Extraction and Analysis of Diagnostically Useful Proteins from Formalin-fixed,Paraffin-embedded Tissue Sections

Abstract

We describe and discuss a method of protein extraction for Western blot analysis from formalin-fixed, paraffin-embedded tissue sections. From 5-mm² 50-μm-thick tissue sections, an abundance of proteins could be extracted by incubating the sections in lysis buffer containing 2% sodium dodecyl sulfate (SDS) at 100C for 20 min followed by incubation at 60C for 2 hr. Extracts yielded discernible protein bands ranging from 10 kD to 120 kD as identified by SDS-polyacrylamide gel electrophoresis (PAGE). Western blot analysis successfully detected membrane-bound protein such as E-cadherin, cytosolic protein such as β-catenin, and nuclear proteins including proliferating cell nuclear antigen (PCNA), mutant-type p53, cyclin D1, cyclin E, and cyclin-dependent kinases (CDKs). With this technique, we could examine cyclin D1 and CDK2 expression in small adenomas compared with cancer tissues and normal mucosa. The simple method of protein extraction described here should make it possible to use large-scale archives of formalin-fixed, paraffin-embedded samples for Western blot analysis, and its application could lead to detailed analysis of protein expression. This new technique should yield valuable information for molecular biology.

Keywords

paraffin-embedded tissue protein extraction Western blot analysis

Recent advances in molecular technology have identified many genes responsible for or associated with malignant transformation of the cell (Vogelstein et al. 1988; Marx 1994). Proteins encoded by tumor-associate genes are suspected to play crucial roles in the formation and progression of human malignancies, so that the need to investigate protein expression is growing. Immunohistochemical staining and Western blot analysis using specific antibodies are the most frequently adopted techniques for identification of these proteins. The former provides valuable information on localization and distribution of the proteins in cells and tissues. However, the evaluation of staining results is not always easy because of the presence of potentially crossreacting proteins, and the results are equally hard to quantify. Western blotting, on the other hand, enables us to assess protein expression by referring to its molecular weight, and makes a quantitative interpretation of the expressed protein possible by means of densitometry (Dell'Orco et al. 1997) although no information is available on the localization and distribution of cells expressing the protein.

Immunohistochemistry and Western blot analysis are therefore complementary techniques for analysis of protein expression. These two methods, however, can not be applied to the same materials because immunohistochemistry requires processing of the tissues by fixatives such as formalin, which is believed to destroy the proteins and to make it impossible to analyze the same tissues by Western blotting (Conti et al. 1988). Recently, however, many authors have reported that DNA and RNA remained well preserved in routinely processed paraffin-embedded tissues and could be extracted for analysis by use of the polymerase chain reaction (von Weizsäcker et al. 1991; Neubauer et al. 1992; Ohue et al. 1994). These findings imply the possibility of protein extraction from formalin-fixed, paraffin-embedded tissue sections.

In this article we describe a method of protein extraction for Western blot analysis from formalin-fixed, paraffin-embedded tissue sections. Extraction efficiency is comparable to that from frozen tissue samples, and membrane-bound, cytosolic, and nuclear proteins were clearly identified by Western blotting. The method described here should make it possible to use large-scale archives of pathological specimens processed through routine fixation not only for immunohistochemistry but also for Western blot analysis, and can be expected to yield valuable information for molecular biology.

Materials and Methods

Tissue Samples

Fresh specimens of colorectal cancer were surgically obtained at the Second Department of Surgery of Osaka University Hospital. Cancer tissues were cut into 5-mm-thick slices and immediately fixed in 10% buffered formalin at 4C for 24–48 hr. Four colorectal adenomas were also obtained at surgery together with the cancer tissues. The size of the adenomas ranged from 5 mm to 14 mm maximal diameter (mean 8.8 mm), and all adenomas exhibited moderate dysplasia. These adenomas were fixed in the same manner as used for cancer tissues. The cancer tissues and adenomas were then washed in water for 1 hr and dehydrated in graded ethanol (60%, 80%, 90%, 95%, and 100%) at 4C. Finally, after permeation in xylene, they were embedded in paraffin. The storage period for paraffin blocks used in this study was from 2 months to several years. The full depth of the colorectal wall containing normal mucosa remote from cancer tissues was also cut into 5-mm-thick and 5-cm-long slices and was fixed in buffered formalin at 4C for 24–48 h, then embedded in paraffin with the same procedure as used for cancer tissues. Fresh cancer tissues were also collected and stored at -80C until use for protein assay and Western blotting.

