Multiple variants of the human lymphocyte homing receptor CD44 generated by insertions at a single site in the extracellular domain.

The human CD44 cell-surface glycoprotein participates in a wide variety of cell-cell interactions including lymphocyte homing and tumor metastasis. The CD44 antigen is known to display extensive size heterogeneity when compared between different tissue sources although the structural basis for this variation is not yet clear. Recently, two further isotypes in addition to the basic hemopoietic form of the CD44 antigen have been cloned and sequenced and these have been found to contain all or part of a 200-400-base pair insert within the extracellular domain, suggesting that the characteristic heterogeneity in the molecule may be generated by a mechanism of alternative splicing. We have obtained further evidence for alternative splicing, and we report here the cloning and sequencing of six different CD44 sequence variants from a variety of cell lines using a combination of expression cloning and the polymerase chain reaction. Comparison of these variants indicates that each is probably assembled by the insertion of five different exon units in tandem into a discrete site within the membrane proximal region of the extracellular domain. One of the variants contains an exon that shares extensive amino acid sequence homology with a recently described rat CD44 variant that mediates tumor metastasis. Another variant contains a new exon that encodes a tandem repeat of the consensus sequence SG for covalent modification with chondroitin sulfate and is expressed predominantly on mammary tumors. We suggest that a mechanism of alternative exon splicing generates much of the observed structural heterogeneity of CD44 and that the particular set of CD44 variants expressed in a single cell may represent a precise postal code directing the final destination of migrating cells and metastatic tumors.     

Trans-acting factors regulate the expression of CD44 splice variants.

Variant isoforms of the cell surface glycoprotein CD44 (CD44v) are expressed during development, in selected adult tissues and in certain metastatic tumor cells. CD44v differ from the standard isoform (CD44s) by up to ten additional exon sequences included by alternative splicing. By cell fusion experiments, we have obtained evidence for the existence of cell-type specific trans-acting factors recruiting CD44 variant exon sequences. Stable cell hybrids of CD44s and CD44v expressing cells indicated a dominant mechanism for variant-exon inclusion. In transient interspecies heterokaryons of human keratinocytes and rat fibroblasts, the ability of the keratinocytes to include all variant exon sequences in CD44 was conferred completely on the rat fibroblast nucleus. Fusions of cells with complex CD44 splice patterns do not permit interpretation of splice control by the relative abundance of a single trans-acting factor, but rather by (a) positively acting factor(s) recruiting variant exon sequences in the 3' to 5' direction and additional factors selecting individual exons. Since the pancreatic carcinoma cell line BSp73ASML (in contrast to the cervix carcinoma cell lines SiHa and ME180) could not transfer its specific splice pattern in cell fusions, we conclude that in some tumors, splicing is also controlled by mutation of cis-acting recognition sites.     

CD44 is not an adhesive receptor for osteopontin

Osteopontin is a secreted glycoprotein with adhesive and migratory functions. Cellular interactions with osteopontin are mediated through integrin receptors which recognize the RGD domain. Recently, CD44, a non-integrin, multifunctional adhesion molecule was identified as an osteopontin receptor. CD44 is a ubiquitous surface molecule that exists as a number of different isoforms, generated by alternative splicing. To analyze which forms of CD44 mediate binding to osteopontin, we used the standard form of CD44 as CD44-human immunoglobulin fusion proteins and several splice variants in enzyme-linked immunosorbant assays. Multiple preparations of osteopontin were used including native osteopontin derived from smooth muscle cells, human urinary osteopontin, full-length recombinant osteopontin, and two recombinant osteopontin fragments expected to be formed following thrombin cleavage. Our data show that although the CD44-hlg fusion proteins could interact with hyaluronic acid as expected, there was no interaction between CD44H, CD44E, CD44v3,v8-v10, or CD44v3 with osteopontin. These studies suggest that CD44-osteopontin interactions may not be common in vivo and may be limited to a specific CD44 isoform(s), and/or a particular modified form of osteopontin.     

Splicing choice from ten variant exons establishes CD44 variability.

The enormous heterogeneity of the surface protein designated CD44 is in part due to posttranslational modification, and in part due to differential splicing. Alternative splicing occurs within one particular region encoding the extracellular portion of the protein. Comparison of various cDNA clones with different 'inserts' in this variable region with sequences of genomic clones from the mouse has revealed the existence of at least ten exons from which sequences are chosen by alternative splicing. Various combinations of these exons account for the tremendous heterogeneity of CD44 molecules expressed in different tissues, and in progressing tumor cells. The existence of different isoforms of CD44 suggests a broad spectrum of yet unknown physiologic functions.     

