|Home | About | Journals | Submit | Contact Us | Français|
Laser-capture microdissection (LCM) enables the selection of a specific and pure cell population from a heterogenous tissue such as tumors. Activity-based protein profiling/profile (ABPP) is a chemical technology using enzyme-specific active site-directed probes to read out the functional state of many enzymes directly in any proteome. The aim of this work was to assess the compatibility of LCM with downstream ABPP for serine hydrolase (SH) in human lung adenocarcinoma. Fresh frozen lung adenocarcinoma tissue was stained with hematoxylin, toluidine blue, or methyl green (MG). Proteome from stained tissue was labeled further with SH-directed probes, and ABPPs were determined on a one-dimensional gel-based approach. This allowed us to assess the impact of staining procedures on their ABPPs. The effect of the LCM process on ABPPs was assessed furthermore using MG-stained lung adenocarcinoma tissue. The staining procedures led to strong changes in ABPPs. However, MG staining seemed the most compatible with downstream ABPP. MG-stained, laser-captured, microdissected tissue showed additional change in profiles as a result of the denaturing property of extraction buffer but not to the microdissection process itself. LCM staining procedures but not microdissection per se interfered with downstream ABPP and led to a strong change in ABPPs of SHs in human lung adenocarcinoma.
Laser-capture microdissection (LCM) is an essential tool in -omics sciences and was first described for proteomics a decade ago.1 This tool enables the selection from a heterogeneous tissue of a specific cell population, for instance, tumor tissue, to be analyzed further.2 As with other microscope-based techniques, histological staining of the tissue section is generally required and can consequently impede downstream proteome analysis. Different staining methods suggested detrimental effects on subsequent protein separation techniques.3,4
Activity-based protein profiling/profile (ABPP) is a chemical technology that allows a direct activity read-out of members of distinct enzyme superfamilies in any given proteome through the use of enzyme class-specific active site-directed probes.5–7 ABPP probes selectively label active enzymes but not their inactive forms (zymogen or inhibitor-bound). As demonstrated by Jessani et al.8–10 for KIAA1363 activity in estrogen receptor(−)/progesterone receptor(−) breast cancers, this approach facilitates the characterization of changes in enzyme activity that occur without corresponding alterations in protein or transcript expression. Consequently, ABPP is a powerful technique to characterize disease-associated enzyme activities that may evade detection by other molecular profiling methods.
Despite the use of LCM in proteomics, coupling LCM with functional cancer proteomics such as ABPP has never been described. This new approach could lead to tumor cell-specific ABPPs determining new targets for enzyme activity modulating drugs. Coupling LCM and ABPP technologies raises the problem of staining procedure duality: On the one hand, a staining procedure may enable a good image quality and an efficient microdissection, but on the other hand, it may impair the recovered proteome with corresponding ABPPs. The aim of this study was then to assess the impact of LCM, including different staining procedures [hematoxylin (H), toluidine blue (TB), methyl green (MG)], on downstream ABPPs for serine hydrolase (SH) in human lung adenocarcinoma.
The general experimental workflow is depicted in Figure 1.
A representative piece of lung adenocarcinoma tissue from one resected patient was embedded in the cryopreservative OCT and frozen in an isopentane bath immediately after surgery. Storage was at −80°C until use. Thereafter, sections of 8 μm thickness were manufactured with a HM 560 Microm cryostat at −20°C, mounted on glass slides or MembraneSlides (MMI, Glattbrugg, Switzerland), and again, stored at −80°C until used.
Four different staining procedures were tested to visualize frozen tumor tissue for LCM: 1. single fixation in 75% ethanol (OH) for 30 s. 2. H staining with sequential immersions in 70% OH (3 s), water (10 s), Mayer's H (15 s), water (10 s), and Scott's tap water substitute (10 s); sections were then dehydrated through graded OH (70%, 95%, 95%, 10 s each; 100%, 100%, 30 s each) and cleared in xylene (2×30 s). 3. TB staining was performed by immersion of slides in 0.5% w/v TB in acetate buffer, pH 3.8 (10 s), followed by a 100% OH rinse. We used a pH-3.8 buffer, as TB has been reported to bind preferentially to nucleic acids rather than protein when used at a pH less than 4.0.11 Sections were then dehydrated in 100% OH (2×10 s) and cleared in xylene (1 min). 4. MG-stained sections were manufactured by fixation in 75% OH (30 s) and staining in 0.5% w/v violet-free MG in 0.1 M sodium acetate buffer for 10 s (pH 4.2). All of the chemicals for staining were from Sigma (Buochs, Switzerland).
In all cases, sections were finally air-dried at room temperature and stored in a dessicator on dry ice until processed further.
A 10× magnification objective of a MMI UVCut microdissector was used. Standard parameters were 75% cutting speed, 50% laser focus, and 100% laser power.
Undissected tissues were scratched from the glass slides using a #11 scalpel blade and suspended in Dulbecco's PBS. Microdissected tissues were extracted from the caps using radioimmunoprecipitation assay (RIPA) EB (20 mM Hepes, pH 7.5, 150 mM sodium chloride, 1% Nonidet P-40, 0.25% sodium deoxycholate, 10% glycerol). All of the chemicals for staining were from Sigma. Repeated incubation into liquid nitrogen and syringing up and down (0.5 ml Becton Dickinson U-100 insulin syringe) were used to solubilize proteome. Protein concentration was determined using the Bio-Rad DC protein assay (Bio-Rad, Glattbrugg, Switzerland).
SHs from solubilized proteome were labeled in the dark and during 1 h using FPs linked to a rhodamine derivate reporter tag at a 2-μM final concentration. The labeled SH trypsin was considered as a positive control. Denatured proteome before labeling was considered as negative controls.
