PATIENTS

Thirteen patients who underwent coloproctectomy consecutively in the period from June 1992 to September 1994 were chosen for these investigations. Ten patients were men and three were women, median age 41 years (29–75 years). They suffered from long standing ulcerative colitis (10–30 years) and were followed up with regular biopsies for evaluation of dysplasia. Nine of the patients were resected because of persisting dysplastic changes, whereas four had developed carcinoma. Three of them had been followed up in other hospitals, dropped out of the surveillance programme for a period of two to three years, and then returned with carcinoma. The fourth patient developed a Dukes’s stage A carcinoma between two control visits 12 months apart. Two patients with a dysplasia associated lesion or mass (DALM)17 had been followed up in other cities until the lesions were detected.

COLECTOMY SPECIMENS

The proctocolectomy specimens showed different types of epithelial changes, including dysplasia and cancer. All specimens displayed more or less inactive colitis with little active inflammation. Four patients had adenocarcinomas: one had a Dukes’s A carcinoma in the caecum, another a Dukes’s C in the sigmoid, a third patient presented a multifocal and mucinous Dukes’s C carcinoma with three tumours in the transverse and descending colon, and the fourth had a Dukes’s D adenocarcinoma with primary tumour in the transverse colon. One patient had a large DALM (diameter 8 cm) with intramucosal carcinoma in the transverse colon, and one patient showed a DALM (4 cm) with high grade dysplasia in the rectum (table 1).

All colectomy specimens were opened longitudinally and rinsed in ice cold phosphate buffered saline (PBS), pH 7.6, immediately after removal, and thereafter placed in a bucket of PBS which was immersed in ice. Within eight hours tissues from eight or nine segments were removed; neighbouring tissue blocks were either fixed in 96% ethanol at 4°C18 or formalin, embedded in paraffin wax or OCT preservative (Miles Inc., Elkhart, Indiana, USA) for freezing. Finally, mucosa samples in close proximity to the tissue blocks were scraped off and prepared as single cell suspensions for subsequent DNA extraction.

HISTOPATHOLOGICAL LESIONS

For each colectomy sample, the histopathological lesions of interest were first identified on routinely stained slides and classified according to Riddell et al.19 In addition, two variants with hypermucinous epithelium were defined, one of them described by Morsonet al: villous surface with high number of goblet cells, but without nuclear atypia.20 The other variant had abundant goblet cells, no nuclear atypia, but a flat surface. Altogether, 161 mucosal lesions were selected (table2).

MICRODISSECTION

Parallel sections were cut with the microtome set at 6 μm, and the slides dried overnight at 37°C. Corresponding areas of interest were delineated and microdissected under a Leica dissection stereomicroscope after rapid staining with haematoxylin and eosin. Thereafter the tissue was scraped off the slide (the sections were covered by 25 μl Tris buffer, 0.05 mol), with the tip of a sealed glass pipette and then sucked into a microcapillary. Figures 1A and B show the same slide before and after microdissection. Tissue samples were then put into Eppendorf tubes and incubated with proteinase K at 37°C overnight. Proteinase K activity was inactivated by heating to 95°C for 10 minutes, and the resulting solutions were used directly as templates for K-rasanalyses.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Section from large bowel mucosa of patient 2, rich in goblet cells, but without dysplasia, (A) before and (B) after microdissection.

DNA EXTRACTION

For DNA extraction from fresh mucosa in close vicinity to the tissue blocks, standard methods were used: after incubation with proteinase K at 37°C overnight the mucosa was extracted twice in phenol and twice in chloroform, followed by ethanol precipitation.

K-ras MUTATION ANALYSIS

K-ras mutations in codons 12 and 13 were identified in DNA from each of the microdissected areas, as well as in mucosa extracted from neighbouring areas, by means of a modification of the “enriched PCR method”7 combined with temporal temperature gradient electrophoresis (TTGE).21

