the immunotherapy of diseases (Rothenfusser S et al. 2003, Curr Opin Mol Ther 5:98-106) and for developing gene-specific therapeutic agents (Tuschl T et al. 2002, Mol Interv 2: 158-167).

For the recognition of long dsRNA, two detection modes are known, the serine threonine kinase PKR (Williams B R, 2001, Sci Signal Transduction Knowledge Environment 89: RE2; Meurs E F et al. 1992, J Virol 66: 5805-5814; Katze M G et al. 1991, Mol Cell Biol 11: 5497-5505) and Toll-like receptor (TLR) 3 (Alexopoulou L et al. 2001, Nature 413: 732-738). Whereas PKR is located in the cytosol, TLR3 is present in the endosomal compartment (Matsumoto M et al. 2003, J Immunol 171: 3154-3162). TLR3 is a member of the Toll-like receptor family that has evolved to detect pathogen-specific molecules (Takeda K et al. 2003, Annu Rev Immunol 21: 335-376).

A second characteristic feature of viral nucleic acids used by the immune system to recognize viral infection are CpG motifs found in viral DNA, which are detected via TLR9 (Lund J et al. 2003, J Exp Med 198: 513-520; Krug A et al. 2004, Blood 103: 1433-1437). CpG motifs are unmethylated CG dinucleotides with certain flanking bases. The frequency of CpG motifs is suppressed in vertebrates, allowing the vertebrate immune system to detect microbial DNA based on such CpG motifs (Krieg A M et al. 1995, Nature 374: 546-549; Bauer S et al. 2001, Proc Natl Acad Sci USA 98: 9237-9242; Wagner H et al. 2002, Curr Opin Microbiol 5: 62-69). Like TLR3, TLR9 is located in the endosomal compartment where it directly binds to CpG motifs (Latz E et al. 2004, Nat Immunol 5: 190-198).

In addition to long dsRNA and CpG DNA, two recent publications suggest a third mechanism by which viral nucleic acids are recognized. These studies demonstrate that single-stranded RNA (ssRNA) of ssRNA viruses is detected via TLR7 (mouse and human) and TLR8 (only human) (Diebold S S et al. 2004, Science 303: 1529-1531; Heil F et al. 2004, Science 303: 1526-1529). Guanine analogues have been identified earlier as specific ligands for TLR7 and TLR8 (Lee J et al. 2003, Proc Natl Acad Sci USA 100: 6646-6651; Heil F et al. 2003, Eur J Immunol 33: 2987-2997). Like TLR9 (receptor for CpG DNA) (Latz E et al. 2004, Nat Immunol 5: 190-198), TLR7 and TLR8 are located in the endosomal membrane (Heil F et al. 2003, Eur J Immunol 33: 2987-2997).

Detection of viral nucleic acids leads to the production of type I IFN (IFN-α and IFN-β). The major producer of type I IFN in humans is the plasmacytoid dendritic cell (PDC, also called interferon producing cell, IPC). The plasmacytoid dendritic cell (PDC) is a highly specialized subset of dendritic cells that is thought to function as a sentinel for viral infection and is responsible for the vast amount of type I IFN during viral infection (Asselin-Paturel C et al. 2001, Nat Immunol 2: 1144-1150). There is increasing evidence that PDC preferentially use nucleic acid-based molecular patterns to detect viral infection. TLR expression of human and mouse PDC is limited to TLR7 and TLR9 (Krug A et al. 2001, Eur J Immunol 31: 3026-3037; Hornung V et al. 2002, J Immunol 168: 4531-4537; Edwards A D et al. 2003, Eur J Immunol 33: 827-833).

IFN-α was the first type of interferon to be identified and commercialized; it is widely used clinically in the treatment of a variety of tumors (e.g., hairy cell leukemia, cutaneous T cell leukemia, chronic myeloid leukemia, non-Hodgkin\'s lymphoma, AIDS-related Kaposi\'s sarcoma, malignant melanoma, multiple myeloma, renal cell carcinoma, bladder cell carcinoma, colon carcinoma, cervical dysplasia) and viral diseases (e.g., chronic hepatitis B, chronic hepatitis C). IFN-α products that are currently in clinical use include the recombinant protein and the highly purified natural protein, both of which have high production costs. Therefore, there is a need for more economical ways of providing IFN-α to patients in need. Furthermore, IFN-α is currently administrated systematically and causes a broad spectrum of side effects (e.g. fatigue, flu-like symptoms, diarrhea). Most alarmingly, IFN-α causes a decrease in bone marrow function which leads to increased susceptibility to life-threatening infections, anemia and bleeding problems. Therefore, there is a need for ways of providing IFN-α in a more localized (i.e., target-specific) matter to reduce the occurrence of side effects.

In addition to inducing an anti-viral interferon response, dsRNA also induces post-transcription gene silencing, a highly conserved anti-viral mechanism known as RNA interference (RNAi). Briefly, the RNA III Dicer enzyme processes dsRNA into short interfering RNA (siRNA) of approximately 22 nucleotides. The antisense strand of the siRNA binds a target mRNA via base pairing and serves as a guide sequence to induce cleavage of the target mRNA by an RNA-induced silencing complex RISC. dsRNA has been an extremely powerful tool in studying gene functions in C. elegence and Drosophila via gene silencing. However, its use in mammalian cells has been limited because the interferon response it elicits is detrimental to most mammalian cells.

Subsequently, it was found that siRNA was also capable of inducing RNAi, causing degradation of the target mRNA in a sequence-specific manner and it was thought to be short enough to bypass dsRNA-induced nonspecific effects in mammalian cells (Elbashri S M et al. 2001, Nature 411:494-498). Since then, siRNA has been widely used as a gene silencing tool in deciphering mammalian gene functions in research and drug discovery, and there has been great interest in its potential in therapeutic applications.

siRNA can be used to reduce or even abolish the expression of disease/disorder-related genes for preventing or treating diseases caused by the expression or overexpression of the disease-related genes. Such diseases include, but are not limited to, infections, metabolic diseases, autoimmune diseases and cancer. However, concern has been raised recently about the potential for siRNA to activate immune responses which may be undesirable for certain indications and thus limit the use of siRNA as a gene silencing agent for therapeutic purposes (Sioud M et al. 2003, Biochem. Biophys. Res. Commun. 312:1220-1225). Therefore, there is a need for methods for predicting the potential of a given siRNA to induce an interferon response and for methods for designing and preparing siRNAs for gene silencing which are devoid of unwanted immunostimulatory activities.

On the other hand, for certain therapeutic applications, for example, the prevention or treatment of cancer and viral infections, immunostimulatory activity may be desirable as an additional functional activity of the siRNA.

In an effort to apply siRNA for the specific downregulation of TLR9 in PDC in our previous publication (Hornung V et al. 2005, Nat Med 11: 263-270), we made the surprising observation that, despite the inability of PDC to detect long dsRNA, certain siRNA sequences were potent in vitro inducers of IFN-α in PDC. We found that i) short interfering RNA (siRNA) induces IFN-α in human plasmacytoid dendritic cells when transfected with cationic lipids, ii) this activity of siRNA is sequence-dependent but independent of the G or U content of the siRNA, iii) the immunostimuatory activity of siRNA and the antisense activity are two independent functional activities of siRNA, iv) the immune recognition of siRNA occurs on the single strand level, v) siRNAs containing the 9mer sequence motif 5″-GUCCUUCAA-3″ show potent immunostimulatory activity, and vi) such siRNAs induce systemic immune responses in mice, and vii) the induction of immune responses by siRNA requires the presence of TLR7 in mice. Our findings suggest that the 9mer sequence motif 5″-GUCCUUCAA-3″ may be a ligand for TLR7.

The natural ligand for TLR7 has not been well defined to date. Guanine analogues have been identified earlier as specific ligands for TLR7 and TLR8 (Lee J et al. 2003, Proc Natl Acad Sci USA 100: 6646-6651; Heil F et al. 2003, Eur J Immunol 33: 2987-2997), whereas guanosine ribonucleoside or a derivative thereof has been identified as TLR7 ligand in WO03086280.