Antibodies

Mouse monoclonal antibodies (MAbs) against PCNA (PC-10, IgG_2a) and p53 (1801, IgG₁) were purchased from Novocastra Laboratories (Newcastle, UK). A mouse MAb against cyclin E (HE-12, IgG₁) was acquired from Oncogene Science (Cambridge, MA). The mouse MAb against β-catenin (IgG₁), which was raised against a 23-kD protein fragment corresponding to amino acid residues 571–781 of mouse β-catenin (Ozawa et al. 1989), was purchased from Transduction Laboratories (Lexington, KY). The mouse MAb (HECD-1, IgG₁) against human E-cadherin, which was obtained from Takara Shuzo (Kyoto, Japan), recognizes an extracellular epitope of human E-cadherin (Nagafuchi and Takeichi 1988). Rabbit polyclonal antibodies against cyclin D1 (M-20, IgG) and CDK2 (M-2, IgG) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The antibody against cyclin D1 recognizes an epitope within the C-terminal domain (amino acid residues 276–295) of human cyclin D1 (Xiong et al. 1991) and the one against CDK2 recognizes the C-terminal domain (amino acid residues 283–298) of human CDK2 (Meyerson et al. 1992). A rabbit polyclonal antibody against CDK4 (IgG), which recognizes the C-terminal half of human CDK4 (amino acid residues 140–303) (Matsushime et al. 1992), was purchased from Pharmingen (San Diego, CA).

Figure 1

H–E staining of a 4-μm-thick paraffin section (A) and a 50-μm-thick serial section after dissection of cancer tissue (B). Bars = 5 mm.

Protein Extraction

Paraffin sections 4 μm thick were cut for hematoxylin–eosin staining and 50-μm-thick serial sections were cut for protein extraction and mounted on plain glass slides. Three 50-μm-thick sections for protein extraction were deparaffinized in xylene, rehydrated in graded ethanol, immersed in distilled water, and air-dried. To exclusively collect 5-mm² cancer tissues, the targeted areas were cut macroscopically with a fine needle referring to the microscopic observation of the morphology of serial H–E sections. After the tissue sections on the glass slide were immersed in distilled water, only the targeted areas of cancer tissue were separated from the glass slide and recovered (Figure 1). Adenomas were also cut and collected in the same manner. Normal mucosa was recovered from 5-cm-long sections of full-depth colorectal wall with a fine needle. The dissected tissues were further cut into small pieces and then placed in Eppendorf tubes. Two hundred μl of RIPA buffer, pH 7.6 [1 M sodium dihydrogen phosphate, 10 mM disodium hydrogen phosphate, 154 mM sodium chloride, 1% Triton X-100, 12 mM sodium deoxycholate, 0.2% sodium azide, 0.95 mM fluoride, 2 mM phenylmethylsulfonyl fluoride, 50 mg/ml aprotinin (Sigma Chemical; St Louis, MO), 50 mM leupeptin (Bayer; Leverkusen, Germany)], containing 0.1% or 2% SDS, was added to each tube, and the contents were incubated under different conditions as follows: at 0C for 2 hr; at 37C for 2 hr; at 60C for 2 hr; and at 100C for 20 min, followed by incubation at 60C for 2 hr. After incubation, the tissue lysates were centrifuged at 15,000 X g for 20 min at 4C. The supernatants were collected and stored at -80C until use for protein assay and Western blot analysis.

From fresh cancer tissues, protein was extracted by the conventional method, in which 100 mg of the fresh frozen tissues was homogenized with a syringe-type, hand-held homogenizer and incubated in 1 ml of RIPA buffer containing 0.1% SDS at 0C. The tissue lysates from fresh frozen tissues were also centrifuged at 15,000 X g for 20 min at 4C and the supernatants were collected and stored at -80C.

Protein Assay

The protein concentration of the lysates was determined with the Bio-Rad DC Protein Assay (Bio-Rad Laboratories: Hercules, CA). The reaction is similar to the well-documented Lowry assay (Lowry et al. 1951). Reagent was added to the lysates and the absorbance was read at 750 nm. The protein concentration was determined based on the standard curve using bovine serum albumin.

