91影库

It is well established that tumors, even those of the same type, exhibit differences in genetics and morphology. This heterogeneity not only exists for tumors from different patients but also across regions within the same tumor. The latter, termed intratumoral heterogeneity, is of particular interest because it directly affects diagnosis and prognosis.

This microscopic photo shows tumor cells from a fine needle aspiration cytology smear of a liver mass. Tumor cells exhibit nuclear enlargement, opened chromatin and multiple nucleoli.Courtesy of Jian-Hua Qiao/NIH flickr

and others at the study intratumoral heterogeneity. “This has important implications for tumor development because certain cells might be more aggressive than others,” Beck said.

Most studies have looked at intratumoral heterogeneity at the genomic level. It remains largely unknown to what extent the local proteome of tumors intrinsically varies. In in Molecular & Cellular Proteomics, Beck and a group of researchers at the EMBL attempt to answer this question. “We were interested to find out if the proteins contained within individual cells of the tumor are the same or different,” Beck said. Since heterogeneity in the tumor microenvironment, such as the presence of a neighboring blood vessel, may drive genetic changes, he reasoned that it might also be reflected on the level of proteins.

The researchers looked at hepatocellular carcinoma, or HCC, the most common type of liver cancer. They used HCC samples biopsied from patients and then formalin-fixed and paraffin-embedded on microscope slides. Such samples, commonly referred to as FFPE, preserve the integrity of the tissue architecture of the original tumor, allowing the researchers to study the spatial differences in protein expression.

FFPE samples, however, present technical challenges for proteomic analysis, particularly because only a limited amount of proteins can be extracted. To overcome this problem, the researchers developed a novel method that efficiently extracts proteins from FFPE samples. To profile the spatial expression of proteins, they combined this method with a technique called laser-capture microdissection to carve out microscopic regions within the tumor. The extracted proteins then were run on a mass spectrometer for identification.

The researchers first looked at the differences of protein expression between the tumor tissue and the normal tissue immediately adjacent to it. They detected consistent changes of multiple proteins known to be associated with HCC. More importantly, they also identified a few proteins that previously were not known to be HCC-related, opening possibilities for candidate biomarker development. Among these were members of the NADH dehydrogenase complex I. This finding was striking because the researchers showed that the changes were not reflected at the gene expression level, underscoring the importance of proteome profiling.

The researchers went deeper and dissected different regions within the tumor bulk. Here they found significant variations in expression of multiple proteins between areas from the center and the periphery of the tumor. “We could show that even between seemingly identical cells, with the same morphology and the same genome, there are surprisingly pronounced differences on the level of the proteins,” Beck said.

These spatial differences of protein expression include proteins that have previously been identified as HCC biomarkers. “In our analysis, we saw that even proteins that have been proposed as such biomarkers are not evenly distributed across the tumor,” Beck said.

This finding is of immediate clinical importance. Only a small fraction of a tumor can be obtained in a diagnostic or pretreatment biopsy, and thus the region of withdrawal could have a direct impact on the acquired expression profile. “It is possible that the tissue sample taken during biopsy does not reflect the actual state of the entire tumor,” Beck said.

Beck believes the method developed in this study not only allows for studying intratumoral heterogeneity but also can improve cancer proteomics research in general. “Proteomic intratumoral heterogeneity should be taken into account for future cancer research,” he said, “for example in the design of biomarker discovery experiments.”