A recent study has demonstrated, for the first time, that it's possible to detect—and automatically measure—key activity indicators of cancer cell signal transduction pathways in intact tissue. The development, based on multiplexing and multispectral imaging, is promising for cancer pathogenesis and drug discovery.
Agoal of many researchers is to achieve better knowledge of the mechanisms of various signaling pathways in order to discover the specific causes of a given disease. Being able to reveal correlations between signaling pathway activity and clinical outcome would better support target validation, trial design, patient selection, response assessment—and, if trials are successful, the diagnostic component of theranostics.
The predictive power of protein expression measurement depends on the precision and accuracy of tissue analysis tools. Many techniques deployed today, however—such as those based on microarray detection, or analysis of sample lysates—provide data that are, in fact, averages from volumes of tissue, including many cells not of interest. These methods obfuscate key proteomic information that resides at the cellular level and relates to the signaling states of individual cells.
Signal transduction pathways in cancer
During the course of tumor progression, cancer cells acquire a number of characteristic alterations. These include the capacity to proliferate independently of exogenous growth-promoting or inhibiting signals; the tendency to invade surrounding tissues and metastasize to distant sites; the penchant for eliciting an angiogenic response; and the ability to evade mechanisms that limit cell proliferation, such as apoptosis and replicative senescence. These properties reflect changes in cellular signaling pathways that in normal cells control proliferation, motility and survival.
Many proteins under investigation as possible targets for cancer therapy are signaling proteins that are components of these pathways. The nature of these signaling pathways and their roles in tumorigenesis are the subject of intense study by pharamaceutical companies—motivated by the hope that progress in understanding cancer as a disease will accelerate drug development. This is a broad research topic, and the task of identifying relevant pathways, understanding them and demonstrating correlation with outcome is a challenging one. An additional level of complexity arises from the fact that the interrelationship between pathway proteins and their localization—rather than the mere presence of a protein—often helps characterize the pathway.
Targeting popular pathway markers
AKT, ERK and S6 are three widely studied pathway markers that play a vital role in cancer pathogenesis.
AKT has recently been found to play a paradoxical role: On the one hand it increases cancer cells' survival capability, while on the other hand it blocks their motility and invasion abilities, thereby preventing cancer from spreading.3 It had been presumed that cancer cell death could he hastened by inhibiting AKT, but it is now clear that the role of AKT must be understood further, so as not to promote metastases by inhibiting AKT expression (Fig. 1).
Activation of the ERK pathway promotes cell division. This pathway is often up-regulated in human tumors and is thought to fulfill multiple roles in the acquisition of a complex malignant phenotype. Accordingly, a specific blockade of the ERK pathway is expected to produce not only an anti-proliferative effect, but also in anti-metastatic and anti-angiogenic effects in tumor cells. Recently, potent small-molecule inhibitors targeting components of the ERK pathway have been developed. Among them, BAY 43-9006 (Raf inhibitor), and PD184352, PD0325901 and ARRY-142886 (MEK1/2 inhibitors) have reached the clinical trial stage. The combination of ERK pathway inhibitors (cytostatic agents) and conventional anticancer drugs (cytotoxic agents) might provide an excellent basis for the development of new chemotherapeutic strategies against cancer.
Finally, S6 is a ribosomal protein involved in translation. It is thought to play an important role in controlling cell growth and proliferation. It is a major substrate of ribosomal protein s6 kinase and plays a role in regulating translation of RNAs that contain an RNA 5' terminal oligopyrimidine sequence. It is regulated by ribosomal S6 kinase.
Automated multiplexed tissue cytometry
Detecting pathway markers using conventional histology or immunofluorescence is a challenge because getting a full understanding of the pathways involved and the phenotypes requires observation of many markers simultaneously (i.e., multiplexing). Conversely, conventional multiplexing techniques, such as microarrays and flow cytometry, fail to provide the contextual information needed to confirm intracellular localization (which is also a requirement for confirming pathway state).
What is needed is simultaneous measurement of multiple proteins, on a per-cell basis, set in the context of the original anatomy.