SDS-PAGE and Western Blot Analysis

The lysates were treated with SDS-PAGE loading buffer (at a final concentration of 65 mM Tris, 5% 2-mercaptoethanol, 3% SDS, and 10% glycerol) at 100C for 5 min. The total lysates extracted from a 5-mm² cancer tissue section or a 5-cm-long normal mucosal section of 50 μm thickness were applied to each lane and electrophoresed in 12% or 7.5% polyacrylamide gel (13 X 13 cm, 1 mm thick). Fifty μg of total cellular protein of extracts from fresh frozen tissue was also applied and electrophoresed. To examine the range of molecular weight of the extracted proteins, we investigated the SDS-PAGE pattern by Coomassie blue staining of the electrophoresed gels before transfer. For Western blot analysis, the electrophoresed proteins were transferred onto an Immobilon polyvinylidene difluoride (PVDF) membrane (Millipore; Bedford, MA) in transfer buffer containing 192 mM glycine, 25 mM Tris-HCl, pH 8.3, 20% v/v methanol, and 0.02% SDS. After overnight treatment with blocking buffer (5% nonfat dry milk, 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% Tween-20) at 4C, the membranes were incubated with the primary antibody at room temperature (RT) for 2 hr at the following dilutions or final concentrations: anti-PCNA MAb 1:400; anti-p53 MAb 1:50; anti-cyclin D1 polyclonal antibody 0.1 μg/ml; anti-cyclin E MAb 2 μg/ ml; anti-CDK2 polyclonal antibody 0.2 μg/ml; anti-CDK4 polyclonal antibody 1:500; anti-β-catenin MAb 1:500; and anti-E-cadherin MAb 2 μg/ml. For negative control, normal mouse IgG (2 μg/ml) or normal rabbit IgG (5 μg/ml) was added instead of the primary antibodies.

After being washed with TBST (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% Tween-20), the filters were incubated for 1 h with alkaline phosphatase-conjugated rabbit anti-mouse IgG or goat anti-rabbit IgG at a dilution of 1:3000 and were developed with the ProtoBlot NBT and BICP Color Development System (Promega; Madison, WI).

To test the sensitivity, sequentially diluted lysates were analyzed by Western blotting for PCNA and the minimal size of dissected tissue required was determined.

To examine the expression of cyclin D1 and CDK2 in normal mucosa, adenomas, and cancer tissues, 50 μg of total cell protein extracted from these tissue sections was applied for Western blotting of cyclin D1 and CDK2. Densitometric analysis of the Western blotting was performed with an Image Scanning device (Molecular Dynamics; Sunnyvale, CA). After the protein bands on the membrane were scanned, the density was measured. The quantity of the protein was calculated integrally. For four adenomas and cancer tissues, we compared the expression levels of cyclin D1 and CDK2 with matched normal mucosa.

Results

Protein Extraction

The total extracted protein from 5-mm² cancer tissue that was dissected from the same area of the serial sections was measured in six cases. This procedure was conducted under various conditions of incubation in RIPA buffer containing 0.1 or 2% SDS. Protein could not be extracted with incubation in RIPA buffer containing 0.1% SDS at 0C, 37C, or 60C for 2 hr. Preincubation at 100C for 20 min before treatment at 60C for 2 hr did not increase the efficiency of protein extraction. The next step was to extract protein from the tissue sections with incubation in RIPA buffer containing 2% SDS. For six samples, an average of only 13.6 μg of protein could be extracted at 0C, 14.3 μg at 37C, and 15.2 μg at 60C for 2-hr incubation (Table 1). On the other hand, when the sections were incubated at 100C for 20 min before incubation at 60C for 2 hr, 121.5 μg of protein was extracted from 5-mm² 50-μm-thick cancer tissue (Table 1). Because the average dry weight of 5-mm² 50-μm-thick cancer tissue was 0.74 mg for 54 samples (range 0.56–1.11 mg, median 0.67 mg), the efficiency of protein extraction was calculated as 164.2 μg/mg of dry cancer tissue.

Table 1.

