|Home | About | Journals | Submit | Contact Us | Français|
Laser capture microdissection (LCM) allows microscopic procurement of specific cell types from tissue sections that can then be used for gene expression analysis. In conventional LCM, frozen tissues stained with hematoxylin are normally used to the molecular analysis. Recent studies suggested that it is possible to carry out gene expression analysis of formaldehyde-fixated paraffin embedded (FFPE) tissues that were stained with hematoxylin. However, it is still unclear if quantitative gene expression analyses can be performed from LCM cells from FFPE tissues that were subjected to immunostaining to enhance identification of target cells. In this proof-of-principle study, we analyzed by RT-PCR and real time PCR the expression of genes in factor VIII immunostained human endothelial cells that were dissected from FFPE tissues by LCM. We observed that immunostaining should be performed at 4°C to preserve the mRNA from the cells. The expression of Bcl-2 in the endothelial cells was evaluated by RT-PCR and by real time PCR. GAPDH and 18S were used as house keeping genes for RT-PCR and real time PCR, respectively. This report unveils a method for quantitative gene expression analysis in cells that were identified by immunostaining and retrieved by LCM from FFPE tissues. This method is ideally suited for the analysis of relatively rare cell types within a tissue, and should improve on our ability to perform differential diagnosis of pathologies as compared to conventional LCM.
A method based on laser capture microdissection (LCM) that would allow for quantitative gene expression analyses of specific target cells in tissue sections could be very useful to several experimental and clinical fields. To date, formaldehyde as a 10% neutral buffered formalin is the most widely used as a fixative for various human tissues. As with DNA, formaldehyde reacts with RNA forming an N-methylol (N-CH2OH) followed by an electrophilic attack to form a methylene bridge between amino groups. Adenine is the most susceptible nucleotide to electrophilic attack and it is likely that the adenines within the mRNA sequence and the poly(A) tail of mRNA will be modified in the formaldehyde-fixated paraffin embedded FFPE sections to varying degrees. Thus, it is normally considered that RNA isolated from FFPE sections are less suitable for reverse transcription (cDNA synthesis), as compared to RNA isolated frozen tissue sections (Srinivasan et al., 2002). However a recent report has suggested that it is possible to perform gene expression analysis from FFPE tissues (Pagedar et al., 2006).
The tumor microenvironment is composed of several cell types, of which the endothelial cells constitute a small fraction of the overall cell number. Therefore, if one needs to quantitatively analyze gene expression specifically in endothelial cells, they have to be somewhat selectively removed from the tissue, without contamination of other cell types. To overcome this technical problem, we developed a novel strategy of immuno-LCM in FFPE tissues, which may greatly facilitate the identification of the target cells (i.e. tumor-associated endothelial cells). Using such strategy, we have recently reported, using a severe combined immunodeficient (SCID) mouse model of human tumor angiogenesis, that Bcl-2 orchestrates a tunor-endothelial cell crosstalk that promotes tumor growth (Kaneko et al., 2007). In the analysis, we used the SCID mouse model of human tumor angiogenesis. It allows for the study of human tumors in murine hosts by implanting defined endothelial and tumor cell populations in biodegradable scaffolds (Nör et al., 1999, 2001a). We have demonstrated that this method allows for the development of functional human blood vessels that anastomize with the mouse vasculature (Nör et al., 2001a). Notably, it is suitable for studying both physiologic (when only human endothelial cells are implanted), or pathologic angiogenic processes (when both tumor cells and endothelial cells are co-implanted). Here, we used the SCID mouse model of human tumor angiogenesis as a platform to characterize in detail the method developed for quantitative gene expression analysis in LCM-retrieved endothelial cells after immunostaining from FFPE engineered tumor tissues (Kaneko et al., 2007).
Human dermal microvascular endothelial cells (HDMEC; Cambrex, Walkersville, MD, USA) stably transduced with Bcl-2 (HDMEC-Bcl-2) and empty vector controls HDMEC-LXSN were cultured in EGM2-MV (Cambrex), as described previously (Nör et al. 1999, 2001a,b; Karl et al. 2005; Kaneko et al. 2007). Oral squamous cell carcinoma cells (OSCC3; gift of M. Lingen, University of Chicago) were cultured in DMEM supplemented with 10% fetal bovine serum (FBS).
