MicroRNA Detection Methods for Mammalian Cell Lines and Their Applications in Therapeutic Protein Production - MicroRNAs as Tools in Biopharmaceutical Production - page 106

with small RNA enrichment typically by PAGE followed by RNA quality check.
Then small RNA libraries are constructed with 5 - and 3 - sequencing tags incorpo-
rated, and RT reactions are performed to convert the tagged RNA to cDNA. Lastly,
the fraction with adequate size is isolated and sequenced using established methods
such as “sequencing by synthesis” (SBS) isothermal bridge amplification (Illumina)
or bead-based emulsion PCR (ABI SOLiD) (Tian et al. 2010 ). NGS is particularly
valuable for miRNA discovery work due to the simultaneous availability of sequence
and quantity (read counts).
Additionally, NGS is the only method that can effectively detect isomir sequences
that represent variations of the mature miRNAs from the same precursor, an important
new aspect in miRNA biogenesis (Morin et al. 2008 ).
The first NGS publication in CHO cells emerged in 2010 (Johnson et al. 2011 ).
Johnson and colleagues reported sequences of 350 mature miRNA species from four
CHO lines in various culture conditions. The NGS was performed using the Illumina
platform from six cDNA libraries, and the annotations were based on homology with
human, mouse and rat. The majority of annotations were made based on alignment
with mature miRNAs in miRBASE 15.0. Differential expression studies between the
cell lines or conditions were not reported in this publication. Nevertheless, this was
the first report of NGS in CHO miRNA studies. Hackl and colleagues published the
second study on NGS sequencing of CHO miRNA transcriptome less than a year
later. They reported 387 annotated mature miRNAs from six CHO lines in various
culture conditions (Hackl et al. 2011 ). While choosing Illumina as the NGS platform,
they took a different approach in annotation by using the stem-loop for conserved
miRNA identification. Sophisticated bioinformatics procedures were used to predict
122 novel miRNA. This work is the first to use NGS for differential expression of
miRNAs in CHO in serum and serum-free culture conditions. A third and very recent
publication successfully used the CHO whole genome sequencing data, in addition
to using miRBASE sequences, to identify 190 conserved miRNA hairpin precursors
(Hammond et al. 2011 ). The authors performed differential expression analyses as
well, and reported more than 80 % (158) of miRNAs as significantly up- or down-
regulated in at least one recombinant protein producing cell line comparing with the
parental CHO K1.
The SoLiD platform, on the other hand, is also employed for NGS miRNA
discovery and profiling for human and mouse (Schulte et al. 2010 ), but has not
yet been reported in CHO cells. Tian and colleagues reported the potential bias in
miRNA quantitative measurements that could be attributed to cDNA library prepa-
ration protocol differences. They observed in miRNA expression profiling from an
unsequenced genome (sheep) higher variability in read lengths and in end sequence
reads, also higher occurrence of end secondary structures when using the SoLiD
method (Tian et al. 2010 ). Some researchers successfully adapted a SOLiD protocol
to suit the SBS platform (Buermans et al. 2010 ).
In conclusion, NGS may require more validation against some of the more mature
methods discussed in this review. Git and Tian both reported good correlation of the
SBS method with qRT-PCR. It is still considered good practice to verify specific
findings from high-throughput studies, NGS in particular, by qRT-PCR. The three
major methods for detecting miRNA in CHO are summarized in Table 7.2 . As with