Total amount of protein extracted from 5-mm² 50-μm-thick cancer tissuea

	Incubating conditions
Sample	0C X 2 hr	37C X 2 hr	60C X 2 hr	100C X 20 min 60C X 2 hr
1	18.1	14.5	23.6	171.1
2	3.4	7.7	20.8	151.3
3	35.5	7.3	16.4	151.7
4	13.8	38.0	16.0	102.2
5	4.5	8.4	10.1	70.3
6	6.4	9.6	4.2	82.3
Mean (μg)	13.6	14.3	15.2	121.5

RIPA buffer containing 2% SDS was used for all incubating conditions.

We examined the molecular weight of extracted protein from paraffin sections compared with fresh frozen tissues. Although the SDS-PAGE pattern of the extracted protein identified with Coomassie blue staining was not identical to that obtained from fresh frozen tissue, it indicated that the protein molecules ranging from 10 kD to 120 kD remained well preserved (Figure 2A). The storage period of paraffin blocks did not affect the protein extraction or the electrophoretic pattern.

Western Blot Analysis

First, PCNA expression was analyzed by Western blotting of extracted proteins. As shown in Figure 2B, extracts from both normal mucosa and cancer tissue produced a strong PCNA band (36 kD). This band was comparable to that produced by the extract from fresh frozen cancer tissue. To determine the sensitivity of the method, we analyzed sequentially diluted lysates by Western blotting for PCNA. The protein band for PCNA could be detected even when the lysate extracted from 5-mm² tissue was diluted to 1:32 (Figure 2B).

Western blotting of nuclear proteins including mutant-type p53 (53 kD), cyclin D1 (36 kD), cyclin E (52 kD), cyclin-dependent kinase 2 (CDK2, 33 kD) and cyclin-dependent kinase 4 (CDK4, 34 kD), cytosolic β-catenin (92 kD), and membrane-bound E-cadherin (120 kD) also demonstrated clear protein bands at their expected molecular weight (Figures 3A and 3B).

Figure 2

SDS-PAGE and immunoblotting of PCNA. (A) Coomassie blue staining of electrophoresed gel. Although the SDS-PAGE pattern of the extract from paraffin section (Lane 1) is not identical to that obtained from fresh cancer tissue (Lane 2), the protein molecules ranging from 10 kD to 120 kD are well preserved. (B) Western blotting probed with anti-PCNA antibody. Extracts from both 5-cm-long 50-μm-thick normal mucosa (Lane 1) and 5-mm² 50-μm-thick cancer tissue (Lane 2) show a strong PCNA band (36 kD). The band is comparable to that produced by the extract from fresh frozen cancer tissue (Lane 8). The extract from cancer tissue is diluted to 1:2 (Lane 3), 1:4 (Lane 4), 1:8 (Lane 5), 1:16 (Lane 6), and 1:32 (Lane 7), with the band being detected even when the lysate is diluted to 1:32.

Figure 3

Western blotting of extracted protein probed with various antisera. (A) Electrophoresis in 12% polyacrylamide gel and immunoblotting of p53 (53 kD, Lane 1), cyclin D1 (36 kD, Lane 2), cyclin E (52 kD, Lane 3), CDK2 (33 kD, Lane 4), or CDK4 (34 kD, Lane 5). All these proteins are detected at their expected molecular weight, although nonspecific or crossreactive protein bands were detected in Lanes 2–5. For negative control, normal mouse IgG (Lane 6) or normal rabbit IgG (Lane 7) was added instead of primary antibodies. (B) Electrophoresis in 7.5% polyacrylamide gel and immunoblotting of β-catenin (92 kD, Lane 1) and E-cadherin (120 kD, Lane 2). These cytosolic or membrane-bound proteins are also detected at their expected molecular weights.

In our study, we could not detect the retinoblastoma gene product (pRB) or carcinoembryonic antigen (CEA) in the extracts from buffered formalin-fixed, paraffin-embedded tissue sections (data not shown).

Application of the New Method

We examined the expression levels of cyclin D1 and CDK2 in small adenomas and cancer tissues of the colorectum in comparison with the matched normal mucosa from four cases (Figure 4; Table 2). The definitive regions exhibiting moderate dysplasia were dissected from the adenomas in sizes ranging from 5 mm to 14 mm, from which enough protein could be extracted for Western blotting. Western blotting and densitometric analysis revealed that the average cyclin D1 expression of four adenomas was 2.1 times higher than that of matched normal mucosa, whereas CDK2 expression was only 1.4 times higher. Cyclin D1 expression in cancer tissues was 3.6 times higher and of CDK2 3.3 times higher.

