RNA Interference Pathways and Therapeutic Exploitation - RNA Interference from Biology to Therapeutics - page 12

strand has the potential to function as a miRNA, thereby potentially affecting
hundreds of mRNAs) [ 62 ]. A wealth of modification types have been tested to circum-
vent these shortcomings including modification of the phosphodiester backbone
(e.g., by phosphorothioate linkages (PS) [ 131- 138 ]), substitution of the ribose 2¢ -
OH group [e.g., by 2`- O -Methyl (OMe), 2¢-Flouro (F), 2¢ -Methoxymethyl (MOE)],
or using bridged nucleic acids (such as LNA [ 139 ] , carbocyclic-LNA [ 112, 140 ] ,
ENA [ 141 ] , carbocyclic-ENA [ 112, 140 ] , and oxetane (OXE) [ 112, 142 ] ), substitu-
tion of the ribose unit with six carbon sugars (such as ANA, HNA, 2¢ -F-ANA, and
CeNA [ 112- 117 ]), by disrupting the ribose ring structure (such as in unlocked
nucleic acids (UNAs) [ 124, 143 ]), or modification of nucleoside bases (5-iodo-,
2-thio- and pseudouracil [ 120, 131, 144 ]) (for a more comprehensive review of
chemical modification types, see [ 126 ] ). Modi fi cations, however, must be compat-
ible with siRNA function, and positional tolerances for siRNA modifications have
been well established by empirical testing; by rule of thumb, the guide strands 5¢
phosphate, 5¢ end (seed region), and central positions are particularly sensitive,
especially to several or bulky modifications. In contrast, the entire length of the pas-
senger strand, the 3¢ end, and overhang of the guide strand are fairly tolerant and can
be chemically modified to enhance siRNA performance [ 112, 126, 131- 136, 141 ] .

1.4.4 Improving siRNA Nuclease Resistance

Most unmodified siRNAs are degraded by ribonucleases within minutes in mam-
malian serum [ 46, 112, 131, 135, 145, 146 ], and enhancing siRNA stability has long
been considered essential for siRNA function in vivo. Indeed, extensive chemical
stabilization was found essential for successful silencing in mouse livers upon
low-pressure intravenous injection of naked siRNA, a strategy relevant to siRNA
therapeutics [ 147 ]. Yet, only few extensively or fully modified siRNAs are reported
to be both highly stable and potent [ 147- 149 ] as extensive chemical modi fi cation of
siRNAs will typically reduce their activity [ 112 ] . Instead, moderate modi fi cation
levels using phosphorothioates [ 131- 136 ], thermostabilization of the siRNA duplex
stem by LNA [ 112, 119, 135, 136, 150 ] , OMe or 4 ¢ thioribose [ 151 ] , or combinations
of these modifications have been successful in creating stable and potent siRNA for
applications in vivo [ 46, 152, 153 ] .
Recently, suggestions to modify only nuclease hypersensitive positions in the
siRNA were put forward to limit modification levels and preserve silencing [ 154,
155 ] . As modi fi cation of siRNA 3 ¢ overhangs are very well tolerated by the RNAi
machinery, numerous modification types will provide 3¢ exonuclease resistance and
modestly enhance siRNA stability [ 112, 131- 136 ] . As most dsRNA-speci fi c endori-
bonucleases are preferentially recognizing UpA, UpG, and CpA dinucleotide motifs
[ 154- 157 ], further siRNA stabilization by modifying these vulnerable positions, for
example, by 2¢-OMe, is a straightforward approach to significantly enhance stabil-
ity [ 154, 155 ] .
It is worth noting that the benefits of enhancing siRNA nuclease resistance
seem primarily to originate from effects during siRNA delivery prior to siRNA