It is an object of the present invention to identify RNA oligonucleotide motifs for stimulating an immune response, in particular, IFN-α induction. It is also an object of the present invention to identify ligands for activating TLR7 and TLR8. It is another object of the present invention to develop a method for determining the immunostimulatory activity, in particular, the IFN-α-inducing activity, of a RNA oligonucleotide. It is yet another object of the present invention to develop a method for predicting the immunostimulatory activity, in particular, IFN-α-inducing activity, of a RNA oligonucleotide. It is a further object of the invention to develop a method for designing and preparing RNA oligonucleotide having or lacking immunostimulatory activity, in particular, IFN-α-inducing activity. It is also an object of the invention to provide RNA oligonucleotides having high immunostimulatory activity which can be used to induce an immune response, in particular, IFN-α production, in patients in need thereof. It is yet another object of the present invention to provide siRNA molecules that either have or lack immunostimulatory activity which can be used to treat disorders caused by the expression or overexpression of disorder-related genes.

SUMMARY

The present invention provides a method for determining the immunostimulatory activity of a RNA oligonucleotide, a method for predicting the immunostimulatory activity of a RNA oligonucleotide, a method for preparing a RNA oligonucleotide with high or low immunostimulatory activity, and a method for preparing a RNA oligoncleotide with gene silencing activity and with high or low immunostimulatory activity.

The present application also provides an in vitro method for inducing IFN-α production from a mammalian cell, and an in vitro method for activating a dendritic cell.

The present invention further provides a RNA oligonucleotide with immunostimulatory activity, a RNA oligonucleotide with gene silencing activity and with high or low immunostimulatory activity, and the therapeutic uses thereof.

In addition, the present invention provides a pharmaceutical composition comprising one or more of the RNA oligonucleotides of the invention.

FIG. 1: PBMC of three individual donors were isolated and stimulated with ssRNA oligonucleotides 9.2sense (5′-AGCUUAACCUGUCCUUCAA-3′) and 9.2antisense (5′-UUGAAGGACAGGUUAAGCU-3′) that were complexed with either Lipofectamine, poly-L-arginine, poly-L-histidine or poly-L-lysine in duplicates. 24 hours after stimulation supernatants were harvested and IFN-α was assessed by ELISA. Data are presented as mean values±SEM.

FIG. 2: PBMC of three different healthy donors were isolated and stimulated with poly-L-arginine complexed ssRNA oligonucleotides in duplicates. 44 hours after stimulation IFN-a production was assessed in supernatant via ELISA. For all tested ssRNA oligonucleotides, the mean values of the measured duplicates were normalized to the positive control ssRNA oligonucleotide 9.2sense (5′-AGCUUAACCUGUCCUUCAA) by dividing the mean value of tested oligonucleotide by the mean value of 9.2sense (=100%). Data from three different donors were summarized and are presented as mean values±SEM.

FIG. 3: PBMC of six different healthy donors were isolated and stimulated with poly-L-arginine complexed ssRNA oligonucleotides in duplicates. 44 hours after stimulation IFN-a production was assessed in supernatant via ELISA. For all tested ssRNA oligonucleotides, the mean values of the measured duplicates were normalized to the positive control ssRNA oligonucleotide 9.2sense (5′-AGCUUAACCUGUCCUUCAA) by dividing the mean value of tested oligonucleotide by the mean value of 9.2sense (=IFN-a index of a given oligonucleotide). Next, all individual IFN-a indices were adjusted to the mean value of all IFN-a indices by subtracting the mean value of all IFN-a indices from the individual IFN-a index of a given oligonucleotide (=adjusted IFN-a index). Data from six individual donors were summarized and were assorted in ascending order displaying the corresponding SEM. In addition, a statistical analysis was performed to assess a putative significant difference for the adjusted IFN-α indices of all top thirty ssRNA oligonucleotides. A two-tailed Student\'s t-test was employed to calculate the p-value off all possible ssRNA oligonucleotide combinations. A p-value>0.01 and <0.05 is depicted by a black box, whereas a p-value<0.01 is depicted as a grey box.

FIG. 4: The occurrence of 1mer motifs (5′-X-3′), 2mer motifs (5′-XX-3′,5′-X*X-3′,5′-X**X-3′) and 3mer motifs (5′-XXX-3′,5′-XX*X-3′,5′-X*XX-3′) in ssRNA oligonucleotides with an IFN-a index below the mean IFN-a index (group 1) or above the mean IFN-α index (group 2) was analyzed. The relative occurrence of a given motif within a group of ssRNA oligonucleotides was calculated by dividing the absolute number of occurrences of a given motif within a group through the absolute number of occurrences of all possible motifs within this group. A significant overrepresentation or underrepresentation of a given motif was analyzed using a chi-square test. The null hypothesis of equal distribution within both groups was rejected when the calculated p-value was below 0.05 (significant differences in distribution are indicated by “*”). For all motifs analyzed, relative occurrences are depicted in FIG. 4 for group 1 oligonucleotides (black bars) and for group 2 oligonucleotides (white bars): 1mer motifs 5′-X-3′ (FIG. 4A); 2mer motifs 5′-XX-3′ (FIG. 4B1), 5′-X*X-3′ Figure (FIG. 4B2), 5′-X**X-3′ (FIG. 4B3) and 3mer motifs 5′-XXX-3′ (FIG. 4C1-4C4), 5′-XX*X-3′ (FIG. 4C5-4C8), 5′-X*XX-3′ (FIG. 4C9-4C12).

FIG. 5: For all possible 1 mer motifs (5′-X-3′), 2mer motifs (5′-XX-3′,5′-X*X-3′,5′-X**X-3′) or 3mer motifs (5′-XXX-3′,5′-XX*X-3′,5′-X*XX-3′) a mean IFN-a index was assigned by calculating a mean IFN-α index of all ssRNA oligonucleotides containing the corresponding motifs (=IFN-a score of a given motif). The IFN-α score of all possible motifs is depicted in FIG. 5±SEM: 1mer motifs 5′-X-3′ (FIG. 5A); 2mer motifs 5′-XX-3′ (FIG. 5B1), 5′-X*X-3′ Figure (FIG. 5B2), 5′-X**X-3′ (FIG. 5B3) and 3mer motifs 5′-XXX-3′ (FIG. 5C1-5C4), 5′-XX*X-3′ (FIG. 5C5-5C8), 5′-X*XX-3′ (FIG. 5C9-5C12).

FIG. 6: A calculated IFN-α index was assigned to each oligonucleotide by using the obtained motif-IFN-a scores. For each set of motifs [1mer motifs (5′-X-3′), 2mer motifs (5′-XX-3′,5′-X*X-3′,5′-X**X-3′) or 3mer motifs (5′-XXX-3′,5′-XX*X-3′,5′-X*XX-3] a predicted IFN-a index was calculated for each ssRNA oligonucleotide. Next, the obtained predicted IFN-a indices were compared to the actual adjusted IFN-α indices. Data are depicted the following way: For all ssRNA oligonucleotides the predicted IFN-α indices are shown as a black bars, whereas data are sorted in ascending order according to the actual IFN-α score that is depicted as a red index line. The y-axis on the left side depicts the scale for the predicted IFN-α score, while the y-axis on the right side depicts the scale for the actual IFN-α score.

FIG. 7: PBMC from healthy donors were isolated and stimulated with poly-L-arginine complexed ssRNA oligonucleotides in duplicates. 44 hours after stimulation IFN-a production was assessed in supernatant via ELISA. For all tested ssRNA oligonucleotides, the mean values of the measured duplicates were normalized to the positive control ssRNA oligonucleotide 9.2sense (5′-AGCUUAACCUGUCCUUCAA) by dividing the mean value of tested oligonucleotide by the mean value of 9.2sense (=1). A: A panel of ssRNA oligonucleotides was tested with different positions of the 5′-GUCA-3′-motif within the 19mer ssRNA oligonucleotide (see table 4). The 5′-GUCA-3′-motif is indicated by bold letters. Data from two independent donors were summarized and are depicted as mean values±SEM. B/C: 16 ssRNA oligonucleotides, which included all possible oligonucleotides with permutated bases at the flanking positions to the 5′- and the 3′-end of the central 5′-GUCA-3′-motif (table 5), were complexed with poly-L-arginine and used to stimulate PBMC. 44 hours after stimulation IFN-a production was assessed in supernatant via ELISA. Data for all 16 oligonucleotides from three independent donors were summarized as mean values±SEM (B). In addition all 16 oligonucleotides were assorted into groups according to the base preceding or following the central 5′-GUCA-3′-motif (C). On the left side oligonucleotides with a common base preceding the central 5′-GUCA-3′-motif were grouped, whereas on the right side oligonucleotides with a common base following the central 5′-GUCA-3′-motif were grouped. Individual ssRNA oligonucleotide IFN-a data were summarized according to the respective group and are depicted as mean values±SEM. A two-tailed Student\'s t-test was used to calculate a statistically significant difference between the various groups (p<0.05 is indicated by a “*”).