Discussion

Buffered formalin is a popular fixative and has been used for fixation of surgically or endoscopically resected specimens (Fox et al. 1985). Many investigators have attempted to extract protein from formalin-fixed, paraffin-embedded sections, but its crosslinking effect made them give up extracting protein for Western blotting (Clark and Damjanov 1986; Conti et al. 1988). In this study, we aimed to extract protein efficiently from a small piece of buffered formalin-fixed, paraffin-embedded section for Western blotting.

Figure 4

Western blot analysis of cyclin D1 and CDK2 in normal mucosa (N), adenomas (A), and cancer tissues (C) in two cases. Cancer tissues show both cyclin D1 and CDK2 overexpression, whereas adenomas show only cyclin D1 overexpression.

The efficiency of protein extraction (164.2 μg/mg of dry cancer tissue) was more efficient than that from frozen cancer tissues (44.7 μg protein/mg of frozen cancer tissue on average for 10 samples) by the conventional method of homogenization and incubation in RIPA buffer containing 0.1% SDS at 0C. Although the extraction efficiency was about 50% of that previously reported for AMeX-processed tissues (Sato et al. 1992), in which tissues were fixed in acetone at -20C overnight, then cleared in methyl benzoate and xylene consecutively, it was enough for Western blotting. The protein band for PCNA could be detected even when the lysate extracted from 5-mm² tissue was diluted to 1:32. This indicated that even a 1-mm² sample of cancer tissue would be sufficient for detection of PCNA. Clark and Damjanov (1986) reported that not enough protein for Western blotting of keratin could be extracted from buffered formalin-fixed, paraffin-embedded tissue sections with incubation at 37C or RT for 1 hr. This was in sharp contrast to results obtained with extraction from tissues processed with Carnoy's solution (Clark and Damjanov 1986). Conti et al. (1988) also demonstrated that keratin could not be extracted from buffered formalin-fixed, paraffin-embedded tissue sections with incubation at 100C for 4 min, although extracts from the tissues processed through non-crosslinking fixatives, such as acetone, ethanol, or modified Carnoy's solution, showed a clear band of keratin. Our results indicate that, for successful protein extraction from formalin-fixed, paraffin-embedded tissues, it is necessary to incubate the tissues at high temperature for a long time.

Table 2.

Densitometric analysis of cyclin D1 and CDK2a

	Cyclin D1		CDK2
Case	A/N	C/N	A/N	C/N
1	2.7	4.0	2.5	5.1
2	1.0	3.0	1.7	3.4
3	3.1	4.0	1.1	2.4
4	1.7	3.5	0.4	2.2
Mean (-fold)	2.1	3.6	1.4	3.3

N, normal mucosa; A, adenoma; C, cancer.

Not only PCNA but also many kinds of nuclear, cytosolic, and membrane-bound proteins were detected at their expected molecular weight. The data indicate that the full length of these proteins was well preserved by formalin fixation and could be extracted without degradation even after several years of storage. This appears to be reasonable in view of the fact that formalin fixation does not alter the secondary protein structure, i.e., the regular arrangement of the polypeptide backbone, whereas it might change the tertiary or quaternary structure, i.e., the three-dimensional structure or the structure of aggregates of globular proteins (Bell et al. 1987; Mason and O'Leary 1991).

To our surprise, P53, CDK4, E-cadherin, and β-cate-nin, which could not be immunohistochemically detected in the formalin-fixed tissue sections by the antibodies employed in this study (Shimaya et al. 1993; Cheng et al. 1996), were clearly identified by Western blotting of the tissue extracts. This suggests that incubation of sections in RIPA buffer containing 2% SDS at high temperature might lead to the dissociation of crosslinking, thus untangling the protein molecules and recovering the antigenicity. Recently, enzymatic and nonenzymatic methods of antigen retrieval for immunohistochemistry of formalin-fixed sections have been reported. Heating by microwave oven is a frequently employed technique (Shi et al. 1991; Van den Berg et al. 1993). However, little is known about the exact conformational changes after formalin fixation or the mechanism of antigen retrieval by microwave oven heating. It is suspected that microwave heating affects tissue proteins mainly by raising their temperature (Cattoretti et al. 1993), and this may disrupt the crosslinks produced by formalin (Norton 1993). Our method of protein extraction appears to be based on a similar mechanism of antigen retrieval for immunohistochemistry.