Technology now offers us the opportunity to access this level of information using an effective, practical and reliable platform for cytometric analysis of intact tissue sections. This can be conceptualized as 'tissue cytometry.' The platform supports preclinical and clinical studies through the integration of multiplexed immunohistochemical (IHC) or immunofluorescent (IF) labeling strategies, robotic slide handling, and automated multispectral image acquisition and analysis. Such a platform integrates:
- easy-to-implement multiplexed staining protocols
- an automated slide analysis system able to isolate marker signals from one another and from autofluorescence
- pattern-recognition based image-analysis software for automatically segmenting images and extracting data from cells-of-interest.
For a recent study that aimed to understand signaling activity and assess the AKT, ERK and S6 pathways, a group of researchers representing various organizations developed a staining panel to target phosphoeptitopes of AKT, ERK and S6. We used antibodies of three different isotypes, with secondaries conjugated to Alexa Fluor (Invitrogen Corp.) fluorophores (488, 555 and 647); in addition, we used DAPI as a counterstain.1 Our imaging platform consisted of multispectral imaging hardware (either the manual Nuance system that mounts on research-grade microscopes, or the fully integrated Vectra system), together with inForm advanced image analysis software (all by Cambridge Research & Instrumentation, CRi, Woburn, MA). (Figs. 2, 3)
Staining protocols were optimized using tissue microarrays (TMA) generated from lung tissue samples prepared using methods designed to preserve phosphoepitopes. We generated spectral unmixing libraries with single-stained control samples. Image analysis algorithms were trained to segment tissue regions (e.g., malignant and normal epithelia, stroma, necrosis, etc.) and then cells and cell compartments within tumor regions, to extract per-cell data for cytological analysis.
The research team validated the platform's ability to quantitate changes of phosphoepitope expression using cell blocks of cell lines treated with the relevant inhibitors. Pilot studies of the TMA and cell lines reveal robust and specific signal levels, localized to tissue and cellular structures appropriate for the target molecules, despite being applied in four-plex. Pattern recognition-based, automated image analysis algorithms reliably detected tumor cells and segmented associated cellular compartments, after having been trained on less than 10% of pilot study images. Tissue segmentation accuracy was estimated at greater than 90%, based on visual review by pathologists.
Data can be further analyzed using cluster analysis. Signals from individual cells are divided into quartiles of pS6, pAKT and pERK expression, and placed into 64 bins (Fig. 4).
Separating multiple markers in the same tissue
Quantitative, independent, and specific multi-label protocols work in conjunction with easy-to-use multispectral imaging systems and advanced learn-by-example software to greatly accelerate clinical and pre-clinical studies.2
Multispectral imaging captures information from many narrow wavelength bands, instead of simply one band, for each fluorescence emission filter. This added information enables automated multispectral tools rapidly to isolate label emissions from each other and also from autofluorescence, which commonly obscures weak but important signals in formalin-fixed, paraffin-embedded tissue sections. This ability to separate signals applies even when, as is often the case, they are spatially and spectrally overlapping. Conventional fixed-bandpass filter approaches cannot do this.
The principal scientific benefits of slide scanning, as offered by the integrated Vectra system (which offers robotic slide handling, automated image acquisition and the ability to work with IHC or IF tissue-sections or TMAs), for instance, is realized when the terabytes of imagery created can be processed automatically to produce relevant and useful information (Fig. 5). This is true whether or not the imagery is multispectral. The learn-by-example image analysis algorithms in inForm software provide trainable tissue segmentation, a key component of automated image analysis and data extraction. The software can be trained to differentiate relevant tissue regions (e.g., malignant and normal epithelia, stroma, necrosis, etc.) and segment cellular compartments (nuclei, cytoplasm and membrane) to allow for detailed, spatially resolved multiparameter quantitation.
Such automation is especially powerful for multispectral imaging because it enables rapid tissue cytometry rapidly on a large scale, using many markers at once. This in turn enables a better understanding of the mechanism of disease, and potentially better, more precise avenues of treatment.
- Hoyt C, et al., Tissue cytometry platform for quantitating multiple signaling pathways proteins in intact tissue sections. Poster at 2009 AACR-NCI-EORTC conference.
- Levenson R M, et al., Med. Diagn. (2008) 2(9):1067-1081.
- Yoeli-Lerner M, et al., Mol Cell. (2005) Nov 23;20(4):539-50.
CLIFFORD C. HOYT is vice president and chief technology officer of Cambridge Research & Instrumentation (CRi), Woburn, MA; www.cri-inc.com; e-mail: firstname.lastname@example.org.