FIG. 8: A: PBMC from two healthy donors were isolated and stimulated with poly-L-arginine complexed ssRNA oligonucleotides (Table 6) in duplicates. 44 hours after stimulation IFN-α production was assessed in supernatant via ELISA. IFN-α data were summarized as mean values and subsequently normalized to the positive control RNA9.2sense (5′-AGCUUAACCUGUCCUUCAA-3′). In addition, respective sequences were analyzed using the IFN-α point score matrix (Table 7) and subsequently normalized to RNA9.2sense. A correlation coefficient of 0.84 was calculated for these two sets of data. Measured IFN-α levels are depicted in white bars, whereas predicted IFN-α scores are shown in black bars (A). Next, IFN-α point score matrix was employed to analyze IFN-α-inducing RNA oligonucleotides that have been described in the literature. Given the fact that in the study performed by Judge et al. (2005, Nat Biotechnol 23:457-462) double-stranded RNA oligonucleotides were tested, a mean value for the individually analyzed single-stranded components was calculated. Data were normalized to the most potent RNA oligonucleotide (=100%) within the respective panel of oligonucleotides (B). For the prediction of single-stranded RNA oligonucleotides reported by Heil et al. (2004 Science 303:1526-1529), the predicted IFN-α point scores are depicted (C).

FIG. 9: PBMC from individual healthy donors were isolated and stimulated with poly-L-arginine complexed ssRNA-oligonucleotides in duplicates. For a detailed list of all tested oligonucleotides see Table 1. 44 hours after stimulation IFN-α production was assessed in supernatant via ELISA. For all tested ssRNA-oligonucleotides, the mean values of the measured duplicates were normalized to the positive control ssRNA-oligonucleotide 9.2sense (5′-AGCUUAACCUGUCCUUCAA) by dividing the mean value of tested oligonucleotide by the mean value of 9.2sense (=100%). Data from nine different (A-D) donors or three different donors (E) were summarized and are presented as mean values±SEM.

FIG. 10: PBMC from individual healthy donors were isolated and stimulated with poly-L-arginine complexed ssRNA-oligonucleotides in duplicates. For a detailed list of all tested oligonucleotides see Table 13. 44 hours after stimulation IFN-α production was assessed in supernatant via ELISA. For all tested ssRNA-oligonucleotides, the mean values of the measured duplicates were normalized to the positive control ssRNA-oligonucleotide 9.2sense (5′-AGCUUAACCUGUCCUUCAA) by dividing the mean value of tested oligonucleotide by the mean value of 9.2sense (=100%). Data from two different donors were summarized and are presented as mean values±SEM.

FIG. 11: PBMC of six different healthy donors were isolated and stimulated with poly-L-arginine complexed ssRNA-oligonucleotides in duplicates. 44 hours after stimulation IFN-α production was assessed in supernatant via ELISA. For all tested ssRNA-oligonucleotides, the mean values of the measured duplicates were normalized to the positive control ssRNA-oligonucleotide 9.2sense (5′-AGCUUAACCUGUCCUUCAA) by dividing the mean value of tested oligonucleotide by the mean value of 9.2sense. Next, for each individual donor a global normalization to the mean was performed by subtracting the mean of all data from a particular donor from the individual raw data. Data from six individual donors were visualized using tree view and are depicted in ascending order (A). In addition all individual data were summarized as mean values±SEM and are depicted in ascending order (B).

FIG. 12: The occurrence of 3mer motifs in all ssRNA-oligonucleotides was analyzed. The mean level of IFN-α induction was calculated by grouping all oligonucleotides that contained a respective 3mer motif. For example the 3mer motif 5′-GUC-3′ was contained in ssRNA oligonucleotides ANP 35, 83, 131, 137, 138, 139 and 179 with respective IFN-α induction levels of 1.33, 0.68, 0.93, 0.79, 0.44, 0.84 and 0.73. The mean IFN-α induction level of the 3mer motif 5′-GUC-3′ was thus calculated to be 0.82 with a standard error of mean of 0.10. 3mer motifs that were gapped by one nucleotide between either the first and the second nucleotide position (5′-N-NN-3′) or the second and third nucleotide position (5′-NN-N-3′) were also included in the analysis. A two-tailed T-Test was used to identify motifs that were either significantly higher or lower in IFN-α induction than the residual motifs. For all motifs analyzed, the mean IFN-α induction level was visualized using tree view. The data were assorted according to the first nucleotide position of the motif in four groups. (p-value<0.05 is indicated by *).

FIG. 13: The top 15 percent of all ssRNA oligonucleotides and the respective mean IFN-α induction levels are shown in ascending order. The presence of the identified potent 3mer motifs 5′GUY-3′ (5′-GUC-3′,5′-GUU-3′), 5′-GUNY-3′ (5′-GUNC-3′,5′-GUNU-3′) and 5′-GNUY-3′ (5′-GNUC-3′,5′-GNUU-3′) is indicated by a grey box. All ssRNA oligonucleotides that contain any of the above motifs are indicated by a black box.

FIG. 14: PBMC of four different healthy donors were isolated and stimulated with the following poly-L-arginine complexed ssRNA-oligonucleotides: ANP143 (5′-AAAAAAAGUUCAAAAAAAA-3′), RNA40 (5′-GCCCGUCUGUUGUGUGACUC-3′), p-Gal control sense (5′-UUGAUGUGUUUAGUCGCUA-3′) and 9.2sense (5′-AGCUUAACCUGUCCUUCAA-3′). 44 hours after stimulation IFN-α production was assessed in supernatant via ELISA. All tested ssRNA-oligonucleotides were normalized to the positive control ssRNA-oligonucleotide 9.2sense by dividing the mean value of tested oligonucleotide by the mean value of 9.2sense. Significant differences were analyzed using a two-tailed T-Test.

FIG. 15: The occurrence of 3mer motifs was analyzed in the 193 oligonucleotide library. For each oligonucleotide the 3mer motif with the highest calculated mean IFN-α induction level was identified and assigned to the respective oligonucleotide. The predicted data are depicted in ascending order (black bars) according to the corresponding measured IFN-α induction levels (black line). In addition, the correlation coefficient for the two data sets was determined.

FIG. 16: The occurrence of 3mer motifs was analyzed in the 193 oligonucleotide library. For each oligonucleotide the 3mer motif with the highest calculated mean IFN-α induction level was identified and assigned to the respective oligonucleotide. The predicted data are depicted in ascending order. Various threshold levels (dotted lines) were tested for both for the positive predictive value and the sensitivity to identify oligonucleotides below the IFN-α induction level of 0 (A). For each threshold level, data were then regrouped according to the predicted IFN-α induction level (selected oligonucleotides: left group, eliminated oligonucleotides: right group). For each group, the mean level of IFN-α induction±SEM is depicted in the lower panel. The positive predictive value and the sensitivity for each threshold is indicated in the upper left.

FIG. 17: The prediction algorithm was used to analyze all possible siRNA duplexes targeting the mRNA of human TLR9 (NM—017442). For the 3868 bp long mRNA of TLR9 all possible 19mer siRNA duplexes were considered and the IFN-α prediction algorithm was applied on both the sense and the antisense strand of each siRNA duplex. The predicted IFN-α induction levels are depicted in stacked columns for the sense (upper columns in black) and the antisense strand (lower column in grey). The relative targeting position of the siRNA duplex is given on the y-axis, whereas the predicted IFN-α induction is depicted on the x-axis (A). In addition, six selected regions of the TLR9 mRNA and the respective predicted IFN-α induction levels are depicted in detail in B.