In our study we could not clearly detect the retinoblastoma gene product (pRB) or carcinoembryonic antigen (CEA) in the extracts from buffered formalin-fixed, paraffin-embedded tissue sections (data not shown). pRB is known to be present at a very low level, and it appeared necessary to apply much greater amounts of the extracts for detection of pRB. CEA is produced in large amounts in colon tissue but is tightly attached to the membrane via a glycosyl–phosphatidylinositol anchor (Hefta et al. 1988; Jean et al. 1988; Sack et al. 1988; Takami et al. 1988), and it is considered essential to digest the extracts with phosphatidylinositol-specific phospholipase C (Matsuoka et al. 1990). We believe that further modification of our technique described here will expand its applicability to a wide range of detection of cell proteins.

This new technique provides many advantages for investigation of protein expression. The most valuable is the ability to extract protein from a small region under histological observation. In this study we could examine the expression levels of cyclin D1 and CDK2 in small adenomas and cancer tissues of the colorectum in comparison with the matched normal mucosa. The values for CDK2 overexpression compare well with our previous data obtained with conventional protein extraction from fresh frozen tissues (Yamamoto et al. 1995), which showed that CDK2 was 2.48-fold over-expressed in colorectal cancer tissue compared with matched normal mucosa. Another advantage of our technique is less contamination of normal cells in the process of tissue sampling. It is possible to exclude the muscularis propia and regions infiltrated by many lymphocytes with this method, whereas it is doubtful that unnecessary tissues can be excluded in macroscopic sampling using the conventional method. Because of this selective sampling, protein expression can be evaluated more accurately in densitometric analysis.

In conclusion, the technique of protein extraction from buffered formalin-fixed, paraffin-embedded tissue described here should provide a powerful tool for protein analysis, because large-scale archives of pathological specimens routinely fixed in formalin can be made available and its application could lead to detailed analysis of protein expression.

References

Bell

Rundquist

Svensson

Collins

(1987) Use of cytofluorometry to evaluate binding of antibodies to the cytoskeleton of cultured cells. J Histochem Cytochem 35:1381–1388

Cattoretti

Pileri

Parravicini

Becher

MHG

Poggi

Bifulco

Key

D'amato

Sabattini

Feudale

Reynolds

Gerdes

Rilke

(1993) Antigen unmasking on formalin-fixed, paraffin-embedded tissue sections. J Pathol 171:83–98

Cheng

Nagabhushan

Pretlow

Amini

Pretlow

(1996) Expression of E-cadherin in primary and metastatic prostate cancer. Am J Pathol 148:1375–1380

Clark

Damjanov

(1986) Immunoblotting of keratin polypeptides extracted from tissues preserved in standard histologic fixatives. J Histochem Cytochem 34:679–682

Conti

Larcher

Chesner

Aldaz

(1988) Polyacrylamide gel electrophoresis and immunoblotting of proteins extracted from paraffin-embedded tissue sections. J Histochem Cytochem 36:547–550

Dell'Orco

Jupe

Manjeshwar

Lie

X-T

McClung

King

Hollingsworth

Lightfoot

Kern

(1997) Phohibitin: a new biomarker for breast tumors. Breast J 3:85–89

Fox

Johnson

Whiting

Roller

(1985) Formaldehyde fixation. J Histochem Cytochem 33:845–853

Hefta

LJS

Lee

Paxton

Shively

(1988) Carcinoembryonic antigen is anchored to membranes by covalent attachment to a glycosyl–phosphatidylinositol moiety: identification of the ethanolamine linkage site. Proc Natl Acad Sci USA 85:4648–4652

Jean

Malapert

Rougon

Barbet

(1988) Cell membrane, but not circulating carcinoembryonic antigen is linked to a phosphatidylinositol-containing hydrophobic domain. Biochem Biophys Res Commun 155:794–800

10.

