| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From the Department of Pathology
*
and the Division of Gastroenterology and Hepatology,
Department of Medicine, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
With the exception of a small number of genes, DNA mutational analysis cannot be carried out by means of a histological stain. Protein expression analysis using well established procedures for immunohistochemical staining is very often constrained by the lack of availability of specific and sensitive antibodies for the expanding repertoire of human cancer-related gene products. In addition, many available antibodies prove to be ineffective on formalin-fixed, paraffin-embedded tissue sections due to irreversible alterations to native proteins resulting from standard practices of fixation and tissue processing. The molecular analysis of RNA, in particular, is severely hampered as a result of RNA loss secondary to the combined actions of tissue RNases and high temperatures commonly used in specimen processing.
An alternative approach, permitting integration of cellular characteristics with underlying molecular alterations, is based on the use of tissue microdissection.1, 2 Cellular areas in histological sections are first delineated under light microscopic and stereomicroscopic observation. Then, selected sites, chosen for their representativeness of biological processes operative in a tissue specimen, are removed with precision so as to minimize the inclusion of unwanted cellular elements not intended for genotyping. This can be accomplished simply and rapidly by manual methods3 or through the use of complex equipment specially designed for this purpose.4 Two commercially available units have recently been described.These are the PixCell II laser capture microdissection apparatus (Arcturus, Mountain View, CA)5 and the Robot-Micro-Beam (PALM) laser microdissector (Wolfratshausen, Germany).6 Both use focused laser energy to effect discrete removal of cellular material from microscopic slides, giving the operator the freedom to select topographic tissue targets. Using nucleic acid amplification, samples as small as a single cell can be analyzed for a broad array of gene alterations.7
The full potential of microdissection-based, polymerase chain reaction (PCR) format genotyping has not yet been realized due to inherent limitations related to tissue fixation and processing.8 Prolonged exposure of tissues to the effects of formaldehyde results in irreversible cross-linking of protein and nucleic acid, usually limiting maximum PCR amplicon size to 200 to 300 bp.8 In addition, tissues excessively fixed in formaldehyde suffer nucleic acid strand scission, diminishing the efficiency of nucleic acid amplification and further reducing maximum amplicon size.8 Although DNA survives fixation and embedding reasonably well, RNA content is seriously decreased due to a combination of RNase activity present in virtually all tissue specimens and the use of excessive heating during tissue processing. The latter occurs at many stages of tissue handling, eg, during melting of paraffin at the embedding step and during oven baking of tissue sections to maximize adherence to glass slides.
Standard practices of tissue fixation and embedding, therefore, impose a significant limitation on the application of molecular methods in everyday clinical practice. Alternative techniques to preserve nucleic acids, such as immediate freezing and cryostat sectioning, are not generally desirable in a clinical context due to the frozen section artifact impairment of cellular morphology that accompanies such an approach. Given the empirical foundation of microscopic analysis for current methods of tissue diagnosis and prognosis, the transfer of molecular methods into pathology practice has been greatly hindered.
The logical solution to this problem is a method for tissue fixation and processing that maintains cellular integrity while quantitatively and qualitatively preserving nucleic acids. Protocols designed to protect nucleic acids are also likely to influence favorably the availability of other cellular constituents including proteins, carbohydrates, and lipids. In this report, we present such a system, which is both easy and cost-effective to apply. The system is based on an established process of cold-temperature embedding of tissue specimens in an inert plastic resin with water-miscible characteristics (Immunobed Kit, Polysciences, Warminster, PA). The technique, termed the Immunobed method, has been used successfully for ultrastructural immunocytochemistry. We report here that the system, which allows tissue sections to be cut as thin as 0.5 µm, confers excellent retention of long DNA and RNA lengths suitable for detailed genetic analysis. Through the use of microdissection techniques, this system permits the positive attributes of both microscopic and molecular analysis to be realized for molecular pathology research purposes as well as for clinical diagnosis and treatment.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Cold-Temperature Embedding in Plastic Resin
Tissues were processed for plastic resin embedding according to
the Immunobed protocol provided by manufacturer. Tissues underwent
three 1-hour incubations in increasing grades of ethanol to 100%
ethanol at 4°C. This was followed by immersion in 50%/50%
ethanol/Immunobed solution for 2 hours at 4°C. Catalyst was added to
the plastic resin and the tissue embedded overnight in catalyzed resin
at 4°C using standard beem capsules as tissue supports. Blocks
were stored at 4°C until the time of microtomy.