FIG. 18: HEK 293 cells were transfected with an expression plasmid coding for human TLR9 with a C-terminal YFP-tag. Various siRNA-duplices targeting human TLR9 mRNA were cotransfected. The starting base of the individual siRNA is given in the lower panel. 20 hours after transfection, TLR9 expression was analyzed by flow cytometry. Data are depicted as percentage of TLR9-expression referring to an irrelevant control siRNA as 100% and siRNA_sb1647 as 0%. Results are shown as mean values±SEM (n=3) (A). In addition, above siRNA duplexes and the respective single stranded components were used to transfect human PBMC from five individual donors. 40 hours after transfection IFN-α induction was measured via ELISA. Data are depicted as mean values±SEM (B).

FIG. 19: Based on the algorithm described in example 17, a computer program was written that applies the algorithm to all possible siRNA duplexes targeting all human RNA transcripts (50421 as of 09/2006) as published by the National Center for Biotechnology Information (NCBI). Each entry into the NCBI database (ftp://ftp.ncbi.nih.gov/refseq/H_sapiens/mRNA_Prot/human.rna.fna.gz) of all listed human RNA transcripts was analyzed the following way: A list of all possible 19mer siRNA duplexes targeting a given RNA transcript was generated. Of all siRNA duplexes the IFN-α induction of both the sense and the antisense strand was predicted using the method described in example 17. The obtained data is stored in a database (CD-ROM) and can be retrieved by a search engine. Using the search interface, the user can pick the transcript of interest (alphabetical index of all RNA transcripts targeted by siRNAs) and then adjust the level of threshold to identify siRNA duplexes that are of either low, intermediate or high in immunostimulatory activity (A). For example, using the threshold of 0.11 as described in example 17, a set of siRNA duplexes was identified for Homo sapiens vascular endothelial growth factor (VEGF) transcript variant 1 mRNA (NM—001025366.1) with low immunostimulatory activity for both the sense and the antisense strand (B).

DETAILED DESCRIPTION

As used herein, “a” and “an” refers to a group or species of entities, rather than one single individual.

oligonucleotide

As used herein, the term “oligonucleotide” refers to a polynucleotide formed from a plurality of linked nucleoside units. Such oligonucleotides can be obtained from existing nucleic acid sources, including genomic or cDNA, but are preferably produced by synthetic methods including chemical synthesis, in vitro and in vivo transcription. In preferred embodiments each nucleoside unit includes a heterocyclic base and a pentofuranosyl, trehalose, arabinose, 2′-deoxy-2′-substituted arabinose, 2′-O-substituted arabinose or hexose sugar group. The nucleoside residues can be coupled to each other by any of the numerous known internucleoside linkages. Such internucleoside linkages include, without limitation, phosphodiester, phosphorothioate, phosphorodithioate, pyrophosphate, alkylphosphonate, alkylphosphonothioate, phosphotriester, phosphoramidate, siloxane, carbonate, carboalkoxy, acetamidate, carbamate, morpholino, borano, thioether, bridged phosphoramidate, bridged methylene phosphonate, bridged phosphorothioate, and sulfone internucleoside linkages. The term “oligonucleotide” also encompasses polynucleosides having one or more stereospecific internucleoside linkage (e.g., (Rp)- or (Sp)-phosphorothioate, alkylphosphonate, or phosphotriester linkages).

The oligonucleotides of the invention can include naturally occurring nucleosides, modified nucleosides, or mixtures thereof. As used herein, the term “modified nucleoside” is a nucleoside that includes a modified heterocyclic base, a modified sugar moiety, or a combination thereof. In some embodiments, the modified nucleoside is a non-natural pyrimidine or purine nucleoside, as herein described. In some embodiments, the modified nucleoside is a 2′-substituted ribonucleoside an arabinonucleoside or a 2′-deoxy-2′-substituted-arabinoside.

As used herein, the term “2′-substituted ribonucleoside” or “2′-substituted arabinoside” includes ribonucleosides or arabinonucleoside in which the hydroxyl group at the 2′ position of the pentose moiety is substituted to produce a 2′-substituted or 2′-O-substituted ribonucleoside. Preferably, such substitution is with a lower alkyl group containing 1-6 saturated or unsaturated carbon atoms, or with an aryl group having 6-10 carbon atoms, wherein such alkyl, or aryl group may be unsubstituted or may be substituted, e.g., with halo, hydroxy, trifluoromethyl, cyano, nitro, acyl, acyloxy, alkoxy, carboxyl, carboalkoxy, or amino groups. Examples of 2′-O-substituted ribonucleosides or 2′-O-substituted-arabinosides include, without limitation 2′-O-methylribonucleosides or 2′-O-methylarabinosides and 2′-O-methoxyethylribonucleosides or 2′-O-methoxyethylarabinosides.

The term “2′-substituted ribonucleoside” or “2′-substituted arabinoside” also includes ribonucleosides or arabinonucleosides in which the 2′-hydroxyl group is replaced with a lower alkyl group containing 1-6 saturated or unsaturated carbon atoms, or with an amino or halo group. Examples of such 2′-substituted ribonucleosides or 2′-substituted arabinosides include, without limitation, 2′-amino, 2′-fluoro, 2′-allyl, and 2′-propargyl ribonucleosides or arabinosides.

The term “oligonucleotide” includes hybrid and chimeric oligonucleotides. A “chimeric oligonucleotide” is an oligonucleotide having more than one type of internucleoside linkage. One preferred example of such a chimeric oligonucleotide is a chimeric oligonucleotide comprising a phosphorothioate, phosphodiester or phosphorodithioate region and non-ionic linkages such as alkylphosphonate or alkylphosphonothioate linkages (see e.g., Pederson et al. U.S. Pat. Nos. 5,635,377 and 5,366,878).

A “hybrid oligonucleotide” is an oligonucleotide having more than one type of nucleoside. One preferred example of such a hybrid oligonucleotide comprises a ribonucleotide or 2′-substituted ribonucleotide region, and a deoxyribonucleotide region (see, e.g., Metelev and Agrawal, U.S. Pat. Nos. 5,652,355, 6,346,614 and 6,143,881).

RNA oligonucleotides discussed herein include otherwise unmodified RNA as well as RNA which have been modified (e.g., to improve efficacy), and polymers of nucleoside surrogates. Unmodified RNA refers to a molecule in which the components of the nucleic acid, namely sugars, bases, and phosphate moieties, are the same or essentially the same as that which occur in nature, preferably as occur naturally in the human body. The art has referred to rare or unusual, but naturally occurring, RNAs as modified RNAs, see, e.g., Limbach et al. 1994, Nucleic Acids Res 22: 2183-2196. Such rare or unusual RNAs, often termed modified RNAs (apparently because these are typically the result of a post-transcriptional modification) are within the term unmodified RNA, as used herein. Modified RNA as used herein refers to a molecule in which one or more of the components of the nucleic acid, namely sugars, bases, and phosphate moieties, are different from that which occurs in nature, preferably different from that which occurs in the human body. While they are referred to as modified “RNAs,” they will of course, because of the modification, include molecules which are not RNAs. Nucleoside surrogates are molecules in which the ribophosphate backbone is replaced with a non-ribophosphate construct that allows the bases to the presented in the correct spatial relationship such that hybridization is substantially similar to what is seen with a ribophosphate backbone, e.g., non-charged mimics of the ribophosphate backbone.

All nucleic acid sequences listed herein are in the 5′ to 3′ direction unless otherwise indicated.

The RNA oligonucleotide of the invention can be single-stranded, double stranded, or partially double-stranded.

A single-stranded RNA oligonucleotide may contain self-complementary sequences and forms a hairpin. For example, 5′-GACCUAGCCUAAAACUAGGUC-3′. The self-complementary sequence may be a palindromic sequence. For example, 5′AAAGAUCCGGAUCAAAA-3′.

A double stranded RNA oligonucleotide may have one- or two-nucleotide overhang at the 5′ or 3′ end of one or both strands.

A partially double-stranded RNA oligonucleotide may comprise two strands of the same or different length, wherein the at least one of the strands contains nucleotides outside the complementary sequence. For example,

Esample 1: 5′-AAAAGUUCAAAGCUCAAAA-3′ 3′-CAAGUUUCGAG-5′ Example 2: 5′-UCAAAGUCAAAAGCUCAAAGUUGAAAGUUUAAA-3′ 3′-GACUUGAAAAUUUCAGUUUUCGAGUUUAAGUUGAAAACUCG-5′ Example 3: 5′-UCAAAGUCAAAAGCUCAAAGUUGAAA-3′ 3′-UUUCAGUUUUCGAGUUUAAGUUGAAAACUCG-5′

The length of a single-stranded RNA oligonucleotide is the number of nucleotides contained in the oligonucleotide.