The care and treatment of experimental animals were in accordance with University of Michigan institutional guidelines. Porous poly-L-lactic acid (PLLA; Boehringer Ingelheim, Germany) scaffolds (6 mm × 6 mm × 1 mm biodegradable) with an average pore diameter of 180 µm were fabricated, as previously described (Nör, J.E., 2001). Just before implantation, scaffolds were seeded with 0.9 × 106 HDMEC and 0.1 × 106 tumour cells (OSCC3) in a 1:1 Matrigel/EGM2-MV mix. Five to nine week-old male SCID mice (n=5) (CB-17, Chales River Laboratories, Wilmington, MA, USA) were anesthetized with ketamine and xylazine, and two scaffolds were implanted subcutaneously in the dorsal region of each mouse. Twenty-eight days after transplantation, mice were euthanized, and the scaffolds were retrieved, fixed for 10 hours in 10% buffered formaldehyde at 4°C. Then, the samples were cut in two pieces, one for paraffin blocks, and the other for frozen blocks. Eight-µm serial sections from each piece were cut, and mounted on PEN foil slides (Leica Microsystems, Bannock Burn, IL) for LCM.
To stain the microvascular networks formed by the implanted human endothelial cells, immunoperoxidase staining was conducted using a mouse anti-human Factor VIII monoclonal antibody (1:100 dilution; Lab Vision, Fremont, CA, USA). Slide-mounted tissue sections were deparaffinized twice with 3 min xylene washes at room temperature, and rehydrated through a series of graded ethanol at 4°C. A working dilution of 0.125% trypsin was used to pre-treat these sections for 30 min at room temperature. Endogenous peroxidase activity was blocked by incubation with 3% hydrogen peroxide with methanol for 3 min at 4°C. Immunostaining was performed as described earlier (Nör et al, 2001a.b., Kaneko et al., 2007). A first incubation was made with mouse anti-Factor VIII (diluted 1:100) for 16 h at 4°C. Sections were then incubated with biotinylated anti-mouse IgG (Vector Laboratories, Burlingame, CA; diluted 1:500) for 1 h at 4°C and incubated with avidin-biotin-peroxidase complex (Elite ABC kit; Vector) at 4°C. Peroxidase activity was developed in diaminobenzidine-H2O2 solution (DAB substrate kit; Vector) for 3 min at room temperature. Sections were then air-dried at 4°C. Negative control staining was always conducted in parallel by the omission of the primary antibody, anti-mouse IgG, or avidin-biotin-peroxidase complex.
Frozen sections were cut in a cryostat (Leica, Germany), and were incubated with anti-Factor VIII antibody (1:500 dilution). This was followed by the same steps of immunostaining for paraffin sections.
After immunostaining, a three-step dissection strategy was performed using a laser capture microdissection microscope (Leica AS LMD; Leica Microsystems) with a pulsed 337 nm UV laser (Leica Microsystems). Blood cells in Factor VIII-positive capillaries were excluded first. Then Factor VIII-positive endothelial cells (approximately 1,500) were dissected and collected into individual tubes filled with RNAlater® (Ambion, Austin, TX, USA) and immediately placed on ice. Cells without a nucleus were excluded from the cell count. Then, tumor cells surrounding the blood vessels were collected similarly. Experiments were always carried out in duplicate. The FFPE blocks used in this study were approximately 1 year-old.
Total RNA from paraffin embedded tissue sections was extracted using TRIzol reagent (Invitrogen), and purified with RNeasy Mini kits (Qiagen), as we described (Kaneko et al, 2007, 2008). Its purity (Mo, SGman: OD260/280 = 1.9–2.1) was determined photometrically (Ultrospec 2000, Pharmacia Biotech, Germany). cDNA synthesis and PCR amplification were done in single tubes with SuperScript one-step reverse transcription-PCR (RT-PCR) and Platinum Taq kit (Invitrogen) using simultaneously a human Bcl-2 primer set and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) primer set. The sequence of the primers used here were as follows: Bcl-2, sense, CTGCGAAGAACCTTGTGTGA and anti-sense TGTCCCTACCAACCAGAAGG; and GAPDH, sense, CATGGCCTCCAAGGAGTAAG and anti-sense, AGGGGTCTACAGGCAACTG. Negative control reaction was always conducted in parallel by using water instead of RNA. The PCR products were detected by standard agarose gel with 1.5% ethidium bromide.
Total RNA was extracted and purified from LCM-retrieved samples, as described above. First-strand complementary DNA (cDNA) synthesis was performed with TaqMan® reverse transcription reagents (Applied Biosystems). Probe and primer sets of TaqMan® Gene Expression Assays (Hs00608023_m1; Bcl-2, Hs00174103_m1; and Hs99999901_s1; 18S) were obtained from Applied Biosystems. Total RNA at 0.02 µg per 30 µl of reaction mixture was prepared by TaqMan® Universal PCR Master Mix (Applied Biosystems). For standard RNA, we used TaqMan® One-step RT-PCR Master Mix Reagents (Applied Biosystems). The reactions were performed in 96-well clear optical reaction plate using ABI7700 Sequence Detection System (Applied Biosystems), and the data were normalized by the data of 18S controls. All reactions were performed in triplicate, each run contained at least one negative and one positive controls, and three independent experiments were performed to confirm reproducibility of results.
