In Vivo Gene Silencing by Virally Delivered MicroRNA - Viral Vector Approaches in Neurobiology and Brain Diseases - page 245

for comparison with active miRNAs, which can be some kind of
universally “nonsense” miRNA not expected to affect any gene.
Presumably, endogenous miRNA should silence target genes.
However, endogenous miRNAs can affect multiple genes to which
they have only partial complementarity [ 11 , 64 - 66 ]. Obviously,
this is undesirable in an experimental setting. Furthermore,
miRNAs with poor complementarity to the target mRNA typically
induce a smaller degree of gene knockdown than fully complemen-
tary miRNAs [ 11 , 14 ]. Additionally, the mismatches in the miRNA
stem can also infl uence the knockdown effi cacy. For example,
Invitrogen claims that replacing 5-nt/3-nt native internal loop
with the artifi cial 2-nt UU loop in the murine mi-155 improves the
knockdown rate [ 67 ]. Therefore, endogenous miRNA may not be
the best suited to be used simply as tools to silence genes of inter-
est rather than to study their own role in brain functions. If an
endogenous miRNA construct is chosen for gene silencing experi-
ment, it is obligatory to validate the miRNA to see if it answers
specifi c requirements and is a good fi t for a specifi c experimental
situation such as knockdown of brain genes in living animals.
Engineered miRNAs, although based on the endogenous
miRNA backbone, usually incorporate some improvements into
the stem design and are fully complementary to their targets to
induce mRNA degradation as well as translational repression.
Nevertheless, miRNA sequences need to be tested and validated
experimentally. Invitrogen provides its predesigned miRNAs as a
package of 4 essentially admitting that even predesigned sequences
are not likely all to be active, at least, not to the desired extent.
When designing custom miRNA, it is a good idea to select 3-4 and
test for activity. Since miRNA cloning vectors are also mammalian
expression vectors, the fi rst round of testing can simply be con-
ducted by transfecting cultured cells with the miRNA-encoding
vectors and testing for the knockdown of endogenous target pro-
teins. Importantly, the cells should be derived from appropriate
species. If such cells are not available or available cells do not
express the protein of interest, an alternative is to cotransfect the
cells with the vector encoding the protein of interest of the appro-
priate species and the miRNA-carrying vector (Fig. 4a ). Since
transfection effi cacy varies from well to well and from plate to plate,
it is important to perform each transfection in duplicates or tripli-
cates and repeat the experiments in several plates. For some reason,
transfected proteins are easier downregulated by coexpressed miR-
NAs than endogenous proteins, so this is not as good a test as that
involving endogenous proteins. Furthermore, if cultured cells
express a full complement of closely related proteins, it is possible
to evaluate likely off-target effects of miRNAs, whereas this is not
always possible when using transfected target protein. However,
this method allows for selection of most active miRNA sequences
to be further tested in more “native” setting. Our own experience