In the case of a double-stranded or partially double-stranded oligonucleotide, the length of the oligonucleotide is the length of the individual strands. In other words, a partially double-stranded oligonucleotide can have two lengths.

For increased nuclease resistance and/or binding affinity to the target, an oligonucleotide can include, for example, 2′-modified ribose units and/or phosphorothioate linkages and/or pyrophosphate linkages. For example, the 2′ hydroxyl group (OH) can be modified or replaced with a number of different “oxy” or “deoxy” substituents.

Examples of “oxy”-2′ hydroxyl group modifications include alkoxy or aryloxy (OR, e.g., R═H, alkyl, cycloalkyl, aryl, aralkyl, heteroaryl or sugar); polyethyleneglycols (PEG), O(CH2CH2O)nCH2CH2OR; “locked” nucleic acids (LNA) in which the 2′ hydroxyl is connected, e.g., by a methylene bridge, to the 4′ carbon of the same ribose sugar; G-AMINE and aminoalkoxy, O(CH2)nAMINE, (e.g., AMINE=NH2; alkylamino, dialkylamino, heterocyclyl amino, arylamino, diaryl amino, heteroaryl amino, or diheteroaryl amino, ethylene diamine, polyamino). It is noteworthy that oligonucleotides containing only the methoxyethyl group (MOE), (OCH2CH2OCH3, a PEG derivative), exhibit nuclease stabilities comparable to those modified with the robust phosphorothioate modification. “Deoxy” modifications include hydrogen (i.e. deoxyribose sugars, which are of particular relevance to the overhang portions of partially ds RNA); halo (e.g., fluoro); amino (e.g. NH2; alkylamino, dialkylamino, heterocyclyl, arylamino, diaryl amino, heteroaryl amino, diheteroaryl amino, or amino acid); NH(CH2CH2NH)nCH2CH2-AMINE (AMINE=NH2; alkylamino, dialkylamino, heterocyclyl amino, arylamino, diaryl amino, heteroaryl amino, or diheteroaryl amino), —NHC(O)R(R=alkyl, cycloalkyl, aryl, aralkyl, heteroaryl or sugar), cyano; mercapto; alkyl-thio-alkyl; thioalkoxy; and alkyl, cycloalkyl, aryl, alkenyl and alkynyl, which may be optionally substituted with e.g., an amino functionality.

Preferred substitutents are 2′-methoxyethyl, 2′-OCH3, 2′-O-allyl, 2′-C— allyl, and 2′-fluoro. To maximize nuclease resistance, the 2′ modifications can be used in combination with one or more phosphate linker modifications (e.g., phosphorothioate). The so-called “chimeric” oligonucleotides are those that contain two or more different modifications. The inclusion of furanose sugars in the oligonucleotide backbone can also decrease endonucleolytic cleavage. An oligonucleotide agent can be further modified by including a 3′ cationic group, or by inverting the nucleoside at the 3′-terminus with a 3′-3′ linkage. In another alternative, the 3′-terminus can be blocked with an aminoalkyl group, e.g., a 3′ C5-aminoalkyl dT. Other 3′ conjugates can inhibit 3′-5′ exonucleolytic cleavage. While not being bound by theory, a 3′ conjugate, such as naproxen or ibuprofen, may inhibit exonucleolytic cleavage by sterically blocking the exonuclease from binding to the 3′-end of oligonucleotide. Even small alkyl chains, aryl groups, or heterocyclic conjugates or modified sugars (D-ribose, deoxyribose, glucose etc.) can block 3′-5′-exonucleases.

Similarly, 5′ conjugates can inhibit 5′-3′ exonucleolytic cleavage. While not being bound by theory, a 5′ conjugate, such as naproxen or ibuprofen, may inhibit exonucleolytic cleavage by sterically blocking the exonuclease from binding to the 5′-end of oligonucleotide. Even small alkyl chains, aryl groups, or heterocyclic conjugates or modified sugars (D-ribose, deoxyribose, glucose etc.) can block 3′-5′-exonucleases.

Single-stranded RNA oligonucleotides which contain self-complementary sequences and form a hairpin structure have enhanced nuclease resistance compared to single-stranded oligonucleotides which do not.

The oligonucleotides of the present invention can be 5′ phosphorylated or can include a phosphoryl analog at the 5′ prime terminus. 5′-phosphate modifications of the antisense strand include those which are compatible with RISC mediated gene silencing. Suitable modifications include: 5′-monophosphate ((HO)2(O)P—O-5′); 5′-diphosphate ((HO)2(O)P—O—P(HO)(O)—O-5′); 5′-triphosphate ((HO)2(O)P—O—(HO)(O)P—O—P(HO)(O)—O-5′); 5′-guanosine cap (7-methylated or non-methylated) (7m-G-O-5′-(HO)(O)P—O—(HO)(O)P—O—P(HO)(O)—O-5′); 5′-adenosine cap (Appp), and any modified or unmodified nucleotide cap structure. Other suitable 5′-phosphate modifications will be known to the skilled person.

The RNA oligonucleotides of the present invention also include those with tethered ligands. The properties of a RNA oligonucleotide, including its pharmacological properties, can be influenced and tailored by the introduction of ligands, e.g. tethered ligands.

The ligands may be coupled, preferably covalently, either directly or indirectly via an intervening tether, to the RNA oligonucleotide. In preferred embodiments, the ligand is attached to the oligonucleotide via an intervening tether.

In preferred embodiments, a ligand alters the distribution, targeting or lifetime of a RNA oligonucleotide into which it is incorporated. In preferred embodiments a ligand provides an enhanced affinity for a selected target, e.g., molecule, cell or cell type, a cellular or organ compartment, tissue, organ or region of the body.

Preferred ligands can improve transport, hybridization, and specificity properties and may also improve nuclease resistance of the resultant natural or modified oligoribonucleotide, or a polymeric molecule comprising any combination of monomers described herein and/or natural or modified ribonucleotides.

A wide variety of ligands may be used. Ligands may include agents that allow for the specific targeting of the oligonucleotide; diagnostic compounds or reporter groups which allow for the monitoring of oligonucletotide distribution; cross-linking agents; nuclease-resistance conferring moieties; and natural or unusual nucleobases. General examples include lipophilic moleculeses, lipids, lectins, steroids (e.g., uvaol, hecigenin, diosgenin), terpenes (e.g., triterpenes, e.g., sarsasapogenin, Friedelin, epifriedelanol derivatized lithocholic acid), vitamins, carbohydrates (e.g., a dextran, pullulan, chitin, chitosan, inulin, cyclodextrin or hyaluronic acid), proteins, protein binding agents, integrin targeting molecules, polycationics, peptides, polyamines, and peptide mimics.

The ligand may be a naturally occurring or recombinant or synthetic molecule, such as a synthetic polymer, e.g., a synthetic polyamino acid. Examples of polyamino acids include polyamino acid is a polylysine (PLL), poly L-aspartic acid, poly L-glutamic acid, styrene-maleic acid anhydride copolymer, poly(L-lactide-co-glycolied) copolymer, divinyl ether-maleic anhydride copolymer, N-(2-hydroxypropyl)methacrylamide copolymer (HMPA), polyethylene glycol (PEG), polyvinyl alcohol (PVA), polyurethane, poly(2-ethylacrylic acid), N-isopropylacrylamide polymers, or polyphosphazine. Example of polyamines include: polyethylenimine, polylysine (PLL), spermine, spermidine, polyamine, pseudopeptide-polyamine, peptidomimetic polyamine, dendrimer polyamine, arginine, amidine, protamine, cationic moieties, e.g., cationic lipid, cationic porphyrin, quaternary salt of a polyamine, or an alpha helical peptide.

Ligands can also include targeting groups, e.g., a cell or tissue targeting agent, e.g., a thyrotropin, melanotropin, surfactant protein A, Mucin carbohydrate, a glycosylated polyaminoacid, transferrin, bisphosphonate, polyglutamate, polyaspartate, or an RGD peptide or RGD peptide mimetic.

