Expression of CD44 in Apc and Tcf Mutant Mice Implies Regulation by the WNT Pathway

Apc and Tcf-4 Mutant Mice

Normal and neoplastic small-intestinal tissue from C57BL/6JIco-Apc⁺/Apc1638N mice (Apc^+/−) and wild-type C57BL/6 (Apc^+/+) control mice was obtained from the Department of Human Genetics, University of Leiden, Leiden, The Netherlands. The Apc⁺/Apc1638N mice are heterozygous for an Apc chain-terminating mutation in Apc codon 1638, although the expected truncated protein is not detectable by conventional Western analysis. 42 The mice were sacrificed between 6 and 12 month of age, after which the entire intestine was opened longitudinally and inspected for neoplastic lesions. Lesions with surrounding normal tissue were sampled for routine processing and fixed in formalin or Notox (Earth Safe Industries, Bellemead, NJ) and embedded in paraffin. Embryos of Tcf-4^−/− and Tcf-4^+/− mice 43 were obtained from the Department of Immunology, University Hospital, Utrecht, The Netherlands, embedded in paraffin, and sectioned at 6 μm thickness.

FAP Patients

Colon mucosa biopsies from a 38-year-old male and a 30-year-old female FAP patient, taken at routine colonoscopy, were obtained from the files of the Department of Pathology, Academic Medical Center, University of Amsterdam. The biopsies were embedded in paraffin, and 6-μm sections were prepared.

Monoclonal Antibodies (MAbs)

The MAbs used were Hermes-3 against an epitope on the constant part of the human CD44 molecule (CD44s), 16 VFF18 against human CD44v6, 44 PGP-1 against mouse CD44s (Pharmingen, San Diego, CA), 10D1 against mouse CD44v4, 45 9A4 against mouse CD44v6, 45 and PCNA against proliferating cell nuclear antigen (PCNA; Dako, Glostrup, Denmark) (Figure 1) ▶ .

Immunohistochemistry

Detection of CD44 and PCNA in mouse and human paraffin-embedded tissue sections was performed as described previously, 35 with the following modifications. In brief, the sections were deparaffinated, rehydrated, and boiled in a citrate buffer (0.01 mol/L, pH 6, Merck 6448) for antigen retrieval. On mouse and human tissue, biotinylated rabbit-anti-rat F(ab′)₂ (Dako) and rabbit-anti-mouse F(ab′)₂ (Dako) were used as secondary antibodies, respectively. For color development, 3,3-diaminobenzidine tetrachloride (DAB; Sigma, Bornem, Belgium) was used.

The immunohistochemical staining was scored semiquantitatively based on the staining intensity of positively stained tumor cells. The samples were scored as follows: −, negative; −/+, equivocal/very weak; +, weak; ++, moderate; +++, strong.

Reverse Transcription Polymerase Chain Reaction and Southern Blot Analysis

RNA isolation and first-strand cDNA synthesis were performed as described previously. 46 Polymerase chain reaction (PCR) was performed with 1.5 U of Taq DNA polymerase (Gibco BRL/Life Technologies, Gaithersburg, MD), 300 μmol/L dNTPs (Pharmacia Biotech, Uppsala, Sweden), and 2 mmol/L MgCl₂ in 1X PCR buffer (both Gibco BRL/Life Technologies). Primers used were M44CU (5′-CCCAGGTAGCTTCCTTAACCC-3′) in combination with M44CD (5′-CGTAGAGAGGACCGTGACCGA-3′). PCR was started with a 5-minutes denaturation step at 95°C, after which amplification was performed at 35 cycles of denaturation at 95°C for 30 seconds, annealing at 55°C for 1 minute, and elongation at 72°C for 2 minutes. After a final elongation step for 10 minutes at 72°C, samples were cooled on ice. PCR products were resolved in 1.5% agarose/Tris-buffered ethanolamine gel and blotted on Hybond-N⁺ membranes (Amersham, Little Chalfont, UK).

To generate ³²P-labeled exon-specific probes, we used the plasmid pZeo SV mCD44v4-v10, containing the murine CD44 exon v4-v10 (a kind gift from Dr. M. Hofmann from the Institut für Genetics, Forschungszentrum, Karlsruhe, Germany). To generate a ³²P-labeled exon-v3 probe, we used DNA from normal mouse skin. For the generation of the CD44s probe we used the plasmid pZeo S mCD44st, containing the murine CD44 standard region. The PCR mixtures for the v3 and v9 exons contained 2 mmol/L MgCl₂, 100 μmol/L dATP, dTTP, and dGTP, and 13.2 μmol/L dCTP. The mixtures for the v6 exon contained 1 mmol/L MgCl₂, 200 μmol/L dATP, dTTP, and dGTP, and 26.4 μmol/L dCTP. The mixtures for the standard region contained 2 mmol/L MgCl₂, 300 μmol/L dATP, dTTP, and dGTP, and 39.6 μmol/L dCTP. In addition, all PCR mixtures contained 0.22 MBq [α-³²P]dCTP, 10 pmol of each oligonucleotide primer, and 1.5 U of Taq DNA polymerase. The primers used for v3 were MV3U (5′ GTACGGAGTCAAATACCAAC3′) and MV3D (5′ TGGTACTGGAGATAAAATCT 3′), for v6 were MV6U (5′ CTCCTAATAGTACAGCAGAA 3′) and MV6D (5′ AGTTGTCCCTTCTGTCACAT 3′), and for v9 were MV9U (5′ CACAGAGTCATTCTAGAAC 3′) and MV9D (5′ TGCTAGATGGCAGAATAGAA 3′). Samples were amplified for 35 cycles in a PTC-100 (MJ Research, Watertown, CA). Each cycle consisted of denaturation at 95°C (for CD44s, v3, and v9) or at 96°C (for v6) for 30 seconds, annealing at 55°C (CD44s, v3, and v9) or 50°C (V6) for 1 minute, and extension at 72°C for 2 minutes (for CD44s, v3, and v9) or 1 minute (for V6), followed by a final elongation step for 10 minutes at 72°C. The PCR products resulted in ³²P-labeled exon-specific probes and were used to hybridize the membranes according to standard procedures.