Multisite and bidirectional exonic splicing enhancer in CD44 alternative exon v3

The human CD44 gene encodes multiple isoforms of a transmembrane protein that differ in their extracellular domains as a result of alternative splicing of its variable exons. Expression of CD44 is tightly regulated according to the type and physiological status of a cell, with expression of high molecular weight isoforms by inclusion of variable exons and low molecular weight isoforms containing few or no variable exons. Human CD44 variable exon 3 (v3) can follow a specific alternative splicing route different from that affecting other variable exons. Here we map and functionally describe the splicing enhancer element within CD44 exon v3 which regulates its inclusion in the final mRNA. The v3 splicing enhancer is a multisite bipartite element consisting of a tandem nonamer, the XX motif, and an heptamer, the Y motif, located centrally in the exon. Each of the three sites of this multisite enhancer partially retains its splicing enhancing capacity independently from each other in CD44 and shows full enhancing function in gene contexts different from CD44. We further demonstrate that these motifs act cooperatively as at least two motifs are needed to maintain exon inclusion. Their action is differential with respect to the splice-site target abutting v3. The first X motif acts on the 3' splice site, the second X motif acts on both splice sites (as a bidirectional exonic splicing enhancer), and the Y motif acts on the 5' splice site. We also show that the multisite v3 splicing enhancer is functional irrespective of flanking intron length and spatial organization within v3.     

CD44v6: a target for antibody-based cancer therapy

The human CD44 gene encodes type 1 transmembrane glycoproteins involved in cell-cell and cell-matrix interactions. The structural heterogeneity of the gene products is caused primarily by alternative splicing of at least 10 out of 20 exons. Certain CD44 variant isoforms, in particular those containing CD44 variant domain 6 (CD44v6), have been implicated in tumourigenesis, tumour cell invasion and metastasis. Here we will give an overview of immunohistochemically determined CD44v6 expression in human malignancies (primary epithelial and nonepithelial tumours as well as metastases) and normal tissues, and review several examples of the clinical use of CD44v6-specific antibodies. In nonmalignant tissues, CD44v6 expression is essentially restricted to a subset of epithelia. Intense and homogeneous expression of CD44v6 was reported for the majority of squamous cell carcinomas and a proportion of adenocarcinomas of differing origin, but was rarely seen in nonepithelial tumours. This expression pattern has made CD44v6 an attractive target for antibody-guided therapy of various types of epithelium-derived cancers.Abbreviations CD44 type 1 transmembrane glycoprotein, cell surface receptor for hyaluronate - CD44s (CD44H) standard form of CD44 - CD44v6 splice variant exon 6 of CD44 - CTC common toxicity criteria - 2F10, VFF4, VFF7, VFF18 (BIWA 1), U36, V6B3, HB-256, Var 3.1 monoclonal antibodies targeting the CD44v6 antigen - SCC squamous cell carcinoma     

The keratan sulfate-enriched region of bovine cartilage proteoglycan consists of a consecutively repeated hexapeptide motif

We have determined the sequence of a cDNA clone encoding the keratan sulfate-rich domain of the large aggregating cartilage proteoglycan core protein. The C-terminal portion of the deduced amino acid sequence is homologous to the chondroitin sulfate-rich region (domain CS1) of the rat chondrosarcoma proteoglycan, and the N-terminal portion is homologous to the second globular domain (G2) of the rat proteoglycan (Doege, K., Sasaki, M., Horigan, E., Hassell, J. R., and Yamada, Y. (1987) J. Biol. Chem. 262, 17757-17767). We could identify, inserted between these regions, a region absent in the rat proteoglycan. This domain corresponds to the keratan sulfate-enriched region of the bovine proteoglycan. It consists of a highly conserved hexapeptide motif consecutively repeated 23 times. Transfer blot analysis of genomic DNA indicated a single gene. The coding region for the keratan sulfate-enriched region was present both in human and bovine DNA, whereas the coding region for this domain appears to be absent in the rat genome. Transfer blot analysis of RNA showed that the keratan sulfate-rich region is present in proteoglycans from fetal as well as adult sources. Furthermore, RNA protection assays of RNA isolated from adult and fetal bovine articular cartilage showed that no alternative splicing occurs within this keratan sulfate-enriched region. These experiments show that the fetal bovine cartilage proteoglycan contains the keratan sulfate attachment domain, although it lacks the keratan sulfate side chains.     

