Machine selection for LCM is largely a user-specific prerogative. There are many suppliers of LCM instruments in the marketplace. Investigators should utilize demo units on typical sample types and stains they hope to utilize when selecting a machine. It may require 2 weeks to fully assess an LCM instrument.

The isolation of RNA from LCM material used to be problematic but is now largely addressed by any one of the low sample input RNA isolation kits available on the market. We utilize the PicoPure RNA Isolation Kit (Life Technologies) for many of our studies. This kit can be utilized down to single-cell inputs, and the final elution volume is compatible with the downstream RNA-Seq library prep protocol. RNA quantitation is performed using the Quant-iT RiboGreen RNA Assay Kit (Life Technologies). RiboGreen is a very sensitive fluorescent detection dye with high specificity for RNA molecules. In some cases, your RNA concentration may be below the limits of detection for RiboGreen (around 200 cells or 1 ng). RNA integrity is assessed using the Agilent 2100 Bioanalyzer. Often, not enough RNA is isolated for RNA integrity measurement on the sample directly. We address this issue by assessing RNA integrity by proxy (see Note 4). Another approach to measuring RNA integrity is the 3′:5′ assay which utilizes quantitative real-time PCR of the GAPDH gene (Nolan et al. 2006).

RNA-Seq library prep approaches and kits are frequently changing with a new approach released on an almost quarterly basis. Therefore, the user should explore multiple kits and decide what approach is best for their particular experimental design. Keep in mind that some approaches only pertain to messenger RNA sequencing while others can do the entire transcriptome—mRNA, long noncoding RNAs, snoRNAs, etc., with the exception of microRNAs, which require a separate dedicated library preparation approach. Library preparation is conducted in two discrete steps—cDNA synthesis and sequencing library preparation. We have experienced good results from both the Ovation RNA-Seq System (NuGEN) and the SMARTer cDNA Synthesis Kit (Clontech). cDNA synthesis products are then typically prepped using the Encore Rapid Library System (NuGEN). We emphasized that the low-input RNA-Seq library preparation kit arena is rapidly evolving and is very much a personal decision that should be made by the wet laboratory researcher(s) involved in the study. Each approach is a balance of efficiency, ease of use, and cost among other variables and should be assessed by each laboratory and their particular situation and experimental study design.

The RNA-Seq approach and data analysis are also aspects that are specific to the researcher/laboratory as well as the experimental design and are beyond the scope of this chapter. In general, we aim for at least ten million unique sequencing “counts” dedicated to each sample—note that if performing paired-end sequencing, the total number of unique counts is equivalent to the number of reads divided by two. If the goal is transcript quantitation and subsequent differential expression analysis, then this depth of sequencing per sample should be sufficient. For deep transcriptome sequencing studies—those that have a primary goal of sequencing even very rare transcripts in the cell—more sequencing depth should be utilized, with some studies approaching 100 million unique counts. Read length and paired-endedness can also be tuned based on experimental design. For a standard differential expression study, we utilize paired-end, 50-mer reads as we find that they provide acceptable mapping statistics and transcriptome reconstruction capabilities as well as being cost- and time-effective. Designs with different goals may want to use longer reads.

RNA-Seq data analysis is another example of a fast-moving discipline. Currently, the main debate is model-based quantitation approaches versus count-based approaches. We have utilized both with acceptable results. We utilize bcbio-nextgen as the basis for our best practices pipeline for the next-generation sequencing analysis. This open-source python toolkit aids in supercomputing resource utilization efficiency. It includes RNA-Seq pipeline modules to perform adapter trimming, read alignment, quality control assessments, and transcript quantitation. The bioinformatic approach to transcript quantitation is an active area of the field, and new approaches are frequently reported and must be tested.

An example of LCM RNA-Seq data comparing two groups for one gene of interest is illustrated in Fig. 1b. In this study, we exposed adult rats either to their home cage conditions (CONT) or to an electroconvulsive shock (MECS) at a level known to activate approximately 100 % of the neurons in the hippocampal formation. Six animals in each group were examined following sacrifice at 60-min postexposure to MECS. We utilized LCM to capture either the cell soma compartment or the corresponding neuropil from three regions within the hippocampus—DG, CA1, or CA3. The top panel of Fig. 1b demonstrates the box and whisker plot results for each subregion, compartment, and experimental group for the Arc (activity-regulated cytoskeleton-associated protein) transcript. The Arc is a well-studied immediate early gene that is known to be activated by neuronal activity and MECS exposure. In each case, you can see upregulation of the Arc transcript in the MECS-treated animals. The bottom panel of the figure is a demonstration of the difference of the median counts between the two experimental groups for each subregion and compartment. Note the larger upregulation of the Arc transcript in the neuropil than in the soma in each subregion, expected based on the role of Arc at the synapse. This is particularly pronounced in the DG.