Briefly, the initial PCR reaction (A) was performed in 25 μl of PCR amplification buffer (50 mmol/l potassium chloride, 10 mmol/l Tris HCl, 1.5 mmol/l magnesium chloride), with either 1 μl of template from microdissection or 50 ng of extracted DNA, 25 μmol of each deoxynucleoside triphosphate, 0.28 μgTaqStart Antibody (Clontec, Palo Alto, California, USA), 1.25 U of Taq DNA polymerase (Promega, Madison, Wisconsin, USA), and 0.04 μmol and 0.2 μmol of the oligonucleotide primers (Genosys Biotechnology, Houston, Texas, USA) 5K1 and 3K1 respectively: 5′-ACTGAATATAAACTTGTGGTCCATG GAGCT-3′ and 5′-TTATCTGTATCAAAGAA TGGTCCTGCACCA-3′. These primers give a PCR product of 170 bp. The 5K1 primer was designed (modified, mismatching bases are underlined) to induce two restriction sites in the wild type (WT) amplification product (N = any base): 5′-CCANNNNNNTGG-3′ which is recognised by endonucleaseBstXI, and involves WT codon 12. 5′-CCANNNNNNNNNTGG -3′ is recognised by endonucleaseXcmI, and involves WT codon 13.

The amplification was performed in a GeneAmp PCR System 9600 (Perkin-Elmer, Foster City, California, USA). The reaction was run for 20 cycles consisting of 30 seconds denaturation at 94°C, one minute annealing at 50°C, and two minutes elongation at 72°C. After the last cycle the tubes were kept at the elongation temperature for 10 minutes. A 5 μl aliquot of the amplification product was incubated with either 10 U of BstXI or 2 U ofXcmI. Digestion was performed as recommended by the supplier of both the endonucleases (New England Biolabs, Beverly, Massachusetts, USA), in a volume of 20 μl. A 5 μl aliquot of the digested amplification product was used directly as a template in an amplification reaction (PCR B) of 25 μl. This time the following primers were used: 5′-ATGACTGAATATAAACTTGTG and 5′-GCCCCGCGCCCGTCCCGCCGCCCCC GCCCG CTC TAT TGT TGG ATC ATA in final concentrations of 0.32 and 0.16 μmol each, together with PCR amplification buffer, 100 μmol of each deoxynucleoside triphosphate, and 1 U of Taqpolymerase.

The reverse GC clamped primer is nested to 3′K1, and these two primers give a PCR product of 111 bp.

After the first annealing step of two minutes at 95°C, the reaction was run for 35 cycles consisting of 30 seconds denaturation at 95°C, 30 seconds annealing at 58°C, and 30 seconds elongation at 72°C. After the last cycle the tubes were kept at the elongation temperature for five minutes. The PCR amplification products were verified on 2% agarose gel containing ethidium bromide, and analysed in ultraviolet light, prior to being subjected to temporal temperature gradient electrophoresis (TTGE).21

TTGE ANALYSIS

The TTGE analyses were performed with the DCode system (Bio-Rad, Hercules, California, USA), using 16 × 16 cm glass plates, with gel containing 10% acrylamide/bis with 6 mol/l urea in 1.25 × TAE buffer (50 mmol/l Tris, 25 mmol/l acetic acid, 1.25 mmol/l EDTA), according to the manufacturer’s recommendations. The PCR B products were electrophoresed at 130 V for five hours with a temperature range of 52–62°C. The gel was then stained in SYBR Green I (Molecular Probes Inc., Eugene, Oregon, USA) and photographed using an ultraviolet transilluminator at 300 nm.

DNA SEQUENCING

All mutation positive samples from the TTGE analysis were sequenced in order to identify the exact base substitution. PCR products were purified using the Qiaquick PCR purification kit (Qiagen, Hilden, Germany) and sequenced using dye primer cycle sequencing and AmpliTaq polymerase FS on an Applied Biosystems 377 DNA sequencer.

CONTROLS

DNA from colon carcinoma cell lines SW480 (Clontech, Palo Alto, California, USA) and HCT116 (American Type Culture Collection, ATCC, Rockville, Maryland, USA) with known K-rasmutations at codon 12 (GTT) and codon 13 (GAC) respectively, were used as positive controls in each of the parallel procedures. Negative controls, without DNA, were run as controls for contamination.

STATISTICAL ANALYSIS

Associations between different variables were examined using the χ² method with Yates’s correction; p values less than 0.05 were considered statistically significant.