Ligands can be proteins, e.g., glycoproteins, lipoproteins, e.g. low density lipoprotein (LDL), or albumins, e.g. human serum albumin (HSA), or peptides, e.g., molecules having a specific affinity for a co-ligand, or antibodies e.g., an antibody, that binds to a specified cell type such as a cancer cell, endothelial cell, or bone cell. Ligands may also include hormones and hormone receptors. They can also include non-peptidic species, such as cofactors, multivalent lactose, multivalent galactose, N-acetyl-galactosamine, N-acetyl-glucosamine, multivalent mannose, or multivalent fucose. The ligand can be, for example, a lipopolysaccharide, an activator of p38 MAP kinase, or an activator of NE-κB. The ligand can be a substance, e.g., a drug, which can increase the uptake of the oligonucleotide agent into the cell, for example, by disrupting the cell\'s cytoskeleton, e.g., by disrupting the cell\'s microtubules, microfilaments, and/or intermediate filaments. The drug can be, for example, taxon, vincristine, vinblastine, cytochalasin, nocodazole, japlakinolide, latrunculin A, phalloidin, swinholide A, indanocine, or myoservin.

In one embodiment, the ligand is a lipid or lipid-based molecule. Such a lipid or lipid-based molecule preferably binds a serum protein, e.g., human serum albumin (HSA). An HSA binding ligand allows for distribution of the conjugate to a target tissue, e.g., liver tissue, including parenchymal cells of the liver. Other molecules that can bind HSA can also be used as ligands. For example, neproxin or aspirin can be used. A lipid or lipid-based ligand can (a) increase resistance to degradation of the conjugate, (b) increase targeting or transport into a target cell or cell membrane, and/or (c) can be used to adjust binding to a serum protein, e.g., HSA.

A lipid based ligand can be used to modulate, e.g., control the binding of the conjugate to a target tissue. For example, a lipid or lipid-based ligand that binds to HSA more strongly will be less likely to be targeted to the kidney and therefore less likely to be cleared from the body. A lipid or lipid-based ligand that binds to HSA less strongly can be used to target the conjugate to the kidney.

In another embodiment, the ligand is a moiety, e.g., a vitamin or nutrient, which is taken up by a target cell, e.g., a proliferating cell. These are particularly useful for treating disorders characterized by unwanted cell proliferation, e.g., of the malignant or non-malignant type, e.g., cancer cells. Exemplary vitamins include vitamin A, E, and K. Other exemplary vitamins include the B vitamins, e.g., folic acid, B12, riboflavin, biotin, pyridoxal or other vitamins or nutrients taken up by cancer cells.

In another embodiment, the ligand is a cell-permeation agent, preferably a helical cell-permeation agent. Preferably, the agent is amphipathic. An exemplary agent is a peptide such as tat or antennapedia. If the agent is a peptide, it can be modified, including a peptidylmimetic, invertomers, non-peptide or pseudo-peptide linkages, and use of D-amino acids. The helical agent is preferably an alpha-helical agent, which preferably has a lipophilic and a lipophobic phase.

In a preferred embodiment, the ligand is an antibody or a fragment thereof which is specific for a moiety present in a cell to be targeted. The moiety may be a protein, a carbohydrate structure, a polynucleotide, or a combination thereof. The moiety may be secreted, associated with the plasma membrane (e.g., on the extracellular or intracellular surface), cytosolic, associated with intracellular organelles (e.g., ER, Golgi complex, mitochondria, endosome, lysosome, secretory vesicle) or nuclear. The antibody may be monoclonal or polyclonal. The antibody may be chemeric or humanized. The antibody may be a single chain antibody. The antibody fragment may be a Fab fragment, a F(ab′)2 fragment, or any fragments that retain the antigen-binding specificity of the intact antibody.

As used herein, “immunostimulatory activity” refers to the capability of a molecule or a composition to induce an immune response. In one aspect, the immunostimulatory activity refers to the type I-IFN-inducing activity, in particular, the IFN-α-inducing activity.

As used herein, “inducing an immune response” means initiating or causing an increase in one or more of B-cell activation, T-cell activation, natural killer cell activation, activation of antigen presenting cells (e.g., B cells, dendritic cells, monocytes and macrophages), cytokine production, chemokine production, specific cell surface marker expression, in particular, expression of co-stimulatory molecules. In one aspect, such an immune response involves the production of type I IFN, in particular, IFN-α, in cells such as PDC.

As used herein, “IFN-α-inducing activity” refers to the capability of a molecule or composition to induce IFN-α production from a cell capable of producing IFN-α. Cells capable of producing IFN-α include, but are not limited to, peripheral blood mononuclear cells (PBMC) (e.g., B cells, dendritic cells (myeloid dendritic cells and plasmacytoid dendritic cells), macrophages, monocytes, natural killer cells, granulocytes), endothelial cells, and cell lines (e.g., THP1; cells transfected with expression vectors for TLR-7 and/or TLR-8 such as CHO cells, COS cells, HEK293 cells). Cells capable of producing IFN-α include those that express TLR7, TLR8, or both TLR7 and TLR8.

As used herein, “gene silencing” refers to the downregulation or the abolition of the expression of a target gene. Gene silencing as used herein, occurs at the post-transcriptional level. Gene silencing may be directly or indirectly mediated by siRNA, shRNA and antisense RNA.

Both the antisense-strand of the siRNA and the antisense RNA have complementary to the target mRNA and are the effector strand of the gene silencing activity. The term complementary is well understood by those skilled in the art. For example, A is complementary to T, G is complementary to C, 5′-AG-3′ is complementary to 5′-CT-3′.

The degree of complementarity between two oligonucleotides is the percentage of complementary bases in the overlapping region of the two oligonucleotides. The degree of complementarily can be determined manually or automatically by various engines such as BLAST. For example, ATCG has 100% complementarity to CGAT and CGATGG, and 75% complementarity to CGTT and CGTTGG. Furthermore, the degree of complementarity between a RNA oligonucleotide and any sequences present in the public databases (e.g., EMBL, GeneBank) can be determined by the BLAST program.

The degree of complementarity between the antisense strand of the siRNA or the antisense RNA and the target mRNA is at least 80% 81%, 82%, 83%, preferably at least 84%, 85%, 86%, 87%, 88%, more preferably at least 89%, 90%, 91%, 92%, 93%, even more preferably at least 94%, 95%, 96%, 97%, 98%, 99%, and most preferably 100%.

The gene silencing activity of a RNA oligonucleotide can be determined experimentally by methods well known in the art. For Example, the RNA oligonucleotide may be introduced into a cell by a method known in the art such as transfection and transduction; the mRNA level of the target gene can be determined by routine methods such as Northern blot analysis, quantitative PCR, RNase protection assay, and branching DNA; and the protein expression level can be determined by routine methods such as Western blotting, ELISA, and biological activity assays specific to the target protein. Furthermore, the mRNA level of all known and hypothetical genes can be determined at the global level using the microarray technology. Technologies in the field of proteonomics allow for the protein levels of a large number of genes to be determined at the global level as well.

Naked RNA oligonceotide may be transfected into a cell via electroporation. RNA oligonucleotide may be complexed with a complexation agent which facilitates the uptake of the oligonucletide into a cell. Such complexation agents include, but are not limited to cationic lipids (e.g., Lipofectamine, Oligofectamine, DOTAP), cationic peptides, and calcium phosphate.

The gene silencing activity of a RNA oligonucleotide can be predicted by algorithms such as the one disclosed in Reynolds et al. 2004, Nat Biotechnol 22:326-330.

siRNA

As used herein, “siRNA” stands for short interfering RNA, and has the same definition as that established in the art. siRNA is double-stranded and is usually between 19 and 27 nucleotide in length. In vivo, siRNA is the product of Dicer activity on long dsRNA. The antisense strand of siRNA is complementary to the target mRNA; it binds the target mRNA and induces RISC-mediated target mRNA degradation. siRNA can be chemically synthesized, produced in vitro by Dicer-mediated enzymatic degradation of long dsRNA, produced by in vitro transcription from linear (e.g. PCR products) or circular templates (e.g., viral or non-viral vectors), or produced by in vivo transcription from viral or non-viral vectors. Commercially available synthetic siRNA usually contain a core of 19 complemetary base pairs and a 2-nucleotide (UU or TT) 3′ overhang on each strand. siRNA may be chemically modified to have enhanced stability in vitro (especially in serum-containing media) and in vivo. siRNA may also be chemically modified to have enhanced uptake by cells in vitro and in vivo. Furthermore, siRNA may be linked to tethered ligands to have enhanced target specificity and improved pharmacological properties (such as half-life, clearance, distribution).

shRNA

As used herein, “shRNA” stands for short hairpin RNA and has the same definition as that established in the art. shRNA is processed inside a cell into siRNA which mediates RNAi as described previously. The loop sequence in shRNA is not thought to be involved in RNAi, and it can be of various lengths and sequences. The preferred lengths and sequences of the loop are known to those skilled in the art.