Preparation of Tissue Sections
Immunobed blocks were cut on a Reichert ultracut S microtome at
thicknesses from 0.5 to 4 µm. Tissue sections were floated in a
DEPC-treated water bath. Sections were picked up on clean glass slides
and allowed to air dry at 4°C, after which they were stored at
-20°C until ready for staining and microdissection. Immediately
before microdissection, tissue sections were stained with
hematoxylin-eosin and examined without coverslipping. To confirm
accuracy of tissue microdissection, postmicrodissected stained
histological sections were dehydrated through graded alcohols and
xylene. Routinely processed tissue sections received a nonaqueous mount
and plastic-embedded sections were covered with Crystalmount (Biomeda
Corp., Foster City, CA).
Tissue Microdissection
Paraffin-embedded tissue sections were deparaffinized in xylene,
rehydrated by passage through graded alcohols, and air-dried.
Plastic resin-embedded tissue sections were simply air-dried in
preparation for microdissection. Microdissection was carried out
under stereomicroscopic observation using a Zeiss stereomicroscope.
Selected areas of liver tissue were removed using a no. 11 scalpel and
placed directly into 500-ml tubes for molecular analysis.
Detection and Genotyping of Hepatitis C Virus (HCV) and Internal
Nucleic Acid Controls
Pairs of oligonucleotide primers were used to amplify by PCR and
to DNA sequence two separate regions of HCV.9, 10
A short
(153 bp) amplicon was generated from the 5' nontranslated
region9
and a longer (364 bp) PCR product was created from
the NS5 region.10
In both instances, the upstream primers
were used to sequence the amplified material according to methods
described below. Positive HCV controls consisted of fresh-frozen blood
and/or tissue known to contain the virus. Amplification and sequencing
of housekeeping genes in each tissue specimen were used as internal
positive controls for DNA and RNA content. DNA amplification reactions
were targeted to a 335-bp fragment derived from the 3' nontranslated
region of the epidermal growth factor receptor gene.11
In
a similar fashion, a 264-bp amplicon was generated from mRNA of the
ß-actin gene.12
Microdissected tissue for RNA genotyping (HCV and ß-actin) was treated with guanidinium isothiocyanate for isolation of RNA.13 The extracted RNA was reverse transcribed with avian myeloleukemia virus, and the resulting cDNA was split into two parts; one portion was used for HCV genotyping and the remaining portion was destined for housekeeping gene amplification. Microdissected tissue for DNA genotyping was treated with Proteinase K (10 mg/ml) for 1 hour. A one-fourth portion of the postdigested material was place directly in a microcentrifuge tube for PCR amplification using the GeneAmp kit (Perkin-Elmer, Norwalk, CT).
The products of PCR amplification were electrophoresed in 2% agarose and amplicon bands, visualized with ethidium bromide, and excised on a transilluminator. PCR-generated DNA was isolated with glassmilk (GeneClean, Bio 101, La Jolla, CA) and manually sequenced according to manufacturer’s instructions (Sequenase, United States Biochemical Corp, Cleveland, OH). The DNA sequence was read from autoradiograms of 6% polyacrylamide gels exposed overnight.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. The quality of cellular morphology in 4-µm-thick sections was
judged to be equivalent between the two methods of tissue handling.
|
PCR amplification of DNA gene targets <250 bp in length occurred in a
reproducible fashion in both routine and molecular processed tissue
(Figures 2and 3). Longer DNA amplicons could not be achieved with routinely fixed
tissue, in contrast to plastic resin-embedded specimens, which gave a
strong PCR product band (Figure 3)
. Only short RNA reverse
transcription PCR (RT-PCR) amplification products were generated from
routinely processed tissue; however, this required nested PCR, which
was more tedious to perform and which carried a significant risk of
cross-contamination. Furthermore, the product bands in these cases
tended to be inconsistent and relatively weak in intensity. In
contrast, RNA extracted from microdissected plastic resin-embedded
tissue sections generated in a robust manner both short and long RNA
amplification products (Figure 3)
. Even microdissection targets as
small as 20 cells from plastic resin sections yielded discernable
ethidium bromide-stained bands on agarose gel electrophoresis (Figures 2
and 3)
. Both short and long RNA products could be detected and
consistently genotyped from plastic resin tissue sections (Figure 4)
.
|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Appreciation of cellular heterogeneity is fundamental to any effective genotyping strategy for solid tissue specimens. Failure to take this essential factor into account invariably leads to averaging of genetic information between divergent cell types. For example, the presence of a highly proliferative cell type, such as reactive fibroblasts, in a tissue sample may contribute to a potentially false impression that a rapidly growing tumor cell population exists in the sample under study. In this way, a benign neoplastic process that evokes an exuberant reactive cellular response may be mistaken for malignancy. Molecular analysis must be correlated on a cell-by-cell basis with the cell population of the sample being studied. Until this basic fact is completely understood, advanced techniques such as DNA sequencing and RNA expression profiling will invariably fail to have the expected impact in the practice of clinical medicine.