The overall strategy used here for quantitative gene expression analysis of immunostained FFPE tissue sections is summerized in Figure 1. After tumor retrieval, samples were processed for FFPE, sectioned, mounted on LCM slides, and immunostained using the avidin-biotin-peroxidase method. Following the LCM, RNA extraction and PCR analysis were performed.
For statistical comparison of two experimental groups, we used Student's t tests using Prism 5.0 (Graphpad, San Diego, California).
The step-by-step LCM procedure is shown in Fig. 2. We first retrieved blood cells, and then Factor VIII-immunostained endothelial cells, and lastly the surrounding tumor cells. Notably, Factor VIII immunostaining allowed for clear identification of the endothelial cells.
To confirm if the tissues dissected by LCM were without contamination, we performed RT-PCR analysis for vascular endothelial cell growth factor receptor 2 (VEGFR2: a specific marker for endothelial cells) and E-cadherin (a specific marker for cells of ectodermal origin). RT-PCR analysis showed that mRNA from dissected tissues was adequate for PCR analysis for target genes. Furthermore, dissected endothelial cells did not express E-cadherin, and tumor cells did not express VEGFR2, confirming the quality and origin of the RNA (Fig. 3). As shown in Fig. 4a, Bcl-2 and GAPDH gene expression in both FFPE and frozen tissues was detected by RT-PCR. Notably, we were able to compare the expression levels between tumor cells and endothelial cells in paraffin samples by real time PCR (Fig. 4b).
LCM is normally used to demonstrate the presence of specific gene transcripts in cryosectioned tissues (Cristobal et al., 2004, 2005). LCM of immunostained frozen tissue sections has been used successfully for gene expression analysis (Fend et al., 1999, Stoebner et al., 2008). We have also reported that gene expression of some antigen presenting cell-related molecules, such as HLA-DR alpha-chain, CD83, and CD86, can be analyzed in LCM-retrieved HLA-DR-expressing cells from immunostained frozen tissue sections (Kaneko et al., 2008). RNA recovery from several kinds of FFPE tissues is now possible (Su et al., 2004), although frozen tissue sections may allow us to retrieve better quality and more amount of mRNA, as compared to mRNA from the FFPE tissues. However, no technical reports have described quantitative gene expression analysis after LCM of immunostained FFPE tissues. In this report, we describe a method for gene expression analysis of factor VIII-expressing endothelial cells microdissected from FFPE tissues. This method allowed us to retrieve mRNA from a pre-determined cell type in histologically defined areas of routinely prepared FFPE tissues. We have also reported that this method can be used in FFPE human surgical materials such as those from head and neck carcinomas (Kaneko et al., 2007).
We here demonstrated that formaldehyde, a crosslinking fixative, do not prevent us from extracting enough RNA for RT-PCR or real time PCR, at least for Bcl-2 and hose keeping gene amplification. These results are in agreement with previous reports assessing PCR amplication of RNA in tissues recovered from paraffin blocks by other methods (Ben-Ezra et al., 1991; Koopmans et al., 1993).
The present method of selective LCM-retrieval may be useful to compare gene expression patterns in tumor cells versus endothelial cells during tumor progression or tumor angiogenesis. This method could also be useful for quantitative gene expression analysis in blood vessels from tumors of patients that have been treated with antiangiogenic drugs, allowing for validation of the effect of drug on the expected targets. Such capability might be exceedingly useful for the evaluation of the bioactivity of new drugs.
In conclusion, we demonstrated that immunostaining-based LCM method presented here allowed us to quantify Bcl-2 and housekeeping gene mRNAs from endothelial cells in FFPE tissue sections. This method may be suited for the analysis of relatively rare cell types within a tissue, and should improve on our ability to perform differential diagnosis of pathologies as compared to conventional LCM.
The authors wish to thank Chris Strayhorn for his assistance with the preparation of tissues for histology. This work was supported by grant P50-CA97248 (University of Michigan Head and Neck SPORE) from the NIH/NCI (JEN); grants R01-DE14601, R01-DE15948, R01-DE16586, R21-DE19279 from the NIH/NIDCR (JEN); and Grants-in-Aid for Scientific Research (Nos. 15791091 and 18791393) from the Japan Society for the Promotion of Sciences (TK).