Similar to siRNA, shRNA can be chemically synthesized, produced by in vitro transcription from linear (e.g. PCR products) or circular templates (e.g., viral or non-viral vectors), or produced by in vivo transcription from viral or non-viral vectors.

As used herein, “antisense RNA” has the same definition as that established in the art. Antisense RNA is complementary to target mRNA and it thought to interfere with the translation of the target mRNA. Antisense RNA molecules are usually 18-50 nucleotides in length. Antisense RNA may be modified to have enhanced stability, nuclease resistance, target specificity and improved pharmacological properties.

Similar to siRNA and shRNA, antisense RNA can be chemically synthesized, produced by in vitro transcription from linear (e.g. PCR products) or circular templates (e.g., viral or non-viral vectors), or produced by in vivo transcription from viral or non-viral vectors.

As used herein, “disorder/disease-related gene” refers to a gene that is expressed or overexpressed in a disease/disorder and that is not expressed or expressed in reduced amount under normal condition. For example, a mutant CF gene is expressed in cystic fibrosis patient but not in an individual without cystic fibrosis; ErbB2 (or Her2) is overexpressed in breast cancer cells compared to normal breast cells; a viral gene is expressed in infected cells but not in uninfected cells. The gene product of the disorder/disease-related gene is referred to herein as the “disorder/disease-related antigen”.

As used herein, the term “mammal” includes, without limitation, rats, mice, cats, dogs, horses, sheep, cattle, cows, pigs, rabbits, non-human primates, and humans.

The vertebrate immune system established different ways to detect invading pathogens based on certain characteristics of their microbial nucleic acids. Detection of microbial nucleic acids alerts the immune system leading to the appropriate type of immune responses that is required for the defence against the respective type of pathogen detected. Detection of viral nucleic acids leads to the production of type I IFN, the key cytokine for anti-viral defence. While it is well established that the recognition of microbial DNA is sequence-specific, involving the so-called CpG motifs, the optimal motif for the recognition of microbial RNA has not been defined yet. The present application provides a technology platform for identifying the optimal motif for the recognition of microbial RNA and the induction of type I IFN.

The technology platform of the present invention comprises three key features: i) the transfection of peripheral blood mononuclear cells (PBMC) from healthy donors with RNA oligonucleotides; ii) the generation of a RNA oligonucleotide library containing 4mer motifs on a poly adenosine (polyA) backbone; iii) the development of algorithms based on the experimental data generated for the RNA oligonucleotide library to predict the immunostimulatory activity of any given RNA oligonucleotide.

The first key feature of the technology platform is the method of introducing RNA oligonucleotides into PBMC. Naked RNA oligonucleotides are not taken up by the cells to any significant degree. RNA oligonucleotides normally need to form complexes with complexation agent (or transfection agent) in order to be introduced into cells. In the literature, cationic lipids such as lipofectamine or DOTAP are routinely used as complexation agent for the transfection of RNA oligonucleotides. However, RNA-cationic lipid complexes lead to rapid cell death of myeloid cells. Although myeloid cells are not the cellular source of IFN-α within PBMC (the source is PDC), death of myeloid cells in the cell culture negatively affects the reproducibility of IFN-α induction in PBMC. Therefore, the use of cationic lipids is limited to isolated PDC. However, isolated PDC are not suitable for large scale screening assays because PDC make up 0.2-0.6% of the PBMC in a normal individual; it is difficult to obtain enough cells for the assays.

To identify a complexation agent that is suitable for use with PBMC, we compared different types of cationic peptides, poly-His, poly-L-Lys, and poly-L-arg. Poly-L-arg was found to provide the most potent support for the immunostimulatory activity of RNA oligonucleotides when compared to other cationic peptides and cationic lipids (FIG. 1). A protocol was then established that allows well-controlled and highly reproducible complex formation between the RNA oligonucleotide and the complexation agent and subsequent RNA transfection into cells. Complex formation could be controlled by salt concentration, phosphate content and incubation time. Complex formation was monitored by the size of complexes and the functional activity over a range of concentrations. The use of poly-L-arg did not affect the viability of myeloid cells and thus could be applied to PBMC without restrictions.

The second key issue of the technology platform was the generation of the RNA oligonucleotide library. An earlier study showed that a minimal length of 19 bases was required for the optimal immunostimulatory activity of an RNA oligonucleotide; furthermore, it showed that poly adenosine (poly A) was completely inactive (Hornung V et al. 2005, Nat Med 11:263-270). Therefore, the motif search was performed with a 19mer oligonucleotide on a poly A sequence background. By adding increasing numbers of uridine (U) in the center of such a poly A oligonucleotide, we found that a 4-nucleotide (4-mer) motif in the center was sufficient to confer marked immunostimulatory activity (FIG. 2). Importantly, after identifying the optimal 4mer sequence motifs for inducing IFN-α production, we found that changing the bases flanking the 4mer motifs did not further enhance the immunostimulatroy activity of the 4mer motifs (FIG. 7B). The library of 193 RNA oligonucleotides used covered all 256 possible 4mer motifs. The reduction from 258 to 193 was possible because of redundant motifs caused by the poly A flanking regions. In additional studies we found that the exact location of the 4mer motif within the poly A backbone is not critical for the immunostimulatory activity (FIG. 7A).

The third key feature of the technology platform was the generation of a data matrix and its mathematical analysis. Algorithms were developed that allowed an excellent prediction of the immunostimulatory activity of RNA oligonucleotides. The frequency of a given 4mer motif at a certain position within an oligonucleotide is only 1:256. Even though the most active 4mer motifs can be used as the core for constructing potent immunostimulatory RNA oligonucleotides, the IFN-α indices of the 4mer motifs are not particularly useful for predicting the activity of a given RNA oligonucleotide, or for designing RNA oligonucleotides with minimal immunostimulatory activity which is desired for an siRNA. Therefore, algorithms were established which based on parts of the 4mer motifs, namely 1, 2 or 3 bases either in a row (XXX) or with spacing (X*XX; XX*X). The highest predictive value was obtained with the algorithm using 3 bases (i.e., 3mer motifs). This 3mer-based algorithm allowed an impressively accurate prediction (correlation coefficient (r)=0.87) of the immunostimulatory activity of the 19mer RNA oligonucleoties carrying 4mer motifs in our library (FIG. 6E) and RNA oligonucleotides previously published in the literature by us and others (FIG. 8A-C).

There are a number of applications for the information generated by our technology platform: a) the 4mer motif data matrix can be used to design oligonucleotides with optimal IFN-α-inducing activity; b) 4mer motifs with minimal IFN-α-inducing activity can be used as the repertoire for selecting potential inhibitory sequence motifs; c) the 3mer-based algorithm (e.g., the IFN-α point score matrix) can be used to predict the immunostimulatory activity of a given RNA oligonucleotide; d) the 3mer-based algorithm (e.g., the IFN-α point score matrix) can be used to design RNA oligonucleotides with maximal immunostimulatory activity and additional sequence requirements for other functionalities such as gene silencing in the case of an siRNA (in this case the use of 4mer motif matrix is not useful since 4mer motifs are not frequent enough); e) the 3mer-based algorithm (e.g., the IFN-α point score matrix) can also be used to design RNA nucleotides with minimal immunostimulatory activity and additional sequence requirements for other functionalities such as gene silencing in the case of an siRNA (in this case the use of 4mer motif matrix is not useful since 4mer motifs are not frequent enough).

In the case of an immunostimulatory RNA, an oligonucleotide containing only one of the most potent 4mer motifs is 80% more active than the most active complex oligonucleotide containing a 9mer motif in the literature (Table 1). In a 19mer oligonucleotide, there is room for several potent 4mer motifs. A 19mer RNA olignucleotide containing more than one potent immunostimulatory 4mer motifs is expected to have even higher activity.

