Skip to main content
Premium Trial:

Request an Annual Quote

Mann Lab Develops SILAC-based Approach for Use in Tissue Samples


This story originally ran on April 8.

By Tony Fong

Since its development about eight years ago, stable isotope labeling by amino acids, or SILAC, has become a standard method for quantitative proteomics.

But the technique developed in Matthias Mann's laboratory has been largely limited by the fact that it can be used only in cell cultures — not tissue or body fluid samples.

In a study published April 4 in Nature Methods, though, Mann and his colleagues at the Max Planck Institute for Biochemistry describe a method they have developed to expand the use of SILAC for tissue samples and eventually to researchers who may not be specialists in mass spectrometry.

The method is based on what they call a "super-SILAC" mixture, which combines five SILAC-labeled cell lines with human carcinoma tissue, resulting in hundreds of thousands of isotopically labeled peptides "in appropriate" amounts. The peptides, the authors said, can thus be used as internal standards for mass spectrometry-based analyses.

In an interview this week, Mann said that the "trick" to the method is that SILAC was used as an internal standard and "by choosing some cell lines that correspond to the tumor or some tissue that we want to represent, we have found that we can get counterparts to the proteins expressed in the tumor in this super-SILAC mix."

While some work had been done using SILAC on tissue, the broad quantitative ratio distribution between cell lines and tissue has rendered such approaches for accurate quantification challenging.

"Therefore, we explored the use of a mix of multiple SILAC-labeled cell lines as [an] internal standard for comprehensive human tumor tissue proteome quantification," the authors wrote.

In the initial phase of the development of the method, they SILAC-labeled the breast cancer cell line HCC1599 and "mixed the lysate with the lysate of mammary carcinoma tissue from an individual with grade 2 lobular carcinoma," they wrote in the study.

After digestion of the protein mixture and separation of the peptides into six fractions, they analyzed each fraction by online reverse-phase chromatography coupled with high-resolution, quantitative mass spectrometry. This resulted in the quantification of more than 4,400 proteins at least once in triplicate analysis.

But because "the ratio distribution was broad and bimodal," they concluded that the proteome of a single cell line does not "adequately represent the tumor tissue proteome and therefore cannot be used for accurate quantification."

The original analysis contained 755 proteins with more than four-fold higher expression in the tumor than in the cell line.

To improve accuracy, they selected four breast cancer cell lines differing in origin, stage, and estrogen receptor and ErbB2 expression, and then added a normal mammary epithelial cell type "to expand the tumor subtypes that can be represented," Mann and his colleagues wrote.

The idea was that by taking different cell lines from patients with different categories — such as estrogen receptor positive and negative, or HER2-negative and -positive — and mixing them together, it wouldn't matter where the tumor comes from.

"It would always be well-represented in the super-SILAC mix," Mann said.

By using such a strategy, he and his co-researchers achieved unimodal distribution and found that 90 percent of quantified proteins — 3,837 out of 4,286 — were within a four-fold ratio between the tumor and the super-SILAC mix.

[ pagebreak ]

In addition, the quantitative distribution was "much narrower" with 76 percent of the proteins in the carcinoma and the super-SILAC mix differing by no more than two-fold. The narrower ratio distribution in the new standard resulted in "superior" quantification accuracy. "Therefore the super-SILAC mix is superior to a single-cell line internal standard and can be used to quantify several thousand proteins with an error of a few percent," the authors said.

According to Mann, his group has been working to expand the applicability of SILAC since it was first described in 2002. In addition, they have been studying breast cancer development, and in a current project they are researching breast cancer progression in different cell lines derived from different stages of mutations in the disease.

"And out of this project this idea came that if we mix those cell lines, we might be able to represent the tumor itself," Mann said.

The super-SILAC approach is "very robust and very cheap … and is very applicable to non-specialist laboratories," he said. While the method continues to be developed, "it can be very easily applied [and] it has some advantages compared to other methods that you can use for quantifying human tissue."

Such other methods include isobaric mass-labeling technologies such as iTRAQ and TMT-based methods. But, Mann said, those technologies can be "quite costly" when used on large amounts of patient sample.

