In this paper, we have described a protocol that enables rapid and quantitative detection of RNA targets via RNA FISH. We found that alcohol-based fixatives provide the necessary probe accessibility for rapid hybridization via increased probe concentration, potentially enabling hybridizations in as little as 30 seconds.
Our experiments show that there is a straightforward tradeoff between concentration of probe and the speed of hybridization. We have found that increasing probe concentration by 20X compared to our normal overnight protocol yields reliable RNA FISH results after just 5 minutes of hybridization. At first glance, this increased probe concentration may not seem economically viable, considering the increased use of probes (which are the most costly reagent in the RNA FISH protocol). However, we note that because of the decreased time for drying, our protocol uses roughly 10 fold less hybridization solution for the hybridization itself, greatly mitigating such concerns. We believe that the ultimate choice of how much probe to use and how fast to drive the reaction will depend on the specifics of the application at hand. In some cases, getting a hybridization time of 5–10 minutes may be perfectly fine, in which case one may not need to use large concentrations of probe. However, in some situations, such as during a surgical procedure, the decreased hybridization times may be a benefit that outweighs the cost of increased probe usage.
Of course, even with rapid hybridizations, we have not addressed the issue of the imaging time itself. Typically, image acquisition may require taking image stacks from multiple positions on the slide to obtain enough cells’ worth of image data to make statistically significant claims about differences in gene expression. Currently, doing so could take on the order of 10–20 minutes per condition. However, we believe that technical advances can reduce the time required for both image acquisition and analysis by at least an order of magnitude. In such a case, one could envision comparing gene expression in two samples in well under 30 minutes from living cells to quantified data.
It may be useful here to make a comparison to other methods such as RT-qPCR. RT-qPCR is the current gold standard for gene expression analysis, widely considered to be the most accurate method for quantifying gene expression to date. It has many benefits, including high dynamic range, low cost per reaction, and the ability to parallelize in 96-well plate format. The qPCR itself usually takes on the order of 1–2 hours to complete, but if one includes both RNA extraction and setup time, the total time required is probably closer to around 3–4 hours. (These extra steps also increase the cost of the experiment as well.) We believe that with rapid hybridization, RNA FISH competes favorably with RT-qPCR on most counts. With respect to quantification, our method provides accurate, absolute counts of gene expression of 3 to 5 genes in individual cells without the explicit need for normalization. Since RNA FISH is a direct detection scheme without any amplification, we are able to detect even small fold-changes with high precision, differences that would be hard to measure accurately with RT-qPCR, at least not without a large number of replicates. The cost per reaction is probably dominated by the cost of the probe, which is currently around $300–$600 per probe set for 10,000 hybridizations ($0.06 per reaction) and is thus comparable to a molecular beacon or Taq-man RT-qPCR probe. Of course, costs of labor, equipment and other reagents are variables that are hard to predict, but will be of the same order of magnitude, although we note that the labor required for RNA FISH is probably lower, whereas the cost of an automated microscope is admittedly higher than most qPCR machines. Both the accuracy and cost comparisons to RT-qPCR were valid even with overnight RNA FISH.
The time required for previous iterations of RNA FISH, however, was considerably longer than for RT-qPCR, and our new method alleviates that discrepancy. If one is just comparing the expression of a few genes in a few conditions, then we believe our method is unequivocally several times faster than RT-qPCR, especially when one factors in RNA extraction and setup time. For analyzing larger numbers of genes in parallel, though, the imaging time will become a factor. If one assumes 5–10 minutes per condition and triplex RNA detection, then analyzing, say, 20–30 genes could require up to 2 hours. With advances in high throughput imaging, we anticipate that one could reduce this time by an order of magnitude, thus further increasing the speed advantages.
Another major advantage of RNA FISH is that it also provides single cell information, something that is much more difficult to obtain with single cell RT-qPCR approaches. This enables one to measure variability in gene expression from cell to cell. Since the measurements yield absolute numbers of RNA, the measurements do not necessarily require normalization to an internal control (such as GAPDH), although one could perform such an analysis if one wished through multiplexing. Normalization can be difficult to perform with RT-qPCR approaches, since one typically uses all the material for a single qPCR reaction, leaving none for further normalization.
Furthermore, RNA FISH also provides spatial information on the localization of RNA. Such information is important both for examining differences from cell to cell within a tissue and even subcellular spatial localization. In tissues, one can easily identify particular cells by labeled RNA specific to those cells with one color and then looking at the gene of interest in another color. Subcellular information can be of particular importance for RNA that localize to particular regions of the cell, such as many non-coding RNA, in which case RNA FISH can reveal much about its behavior.
We have also shown that one can perform iceFISH and SNP FISH to visualize chromosomes and single base changes, respectively, with rapid hybridization. Such techniques could be useful for rapidly diagnosing chromosomal abnormalities and for rapid genotyping of particular single nucleotide variants.
In summary, our method for rapid hybridization results in orders of magnitude improvements in hybridization time for single molecule RNA FISH, enabling a new set of high throughput and rapid diagnostic applications.