HISTOPATHOLOGY

Each of the 161 microdissected areas was diagnosed by one experienced pathologist (SNA). To check for interobserver reproducibility, 59 adjacent sections were blindly and independently evaluated for different grades of dysplasia according to standard criteria for comparison,19 by two histopathologists (SNA and OPFC) with a κ value of 0.64, which means substantial agreement.22 The remainder of the slides were afterwards agreed on by OPFC. Intraobserver agreement (SNA) after three months was “almost perfect” (κ=0.82).

One third of the microdissected areas (n=57) consisted of carcinoma or high or low grade dysplasia, while 20 showed flat mucosa, indefinite for dysplasia. Thirteen areas from six large bowels contained a villous growth pattern with hypermucinous, distended goblet cells (fig 2), and 14 were non-villous, hypermucinous, and without dysplasia. Table 2shows the different types of lesions.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Mucosal specimen from patient 11, with villous growth pattern and elongated, distended goblet cells without cytological atypia. This area was K-ras mutation positive.

K-ras MUTATIONS

Altogether 261 K-ras analyses were performed on tissue from 13 colectomy specimens: 161 from microdissected tissue and 100 from DNA extracted from nearby mucosa. K-ras mutations were found in 21 (13%) and 10 (10%) of these, respectively. Table 3 shows a survey of microdissected lesions and mucosa specimens in each of the patients, and the number of K-ras mutation positive lesions. Six of the 13 large bowels did not show any K-ras mutation by analysis of microdissected lesions, while DNA from mucosa specimens was K-ras mutation negative in five bowels.

Twenty one of the 161 microdissected areas (13%) were K-ras mutation positive, 16 in codon 12 and five in codon 13. Only one of the four patients (25%) with a total of six carcinomas had a tumour containing a K-ras mutation. Four of the 20 carcinoma specimens analysed were from this tumour, taken at different levels of infiltration (table 4), and showing identical mutation. Four of 14 (28.5%) high grade lesions were K-rasmutation positive, while only one of 23 (4%) low grade lesions and one of 20 (5%) lesions indefinite for dysplasia in flat mucosa were positive (table 2).

Eight of 13 K-ras mutation positive lesions (61%) were found in mucosa with villous growth pattern with “hypermucinous” epithelium, very rich in large, distended goblet cells, having nuclei without definite cytological atypia. These had previously been defined as “indefinite for dysplasia”, as delineated by Morson et al.20The frequency of K-ras mutations in this entity was significantly higher than in low grade dysplasia (p<0.001, table 2), but did not reach significance versus high grade dysplasia. The mutation frequency in villous, hypermucinous mucosa was also significantly higher than compared with flat mucosa, indefinite for dysplasia (one of 20, p=0.002) and areas negative for dysplasia, with regenerating or atrophic mucosa (one of 57, p=0.001). Two of 14 (14%) flat, hypermucinous lesions without dysplasia were K-ras mutation positive (p=0.03). Figure 2shows one of the K-ras positive lesions in patient 11.

Table 4 shows the distribution and type of K-ras mutations in DNA from microdissected areas. Among the eight villous, hypermucinous lesions, four were from patient 11 and showed an identical mutation (12 AGT—serine), despite considerable distance (several cm) between each of them. The last four were from three different patients, two from patient 9 (with the same 12 AGT mutation), and the two last (patients 10 and 3) showing an aspartate mutation in codons 12 (GAT) and 13 (GAC), respectively.

Among the two patients with DALM, one was K-ras mutation negative (patient 5), and the other K-ras positive (patient 6), shown both in DNA from microdissection and mucosal “slicing”. Figure 3illustrates gels with mutated bands from K-ras investigations, while fig 4 shows DNA sequence analysis of wild type (A) and mutated (B) K-ras.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

TTGE gel showing migration patterns of K-ras exon 1 fragments. All lanes show different migration patterns at the upper bands (heteroduplex bands), indicating different K-ras mutations. The lower band is the homoduplex band. Lanes 1–5 show codon 12 mutations: (1) GTT (valine); (2) TGT (cysteine); (3) GAT (aspartate); (4) GCT (alanine); and (5) AGT (serine). Lane 6 shows a codon 13 GAC mutation (aspartate).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

(A) Automated sequence analysis of K-ras, showing wild type sequence GGT in codon 12. This is an inverse analysis, and should be read from right to left. (B). K-ras mutation in the second base in codon 12, giving the sequence GAT (aspartate) instead of GGT (glycine).