Analysis of the genomic structure of the porcine CD1 gene cluster

CD1 is an MHC class I-like protein that presents lipid antigens to T cell receptors. We determined 470,187 bp of the genomic sequence encompassing the region encoding porcine CD1 genes. We identified 16 genes in this region and newly identified CD1A2, CD1B, CD1C, CD1D, and CD1E. Porcine CD1 genes were located in clusters between KIRREL and olfactory receptor (OR) genes, as observed in humans, although they were divided into two regions by a region encoding OR genes. Comparison of the genomic sequences of CD1 gene loci in pigs with other mammals showed that separation of the CD1 gene cluster by ORs was observed only in pigs. CD1A duplication in the porcine genome was estimated to have occurred after the divergence of the human and porcine. This analysis of the genomic sequence of the porcine CD1 family will contribute to our understanding of the evolution of mammalian CD1 genes.     

Proteoglycan forms of the lymphocyte homing receptor CD44 are alternatively spliced variants containing the v3 exon

The CD44 cell surface glycoprotein is expressed on a broad range of different tissues as multiple isoforms containing from one to ten alternatively spliced exons v1-v10 inserted within the extracellular domain. Differential glycosylation generates still further variability, yielding both N- and O-glycan-modified forms of CD44 in addition to proteoglycan-like variants containing chondroitin sulphate and heparan sulphate. These high molecular mass proteoglycan-like variants, previously identified in lymphocytes, melanomas, and keratinocytes have been implicated in cell-matrix adhesion, cell motility, and invasiveness. More recently, monocyte CD44 molecules presumed to carry glycosaminoglycan chains were shown to bind the chemokine MIP-1 beta (Tanaka, Y.,D. H. Adams, S. Hubscher, H. Hirano, U. Siebenlist, and S. Shaw. 1993. Nature (Lond). 361:79-82.) raising the intriguing possibility that proteoglycan-like CD44 variants might play a role in regulating inflammatory responses. Here we have investigated the molecular identity of these proteoglycan-like CD44 variants by generating a panel of recombinant CD44 isoforms using a novel cassette cloning strategy. We show that both chondroitin and heparan sulphate modifications are associated specifically with isoforms (CD44v3-10 and CD44v3,8-10) containing the v3 alternative exon which encodes a consensus motif SGXG for GAG addition. Other isoforms (CD44v10, CD44v8- 10, CD44v7-10, and CD44v6-10) are shown to lack these GAG chains but to carry extensive O-glycan modifications, most likely within the mucin- like alternative exon inserts. We also demonstrate that the majority of endogenous GAG-modified CD44 isoforms present in epithelial cells constitute v3 isoforms thus establishing that in these cells the majority of proteoglycan-like CD44 variants are generated by alternative splicing. Finally we present evidence using transfected B lymphoma cells that the GAG-modified CD44 isoforms CD44v3-10 and CD44v3,8-10, unlike CD44H, bind only weakly to hyaluronan. Together with the demonstration in the accompanying paper (Bennett, K., D. G. Jackson, J.C. Simon, E. Tanczos, R. Peach, B. Modrell, I. Stamenkovic, G. Plowman, and A. Aruffo. 1995. J. Cell Biol. 128:687-698.), that CD44 molecules containing the v3 exon bind growth factors, these results highlight a new and potentially important role for CD44 alternative splicing in the control of cell-surface proteoglycan expression.     

Anti-tumor effect against human cancer xenografts by a fully human monoclonal antibody to a variant 8-epitope of CD44R1 expressed on cancer stem cells

Background

CD44 is a major cellular receptor for hyaluronic acids. The stem structure of CD44 encoded by ten normal exons can be enlarged by ten variant exons (v1-v10) by alternative splicing. We have succeeded in preparing MV5 fully human IgM and its class-switched GV5 IgG monoclonal antibody (mAb) recognizing the extracellular domain of a CD44R1 isoform that contains the inserted region coded by variant (v8, v9 and v10) exons and is expressed on the surface of various human epithelial cancer cells.