Proteome (20 μg) was loaded per gel lane after chemical (loading dye) and heat denaturation (8 min at 90°C). Proteins were separated by 10% one-dimensional (1D)-SDS-PAGE. Fluorescence was detected by IGFS using a 9400 Amersham Typhoon scanner.
Proteome from tumor tissue sections, which underwent solubilization and labeling without any staining procedure steps, was qualified as a reference or native proteome (no). Other tissue sections were processed according to different protocols, including OH, H, TB, and MG, followed by proteome solubilization and FP labeling (Fig. 2).
FP-labeled trypsin was considered as the positive control (data not shown). Negative controls for each staining procedure could exclude unspecific SH labeling as well as any dyes' autofluorescence (data not shown).
The reference or no profile showed at least 11 distinct bands. In the case of OH, only five bands appeared. Protein profiles of H- or TB-stained sections showed only one unique band. Three bands were conserved for the ABPP from MG-stained sections.
To obtain 20 μg protein, 15 MG-stained tissue sections were microdissected randomly during 354 min, and a total of 99.5 mm2 tissue or 7602 shapes was recovered. RIPA EB was used to collect microdissected cells from the caps (Fig. 3, lane 5). This processing gave the poorest profile, with only one band left. To assess the influence of dissection time and EB on that last major variation in profile, these two conditions were mimicked separately (Fig. 3, lanes 4 and 5, respectively). Varying microdissection time did not influence the ABPP, and the use of EB led to the same profile as when the tissue was microdissected.
Until now, LCM has never been coupled to downstream ABPP. In this study, the impact on ABPPs of different LCM staining procedures as well as the microdissection process was evaluated.
On the one hand, a tissue staining procedure is encouraged, as it increases cell discrimination and microdissection efficiency.2 However, on the other hand, Craven et al.3,4 suggested a detrimental effect of staining components on downstream conventional proteome analysis. In this study, we assessed the three most commonly used LCM staining procedures for proteomic analysis. As the eosin (E) component of the H&E staining procedure was suggested to affect protein focusing on a 2D gel-based proteome analysis of cervix and kidney cortex, we used a modified procedure containing H alone.2,4 The TB staining procedure was the same as suggested by Mouledous et al.12 for microdissection of rat brain samples. In our experiments, ABPPs of tissue treated with both staining procedures did not conserve more than one common band at 70 kDa. H and TB were then not good candidates as LCM staining procedures when coupled to ABPP. A MG staining procedure consisting of 30 s in 75% OH and a dip in 0.5% MG solution was reported for LCM of nasopharyngeal carcinoma and lung adenocarcinoma.13,14 This original procedure did not give a homogeneously stained tissue all over the slide in our experiments. The reason why is not clear, as we used the same tissue type and section thickness as described by Liu and collaborators.13 Mouledous and colleagues12 described that varying stain concentration and incubation time were required to optimize the ratio between efficient LCM and detrimental effects on proteome. We then modified this original procedure by increasing the incubation time to 10 s. The profile of MG-stained tissue was the most conserved of the three dyes, with three conserved bands at approximately 70, 60, and 20 kDa.
ABPPs were strongly influenced by the three different LCM staining procedures and lost the vast majority of bands. This phenomenon was a result of the chemical denaturing properties of the staining compounds, such as OH. For example, when tissue was treated with a single OH bath, about one-half of the signal disappeared. Indeed, proteins lost their biological activity as a result of the loss of their secondary and tertiary structure following the application of OH (used in all three procedures for fixation and/or dehydration). In this context, OH-denatured enzymes, including SHs and their substrates (FPs), could no longer bind to their active site. In conventional proteomics, where the biological activity of proteins does not play such a key role, 30 s in 70% OH fixation was recommended to microdissect brain regions, as only a minimal effect on protein extraction and recovery was described.12
As the MG staining procedure seemed to have the less adverse effect on ABPPs of SH, it was used to further assess the compatibility of microdissection with ABPP.
Profiles from microdissected tissue gave a poor profile with only one band left. This led to the conclusion that the EB was responsible for the adverse effect on protein activity and not microdissection itself. These results confirmed our previous suggestions concerning the influence on profiles of protein denaturing compounds, as RIPA buffer also contains protein denaturants. We faced one of the major limitations of the use of LCM with downstream ABPP, as EB contributes to maintain a sufficient proteome yield after LCM. PCR allowed analysis down to the single-cell level in the case of genomic studies, but such amplification is impossible with a protein-based approach. Hence, proteome analysis of microdissected cells without the use of EB would lead to an extremely long dissection time. In our experience, approximately 6 h microdissection allowed a recovery of 20 μg proteome required for downstream 1D-SDS-PAGE. From the literature, recent protocols reported even days of microdissection per sample to generate enough material (>25 μg) for subsequent analysis by 2D-SDS-PAGE.15
Microdissection time was suggested to have an impact on downstream proteome analysis, and less than 1 h microdissection was suggested.2 In our experiments, mimicking variations of microdissection time from 0 to 25 min did not influence the ABPP from MG-stained tissue.
LCM is a powerful tool for isolating highly pure cell populations from a heterogeneous tissue section via direct visualization of the cells. However, staining procedure steps as well as the use of EB led to strong changes in ABPPs of SHs as a result of their denaturing components. The modified MG staining procedure seemed to be the best candidate for LCM when coupled with downstream ABPP.
There were no financial support or associations that may pose a conflict of interest in this paper.
This study has been approved by the ethical commitee of the University Hospital of Zurich.
Human subjects participating in the study have given the requisite informed consent.