Lowry

Rosebrough

Farr

Randall

(1951) Protein measurement with the Folin phenol reagent. J Biol Chem 193:265–275

11.

Mason

O'Leary

(1991) Effects of formaldehyde fixation on protein secondary structure: a calorimetric and infrared spectroscopic investigation. J Histochem Cytochem 39:225–229

12.

Matsuoka

Matsuo

Sugano

Ohkura

Kuroki

(1990) Characterization of carcinoembryonic antigen-related antigens in normal adult feces. Jpn J Cancer Res 81:514–519

13.

Matsushime

Ewen

Strom

Kato

Hanks

Roussel

Sherr

(1992) Identification and properties of an atypical catalytic subunit (p34PSKJ3/CDK4) for mammalian D-type G1 cyclins. Cell 71:323–334

14.

Marx

(1994) How cells cycle toward cancer. Science 263:319–321

15.

Meyerson

Enders

C-L

L-K

Gorka

Nelson

Harlow

Tsai

L-H

(1992) A family of human cdc2-related protein kinases. EMBO J 11:2909–2917

16.

Nagafuchi

Takeichi

(1988) Cell binding function of E-cadherin is regulated by the cytoplasmic domain. EMBO J 7:3679–3684

17.

Neubauer

Effert

Iglehart

Frye

Liu

(1992) Analysis of gene amplification in archival tissue by differential polymerase chain reaction. Oncogene 7:1019–1025

18.

Norton

(1993) Microwave oven heating for antigen unmasking in routinely processed tissue sections. J Pathol 171:79–80

19.

Ohue

Tomita

Monden

Fujita

Fukunaga

Tamaki

Yana

Ohnishi

Enomoto

Inoue

Shimano

Mori

(1994) A frequent alteration of p53 gene in carcinoma in adenoma of colon. Cancer Res 54:4798–4804

20.

Ozawa

Baribault

Keimer

(1989) The cytoplasmic domain of the cell adhesion molecule uvomorulin associates with three independent proteins structurally related in different species. EMBO J 8:1711–1717

21.

Sack

Gum

Low

Kim

(1988) Release of carcinoembryonic antigen from human colonic cancer cells by phosphatidylinositol-specific phospholipase C. J Clin Invest 82:586–593

22.

Sato

Mukai

Furuya

Kameya

Hirohashi

(1992) The AMeX method: a multipurpose tissue-processing and paraffin-embedding method. Am J Pathol 140:775–779

23.

Shi

Key

Kalra

(1991) Antigen retrieval in formalinfixed, paraffin-embedded tissue: an enhancement method for immunohistochemical staining based on microwave oven heating of tissue sections. J Histochem Cytochem 39:741–748

24.

Shimaya

Shiozaki

Inoue

Tahara

Monden

Shimano

Mori

(1993) Significance of p53 expression as a prognostic factor in esophageal squamous cell carcinoma. Virchows Arch [A] 422:271–276

25.

Takami

Misumi

Kuroki

Matsuoka

Ikehara

(1988) Evidence for carboxyterminal processing and glycolipid-anchoring of human carcinoembryonic antigen. J Biol Chem 263: 12716–12720

26.

Van den Berg

Baas

Polak

Offerhaus

(1993) Detection of p53 overexpression in routinely paraffin-embedded tissue of human carcinomas using a novel target unmasking fluid. Am J Pathol 142:381–385

27.

Vogelstein

Fearon

Hamilton

Kern

Preisinger

Leppert

Nakamura

White

Smith

AMM

Bos

(1988) Genetic alterations during colorectal-tumor development. N Engl J Med 319:525–532

28.

von Weizsäcker

Labeit

Koch

Oehlert

Gerok

Blum

(1991) A simple and rapid method for the detection of RNA in formalin-fixed, paraffin-embedded tissues by PCR amplification. Biochem Biophys Res Commun 174:176–180

29.

Xiong

Connoly

Futcher

Beach

(1991) Human D-type cyclin. Cell 65:691–699

30.

Yamamoto

Monden

Ikeda

Izawa

Fukuda

Fukunaga

Tomita

Shimano

Shiozaki

Monden

(1995) Coexpression of cdk2/cdc2 and retinoblastoma gene products in colorectal cancer. Br J Cancer 71:1231–1236