Methods and Principal Findings

We demonstrated the growth inhibition of human cancer xenografts by a GV5 IgG mAb reshaped from an MV5 IgM. The epitope recognized by MV5 and GV5 was identified to a v8-coding region by the analysis of mAb binding to various recombinant CD44 proteins by enzyme-linked immunosorbent assay. GV5 showed preferential reactivity against various malignant human cells versus normal human cells assessed by flow cytometry and immunohistological analysis. When ME180 human uterine cervix carcinoma cells were subcutaneously inoculated to athymic mice with GV5, significant inhibition of tumor formation was observed. Furthermore, intraperitoneal injections of GV5markedly inhibited the growth of visible established tumors from HSC-3 human larynx carcinoma cells that had been subcutaneously transplanted one week before the first treatment with GV5. From in vitro experiments, antibody-dependent cellular cytotoxicity and internalization of CD44R1 seemed to be possible mechanisms for in vivo anti-tumor activity by GV5.

Conclusions

CD44R1 is an excellent molecular target for mAb therapy of cancer, possibly superior to molecules targeted by existing therapeutic mAb, such as Trastuzumab and Cetuximab recognizing human epidermal growth factor receptor family.     

Complete exon-intron organization of the human leukocyte common antigen (CD45) gene

Ten genomic DNA clones encoding the human leukocyte common Ag (LCA, CD45) gene were isolated by screening human genomic DNA libraries with LCA cDNA probes. One genomic DNA clone contains the promoter region and the first two exons, as determined by primer extension analyses and S1 nuclease protection studies as well as nucleotide sequence determination. The first exon does not encode a peptide, while the second exon contains the initiation ATG codon and encodes the signal peptide. The other nine genomic DNA clones, which are separated from the first genomic clone by an unknown distance, are connected and span a total of 73 kb. The nine connected genomic clones encode a total of 31 exons. The 33 exons encoded by these 10 genomic clones account for the entire cDNA sequences including the 5' and 3' untranslated sequences. Exon 3 and exons 7 through 15 encode the extracellular domain sequences that are common to all LCA isoforms. Differential usage of exons 4, 5, and 6, generates at least five distinct LCA isoforms. Exon 16 encodes the transmembrane peptide. The cytoplasmic region of the leukocyte common antigens is composed of two homologous domains. Exons 17 through 24 encode the first domain, and exons 25 through 32 encode the second domain. The comparison of these exons indicated that the homologous domains were generated by duplication of several exons. The most 3' exon (exon 33) encodes the carboxy terminus of the LCA molecules and includes the entire 3' untranslated sequence.     

Alternative splicing of the primary transcript generates heterogeneity within the products of the gene for Bombyx mori chitinase

The gene of chitinase in the silkworm, Bombyx mori, generates four mRNA products by alternative splicing. Nucleotide sequences of the entire gene for chitinase and respective cDNAs demonstrate that the pre-mRNA undergoes alternative splicing at both the 5' and 3' regions. At the 5' region, the pre-mRNA experienced differential splicing through two alternative 5'-intron consensus splicing sites. These products differ in the last amino acid of the signal peptide and the first amino acid of the mature N-terminal sequences: one with Cys(20)-Ala(21) and the other with Ser(20)-Asp(21). The product with Cys(20)-Ala(21) residues is one amino acid larger than the other with Ser(20)-Asp(21). At the 3' region the pre-mRNA of the chitinase gene undergoes alternative splicing in three different fashions. It is spliced either through retaining or excluding the upstream 121-bp direct repeat found at the 3' region of the coding sequences or through retaining or excluding of an insertion of 9 bp in a combinatorial manner. Retention or exclusion of the upstream 121-bp direct repeat results in a protein with a deduced amino acid sequence similar in size to the one retaining both direct repeats. However, exclusion of the insert of the 9 bp from the mRNA results in a protein with 22 extra amino acids. All of the mRNA products appear to be generated from a single gene as demonstrated by testing the 3' region of the genomic DNA and variant chitinase mRNA products. B. mori chitinase expression in the fifth instar larvae epidermal tissues appears to be developmentally regulated, but the phenomenon of alternative splicing of the pre-mRNA is not stage-dependent. Furthermore, the four mRNA products showed chitinase activity when expressed in Escherichia coli, which demonstrates the role of the alternative splicing process in generating multiple isoforms of the silkworm's chitinase.