Furthermore, inhibitory motifs may exist that inhibit the immunostimulatory activity of a RNA oligonucleotide as in the case of CpG oligonucleotides. Such inhibitory motifs, by definition, are among the motifs with weak IFN-α-inducing activity. In the field of RNA interference, type I IFN induction usually is unwanted. The 3mer-based algorithm (e.g., the IFN-α point score matrix) described above can be used to select siRNA sequences with minimal immunostimulatory activity. A sequence analysis of cyclophylin B mRNA, one of the best studied targets for siRNA, identifies a number of siRNA sequences for which our algorithm (e.g., the IFN-α point score matrix) predicts minimal type I IFN induction and which still are known to be potent in gene silencing (Reynolds A et al. 2004, Nat Biotechnol 22: 326-330). This confirms our previous finding that RNA interference and IFN-α induction are two independent functional activities of a siRNA molecule.

Of note, the motif search performed in the present study focuses on the activity of RNA oligonucleotides to induce IFN-α. From previous studies, it is known that the cellular source of IFN-α within the PBMC is PDC. By analysing the level of IFN-α induction in PBMC, other activities of the RNA oligonucleotides on other cellular subsets of the PBMC, such as myeloid cells, are not addressed. Myeloid cells express TLR8 in addition to TLR7 and thus may show different nucleotide sequence specificities and may be induced to exhibit additional activities than IFN-α production. It therefore needs to be born in mind that ssRNA oligonucleotides are capable of inducing both PDC-dependent (i.e. IFN-α production) and PDC-independent activities (e.g., activation of myeloid cells). In contrast, we found that dsRNA oligonucleotides, such as siRNA, are only recognized by PDC but not myeloid cells. As a result, it is valid to predict the immunological activity of siRNA oligonculeotides based on their ability to induce IFN-α production.

The present invention provides a method for determining the immunostimulatory activity, in particular, the IFN-α-inducing activity, of a RNA oligonucleotide, comprising the steps of:

(a) complexing the RNA oligonucleotide with a complexation agent;

(b) contacting a cell with the complexed RNA oligonucleotide, wherein the cell expresses TLR7 or TLR8 or both TLR7 and TLR8; and

(c) determining the amount of IFN-α produced by the cell of step (b), an increase of IFN-α production indicating immunostimulatory activity of the RNA oligonucleotide.

In one embodiment of the invention, the complexation agent is a polycationic peptide, preferably poly-L-arginine (poly-L-arg). In one embodiment, the polycationic peptide, in particular, poly-L-arg, is at least 24 amino acids in length. The polycationic peptide, in particular, poly-L Arg, may be a heterogeneous mixture of peptides of different lengths.

The cells expressing TLR7 or TLR8 or both TLR7 or TLR8 include, but are not limited to, peripheral blood mononuclear cells (PBMC), plasmacytoid dendritric cells (PDC), myeloid dendritic cells (MDC), B cells, macrophages, monocytes, natural killer cells, granulocytes, endothelial cells, cell lines such as THP1, and cells containing exogenous DNA which directs the expression of TLR7 or TLR8 or both TLR7 or TLR8 such as transfected CHO, HEK293, and COS cells.

In one embodiment of the invention, the cell is a mammalian cell, preferably a human cell or a cell of human origin.

The RNA oligonucleotide can be single-stranded, double-stranded or partially double-stranded.

The present invention provides a method for predicting the immunostimulatory activity, in particular the IFN-α-inducing activity, of a RNA oligonucleotide, comprising the steps of: (a) identifying all possible 3-nucleotide (3mer) motifs contained in the oligonucleotide; (b) assigning an IFN-α point score for each individual 3mer motif; (c) assigning the sum of the IFN-α point scores of individual 3mer motifs as the IFN-α score of the oligonucleotide; and (d) assigning to the oligonucleotide a high immunostimulatory activity if the IFN-α score is at least 23, an intermediate immunostimulatory activity if the IFN-α score is between −4 and 23, and a low immunostimulatory activity if the IFN-α score is at most −4, when n=6; assigning to the oligonucleotide a high immunostimulatory activity if the IFN-α score is at least 26, an intermediate immunostimulatory activity if the IFN-α score is between −4 and 26, and a low immunostimulatory activity if the IFN-α score is at most −4, when n=7; assigning to the oligonucleotide a high immunostimulatory activity if the IFN-α score is at least 28, an intermediate immunostimulatory activity if the IFN-α score is between −5 and 23, and a low immunostimulatory activity if the IFN-α score is at most −5, when n=8; assigning to the oligonucleotide a high immunostimulatory activity if the IFN-α score is at least 30, an intermediate immunostimulatory activity if the IFN-α score is between −5 and 30, and a low immunostimulatory activity if the IFN-α score is at most −9, when n=9; assigning to the oligonucleotide a high immunostimulatory activity if the IFN-α score is at least 1.4909×n+22.014, an intermediate immunostimulatory activity if the IFN-α score is between 0.005×n2-0.2671×n−3.5531 and 1.4909×n+22.014, and a low immunostimulatory activity if the IFN-α score is at most 0.005×n2−0.2671×n−3.5531, when n is greater than 9, wherein n is the length of the oligonucleotide.

The present invention also provides a method for assigning the IFN-α score of a RNA oligonucleotide comprising steps (a)-(c) described above.

A single-stranded RNA oligonucleotide of the length n (n≧6) is broken up into all possible 3mer motifs starting a the 5′ end. This will result in a total number of n-2 possible 3mer motifs. For example the 20mer ssRNA oligonucleotide 5′-CAGAGCGGGAUGCGUUGGUC-3′ can be broken up into the following 18 3mer motifs (5′-->3′): CAG, AGA, GAG, AGC, GCG, CGG, GGG, GGA, GAU, AUG, UGC, GCG, CGU, GUU, UUG, UGG, GGU, GUC.

Subsequently, all of the 3mer motifs are checked against the IFN-α point score matrix (Table 7).

TABLE 7 IFN−α point score matrix 3mer motif IFN−α point (5′→3′) score

Download full PDF for full patent description/claims.




You can also Monitor Keywords and Search for tracking patents relating to this Compositions and methods for immunostimulatory rna oligonucleotides patent application.

Patent Applications in related categories:

20130121969 - Biodegradable polyurethane/urea compositions - The present invention relates to biocompatible, biodegradable polyurethane/urea polymeric compositions that are capable of in-vivo curing with low heat generation to form materials suitable for use in scaffolds in tissue engineering applications such as bone and cartilage repair. The polymers are desirably flowable and injectable and can support living biological ...

20130121971 - Compositions and methods of vascular injury repair - The present invention relates to pharmaceutical compositions comprising a chemotactic hematopoietic stem cell product comprising an enriched population of CD34+ cells containing a subpopulation of CD34+/CXCR-4+ cells having CXCR-4-mediated chemotactic activity, methods of preparing these compositions and use of these compositions to treat or repair vascular injury, including infarcted myocardium. ...

20130121973 - Differentiation of mesenchymal stem cells into fibroblasts, compositions comprising mesenchymal stem cell-derived fibroblasts, and methods of using the same - Methods and compositions are provided for the differentiation and characterization of mammalian fibroblast from mesenchymal stem cells. The methods of the invention provide a means to obtain mesenchymal stem cell-derived fibroblast populations, e.g., seeded on a scaffold, which may be used in wound healing. ...

20130121970 - Method for elimination of space through tissue approximation - Acellular tissue matrix compositions for use in tissue approximation are provided. Also provided are methods for making and using the compositions to approximate tissue. ...

20130121974 - Methods of making enhanced, autologous fat grafts - Cells present in processed lipoaspirate tissue are used to treat patients. Methods of treating patients include processing adipose tissue to deliver a concentrated amount of stem cells obtained from the adipose tissue to a patient. The methods may be practiced in a closed system so that the stem cells are ...

20130121972 - Native wharton's jelly stem cells and their purification - Noncultured Wharton's Jelly stem cells and methods of their purification, storage and use are provided. ...

20130121975 - Use of mesenchymal stem cells for completely repopulating host tissue - A method of treating a genetic disease or disorder such as, for example, cystic fibrosis, Wilson's disease, amyotrophic lateral sclerosis, or polycystic kidney disease, in an animal comprising administering to said animal mesenchymal stem cells in an amount effective to treat the genetic disease or disorder in the animal. ...


###


Other recent patent applications listed under the agent :