The approach he and his colleagues developed is "inherently very accurate because you quantify with all the peptides and with many scans over each peptide, whereas in iTRAQ you typically quantify with one MS/MS [run] or just a few for each peptide," Mann said. "And a lot of times, you have to discard [some results] because there were two peptides that you fragmented together, and then you cannot use iTRAQ ratios. So there are a lot of practical issues that completely don't exist with SILAC."

There is no complicated or expensive chemistry involved, he added. The super-SILAC mix can be characterized at great depth in specialized labs, such as his own, Mann said, but those who don't need that depth of characterization can download the database from Mann's lab for their own purposes.

SILAC remains inappropriate for human body fluid samples, however. "At this moment we don't have a labeled standard that would correspond to body fluid." While SILAC has been used to quantify proteins in mice, doing the same in humans requires a person to ingest labeled amino acids, which Mann said is unethical.

Continuing work to further develop the super-SILAC approach has focused on practical issues, such as use of the method in small tissue-sample amounts.

"And that looks very doable," Mann said, adding that the researchers have used frozen material and paraffin-embedded samples. While that work has meant using the approach on clinical samples, he and his group have not performed a clinical study using a super-SILAC mix.

For non-clinical proteomics research, the method simplifies the "normal use of SILAC." In cell-line work, a researcher could mix the super-SILAC with the case and the control.

"So they don't need to know anything about SILAC and labeling their cells [or] mass spectrometry," Mann said. "They get their results and each of them is labeled against the standard, so then they have the changes from the control to the case automatically.

"It actually simplifies the normal SILAC," he added. "That's another direction that we're going, for use of SILAC for normal proteomics work. That will make it that much more accessible to a wider range of people"

In the work described in Nature Methods, the authors quantified almost 100 phosphorylated peptides without any enrichment steps. Since the submission of their manuscript, Mann and his co-authors have performed a specific experiment in which they enriched for phosphorylated peptides and "we can see that we can quantify thousands of phosphorylation sites this way. That then allows you to address the signaling stages of the tumor," Mann said.

[ pagebreak ]

Yoshiya Oda from the Biomarkers and Personalized Medicine Core Function Unit at Japanese pharma Eisai has been using SILAC to research specific interactions among non-specific binders. In an e-mail, Oda said that the importance of the super-SILAC mix approach is that it allows for quantification of very low-abundance and small fluctuating proteins.

Such proteins may have a significant impact on disease, he said, "Therefore accurate quantitation is very important. … Now we need to know small differences in protein expression levels."

"The advantage of metabolic labeling … is to normalize the variation from the beginning of sample [processing] to the end of analysis," he added.

Matthias Selback, a group leader in intracellular signaling and mass spectrometry at the Max Delbrück Center for Molecular Medicine in Berlin, said the main advantage of the super-SILAC method is that it will result in more accurate quantification and the quantification of more proteins, compared to iTRAQ and TMT.

"By having higher accuracy you can hope that it will facilitate identification of biomarkers," he said. "I'm not sure whether this will directly lead to useable biomarkers, but it's certainly a prerequisite for biomarker identification."

Anouk Emadali, from the Atomic Energy Commission in Grenoble, France, where she is a member of the Laboratory for the Study of Dynamics of Proteomes, called the study convincing in its demonstration of the overall performance of the super-SILAC approach.

The approach also opens a new path for tissue-based proteomics research, she added. "Regarding cancer research, it is obviously more relevant to study tumor tissues … than cell models."

However, she also said that it is unclear whether or how the approach would "answer clinically relevant questions," especially the identification of relevant cancer biomarkers.

"The authors state that the super-SILAC mix can be used as an internal standard to compare both different tumors and normal versus tumor tissues," she said in an e-mail. "However, the super-SILAC mix is not optimized for normal tissue, and [the example given in the study in which the authors applied the method to breast cancer tissue] is not developed enough to convince [me] that it can be used to compare normal and tumor tissue (especially to lead to the identification of low-abundance cancer markers.)"

While Mann and his colleagues directed their study at brain cancer and breast cancer tumors, he said they are thinking of developing a super-SILAC approach for other cancers, such as prostate and blood cancer.

The technology, "in principle," should be effective for quantitative studies of any tissue, he added.

He and his colleagues are in the initial stages of commercializing the technology and have received interest from undisclosed vendors, Mann said.