The best approach might take the form of a microscope slide staining procedure capable of revealing molecular aberrations in neoplastic and supporting cells. Unfortunately this is not feasible, because detailed molecular analysis of multiple genes cannot be achieved by the application of any one staining technique. Rather, multigene analysis has become a reality through the use of non-slide-based high-throughput technologies that are effective with samples as small as a single cell.7 In response, tissue microdissection approaches have gained considerable interest as the best way to bridge the gap between cellular morphology and molecular diagnostics.1, 2, 3, 4, 5, 6
Tissue microdissection logically places genetic analysis after histopathological evaluation. This ensures that the full measure of information derived from microscopic evaluation encompassing concepts related to cellular pathogenesis can be channeled into obtaining the very best sample for molecular analysis. This process of sample selection is far less subjective than microscopic diagnosis, which often is imprecise in establishing a tissue diagnosis and even more inexact in predicting tumor behavior and treatment responsiveness. In an era of staggering achievements in our understanding of the molecular basis of medicine, the role of the pathologist will inevitably grow to include the careful selection of representative tissue samples for genetic analysis. This should lead to a more objective means of detecting and classifying cancer, better prognostication of tumor behavior, and better rationalization of treatment.
At present, the major impediment to achieving this goal is the well recognized detrimental effect of routine tissue fixation and processing on nucleic acid and protein content and structure. Ten-percent neutral buffered formalin fixation followed by paraffin wax embedding produces histological sections of excellent quality, albeit at the expense of virtually total RNA loss and serious cross-linking and strand scission of DNA.8 The latter can be attributed to prolonged formaldehyde fixation beyond the time sufficient to achieve optimal cellular morphology. Although small loss of RNA may be unavoidable given the ubiquitous presence of RNases in tissues and environment, major degradation of RNA is preventable by simple measures such as the use of RNase-free water and cold temperature for tissue processing and embedding. This rationale served as the basis for a system of molecular processing designed to maintain cellular integrity while preserving tissue nucleic acid and proteins.
Cold-temperature plastic resin embedding was chosen as it already has proven to be an effective technique for ultrastructural analysis.15, 16, 17 Two properties of Immunobed made it especially attractive. First, Immunobed can be polymerized in the cold without the need for ultraviolet radiation, which may exert damaging effects on nucleic acids. Second, Immunobed is a water-miscible plastic resin that allows full access by water-soluble molecules to tissue constituents. Its most common use currently is in the context of ultrastructural immunocytochemistry. In addition to these properties, Immunobed can be thin-sectioned for both light and electron microscopic analysis. We have shown in this report that the microscopic appearance of standard 4-µm-thick histology sections is equivalent to that seen with paraffin embedding. In fact, cellular morphology can be improved even further by virtue of Immunobed’s thin sectioning properties.
Cold-temperature embedding as described in this study was primarily
responsible for preserving long DNA and RNA lengths for PCR format
genotyping. Exogenous (HCV) and endogenous (ß-actin) RNA was easily
amplified in a highly sensitive fashion (Figures 2
and 3)
. Samples as
small as 20 cells microdissected by manual methods were sufficient to
detect and sequence HCV RNA from liver biopsies (Figure 4)
. A similar
beneficial effect of Immunobed can be expected on tissue histochemical
and immunohistochemical staining due to the improved preservation of
cellular constituents and antigen reactivity already demonstrated for
this embedding medium.15, 16, 17
Thus, the methods used here
to integrate genomic and RNA expression analysis with histopathology
may also be expected to facilitate proteome analysis into
standard tissue pathology.
Tissue fixation was modified by substituting 2% parafomaldehyde for 10% neutral buffered formalin. The water used to constitute the paraformaldehyde fixative was rendered free of RNases through the action of 0.1% diethyl pyrocarbonate. Fixation was limited to 2 hours to avoid excessive cross-linking of cellular constitutents. These simple measures, together with cold-temperature plastic resin embedding, can easily be adapted to anatomical pathology practice without significant increase in cost or turnaround time. The effort involved in its setup is more than justified by the benefits resulting from the potential for full application of molecular analysis.
Tissue microdissection was performed manually on plastic resin-embedded sections. In contrast to deparaffinized sections, plastic-embedded sections did not require exact wetting conditions to effect tissue removal. The action of sample removal was found to be greatly facilitated due to the support properties of the plastic resin. The entire process was found to be much easier and more quickly carried out than mechanical microdissection systems. Continued experience with this plastic resin embedding, including modifications for its introduction and possible removal before microdissection, may be expected to facilitate the further integration of histopathology and molecular biology. Thus, we recommend cold-temperature plastic resin embedding as a molecular tissue processing approach well suited to molecular analysis in anatomical pathology.
|
Footnotes |
|---|
Accepted for publication August 26, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