Expression of CD44 in the Normal and Neoplastic Intestinal Mucosa of Apc⁺/Apc1638N Mice

CD44 expression in the epithelium of the histologically normal small-intestinal mucosa of Apc⁺/Apc1638N (Apc^+/−) mice and wild-type (Apc^+/+) mice was identical and was restricted to the crypts (Figure 2A ▶ ; Table 1 ▶ ). In these crypt areas, CD44 expression was localized to the basolateral membranes of the cells. Epitopes encoded by the constant (standard) part of the CD44 (CD44s) and by the alternatively spliced exons CD44v4 and v6 were expressed in a similar pattern. However, the intensity of staining differed; whereas the MAbs against CD44s and CD44v6 stained with high intensity, staining with the anti-CD44v4 MAb was weak (Table 1) ▶ . In the lamina propria, CD44s, but not CD44v4 or CD44v6 expression, was observed on stromal cells, lymphocytes, and macrophages.

Intestinal tumors arising in Apc^+/− mice invariably showed a strong homogeneous expression of CD44 (Figure 2, C and D ▶ ; Table 1 ▶ ). This overexpression included CD44s as well as CD44v4- and v6-encoded epitopes and, importantly, was already observed in the earliest detectable neoplastic lesions, ie, in ACFs with dysplasia (Figure 2B) ▶ . It remained present in fully developed adenomas (Figure 2C) ▶ and invasive carcinomas (Figure 2D) ▶ . Analysis of CD44 mRNA in tumors versus normal mucosa by reverse transcription PCR and Southern blotting with exon-specific probes showed a preferential up-regulation of high molecular weight CD44 isoforms in tumor tissue (Figure 3) ▶ .

ACFs with Dysplasia in FAP Patients Overexpress CD44

Previous studies in human colorectal cancer have shown that CD44 overexpression is present in adenomas as well as in invasive carcinomas. 25,35,37 Based on our present observations in Apc-mutant mice, we explored whether dysplastic ACFs in familial adenomatous polyposis (FAP) patients also overexpress CD44. Indeed, a strong up-regulation of CD44, including both CD44s- and CD44v6-encoded epitopes, was observed in ACFs (Figure 4, B and D ▶ ; Table 1 ▶ ). Hence, as in Apc^+/− mice, in FAP patients, deregulation of CD44 expression is present in the earliest detectable neoplastic lesions of colorectal cancer.

Tcf-4 Mutant Mice Lack CD44-Expressing Cells in the Epithelial Lining of the Small Intestine

The above data imply an intimate link between loss of Apc tumor suppressor-protein function and CD44 overexpression in the intestinal mucosa and suggest that CD44 expression is directly or indirectly controlled by β-catenin/Tcf-4 mediated transcription. To further explore this hypothesis, we studied the expression of CD44 in the intestinal mucosa of Tcf-4^−/− mice. These mice have a striking small-intestinal phenotype with a selective loss of cycling intestinal crypt cells. As a result of tearing of the intestinal epithelial lining and fluid imbalance, this leads to perinatal death. 43 In Figure 5 ▶ , this absence of cycling cells is illustrated by using the proliferation marker PCNA; at day E18, numerous proliferating cells were present in the intervillous epithelium of (Tcf-4^+/+ and Tcf-4^+/−) control mouse embryos (Figure 5A ▶ ; Table 1 ▶ ). By contrast, the epithelium of the small intestine of the Tcf-4^−/− embryos was devoid of proliferating cells (Figure 5B) ▶ . Proliferation of cells in the lamina propria directly underneath the epithelium as well as in other organs (not shown) was not affected by the Tcf-4 disruption. In the (Tcf-4^+/+ and Tcf-4^+/−) control mice, CD44s (not shown) as well as CD44v6 expression (Figure 5C ▶ ; Table 1 ▶ ) was readily detectable at the basolateral surfaces of the pseudostratified intervillous epithelium. In sharp contrast, the intestinal epithelium of Tcf-4^−/− mice did not show any CD44 staining (Figure 5D) ▶ . This lack of CD44 expression was specific for the epithelium of the small intestine, as Tcf-4^−/− and control mice showed identical expression of both CD44s and CD44v6 in all other tissues, including the lamina propria of the small and large intestine, the epithelia of the large intestine, stomach, and epidermis and in lymphoid organs (not shown).