This story has been updated to include comments from an Agilent official.
A team led by researchers at Utrecht University and the Hubrecht Institute has identified a molecular signature specific to Lgr5+ cycling crypt base columnar cells, a class of stem cells thought to produce all the cell lineages of the mammalian small intestine epithelium.
In a study published last month in the EMBO Journal, the researchers used a combination of transcriptomics and proteomics to identify a signature of 510 genes enriched in intestinal stem cells.
In addition to providing insight into the composition of the intestinal stem cell population and a resource for future study of these cells, the study also demonstrated the value of adding proteomic analyses to investigations that might once have been done solely via genomic or transcriptomic methods, Albert Heck, chair of the Biomolecular Mass Spectrometry and Proteomics group at Utrecht University and author on the study, told ProteoMonitor.
To define the stem cell signature, the researchers measured mRNA levels of mouse Lgr5 CPC cells and their daughter cells on microarray platforms from Agilent and Affymetrix and protein levels on a mass spec platform based on a Thermo Fisher Scientific LTQ-Orbitrap Velos. They applied a "minimum two out of three" strategy, in which they added molecules to the signature only if enrichment was detected by two out of the three platforms used – a technique that, the authors wrote, "outperformed the use of absolute cutoff values or statistical tests."
"I think this is one of the first times that [data from] two [microarray] platforms and a proteomics platform has been looked at in a combined study, and it shows that they can really complement each other," Heck said.
He noted that in the case of discrepancies between the Agilent and Affymetrix microarray platforms, the researchers were many times able to use the proteomic data to determine which result was correct.
The platforms "agree on most [transcripts], but on some they disagree, and we saw that the proteomics data would sometimes agree with the Agilent data and sometimes with the Affymetrix data," Heck said. "And so [based on the proteomic data] we looked into the probes being used on the [two] arrays, and we could define which of the probes were not so good."
The study also demonstrated that the proteomic approach was able to identify markers not as clearly picked up by the microarrays, Heck added. "Some of the [markers] came out of the proteomic data and were less accurately defined by the transcript data … so we were really able to show that the proteomic data was a valuable source for finding new stem cell markers."
Heck noted that, while he had previously believed transcriptomics retained significant advantages over proteomics, he has reconsidered this view in light of the study.
"If you do proteomics, you always think that transcripts are cheaper, faster, better, but I must say that I am not convinced of that anymore," he said. "I saw that also with transcriptomics data you have the same problems with false discovery rates and probes that don't work. Also, time-wise and expense-wise I think that proteomics is coming closer and closer to transcriptomics."
Furthermore, he said he was "surprised, pleasantly, to find that proteomics could generate data that was of the same depth and quality as the transcript data, and that the overlap or disagreement between the [proteomic and transcriptomic] platforms is no [more] different than between the two different transcript platforms."
Overall, the researchers identified via proteomics 278 genes enriched in the intestinal stem cells, 213 of which could be corroborated by one or both of the other two platforms. They identified 281 enriched genes via the Agilent microarray system – 210 of which could be corroborated – and 356 enriched genes via the Affymetrix system, of which 302 could be corroborated.
In addition to the 510-gene signature, the researchers also detected the presence in the Lgr5 CPC cells of four genes previously proposed as potential markers of quiescent/'+4' stem cells, suggesting that, in fact, these molecules are not markers of this cell population.
The small stem cell sample sizes available for study presented a technical challenge for the proteomics analysis, Heck said, noting that a multidimensional chromatography step that combined hydrophilic interaction chromatography and reverse phase chromatography enabled the researchers to increase their platform's dynamic range while dimethyl labeling enabled good quantitation even with small amounts of sample.
The researchers used 300,000 cells in the proteomic analysis, quantitating a total of 4,817 unique proteins. Heck said the platform could measure proteins in samples as small as 5,000 cells, but, he said, at that sample size they were able to reliably quantify only around 1,000 proteins.
Heck said that he has been approached by a number of scientists wanting to pursue similar studies, noting that the ability to actually measure protein expression levels as opposed to genes or transcripts was drawing researcher interest.
"If you can do proteomics, the data you are measuring is really at the protein level, and so I think people are really moving in that direction," he said.
In addition to the microarray product used in the EMBO study, Agilent, of course, also offers proteomics tools including mass spectrometers. Kathleen Shelton, senior director, Genomics at Agilent, told ProteoMonitor that the company is seeing increasing researcher interest in combining genomic and transcriptomic data with proteomics.
"We actually have an integrative biology initiative to address that [interest] because of the need that we've been seeing from different customers," she said.
Shelton cited as one particular offering the company's GeneSpring software tool, which allows researchers to integrate data from different -omics disciplines. Agilent also featured the software during its presentation at this year's American Society of Mass Spectrometry annual meeting, where company officials highlighted a recent collaboration between Agilent and University of Washington researcher Mike MacCoss that combined GeneSpring with MacCoss's Skyline SRM-MS software to allow researchers to automatically generate SRM assays for target proteins identified via other -omics studies (PM 5/25/2012).