sel proliferation

8/8/2019 Sel Proliferation

1/17

JOURNAL OF CELLULAR PHYSIOLOGY 208:2338 (2006)

SEL1L a Multifaceted Protein Playing a Role inTumor Progression

IDA BIUNNO,1,3 MONICA CATTANEO,1 ROSARIA ORLANDI,2 CRISTINA CANTON,1LAURA BIAGIOTTI,1 STEFANO FERRERO,4 MASSIMO BARBERIS,5 SERENELLA M. PUPA,2

ALDO SCARPA,6 AND SYLVIE MENARD2*1Istituto di Tecnologie Biomediche, CNR, Segrate-Milano, Italy

2Department of Experimental Oncology, Molecular Targeting Unit,Istituto Nazionale Tumori, Milan, Italy

3BioRep, Milano, Italy4Department of Pathology, University of Milan School of Medicine,

AOS Paolo, Milano, Italy5MultiLab, Gruppo MultiMedica, Milano, Italy

6Dipartimento di Patologia, Universita di Verona, Verona, Italia

Sincethecloningin1997ofSEL1L,thehumanorthologofthe sel-1 geneofC. elegans,moststudieshavefocusedonitsroleincancerprogression and have provided significant evidences to link its increased expression to a decrease in tumor aggressiveness. SEL1L

resides on a Genome Desert area on chromosome 14q24.3-31 and is highly conserved in evolution. The function of the SEL1Lencoded protein is still very elusive although, several evidences from lower organisms indicate that it plays a major role in proteindegradation using the ubiquitin-proteosome system. SEL1L has a very complex structure made up of modules: genomically itconsists of 21 exons featuring several alternative transcripts encoding for putative protein isoforms. This structural complexityensures protein flexibility and specificity, indeed the protein was found in different sub-cellular compartments and may turn on aparticular transcript in response to specific stimuli. The overall architecture of SEL1L guarantees an exquisite regulation in theexpression of the gene. J. Cell. Physiol. 208: 23 38, 2006. 2005 Wiley-Liss, Inc.

SEL1L, (sel-1 like gene, Table 1), the human orthologof the C. elegans sel-1 gene, opens a new line of researchaimed to define the role of the encoded protein in theprocesses of normal cell fate decision, cell differentia-tion, and transformation. SEL1L may be the first of anew class of molecules associated with cell fate determi-nation and involved in cancer progression. While mucheffort has been devoted to define the genetic changesthat can lead a cell to proliferate, little is known aboutthe role that cell determinant molecules play in cancerprogression. Most of the genes found to be relevant incell differentiation and cell fate determination areevolutionarily conserved from simple organisms suchas C. elegans andDrosophila to higher primates, SEL1Lmay be one of them. SEL1L shows a high degree of cross-species conservation in its nucleotide and proteinsequence (Biunno et al., 2002) indicating the remark-able selective evolutionary pressure to keep the genefairly intact. To date, neither functional mutations nor

genomic alterations have been reported for SEL1L,making it an essential protein whosealteration(s) maybe lethal for the survival of an organism (Larsen et al.,2001; Saltini et al., 2004).

SEL1L is an intriguing gene whose function is still ata speculative or embryonic stage, but displays severalinteresting characteristics. The C. elegans gene, sel-1(suppressor-enhancer-lin), is consideredto be a negativeregulator of the lin-12/Notch family of proteins (Grantand Greenwald, 1996), probably by acting on theirturn over. The Notch pathway is involved in cellfate decisions in a variety of tissues and in multipleorganisms, and in the maintenance of the proliferative/differentiativebalance in many cell lineages (Artavanis-Tsakonas et al., 1999); in human can cause T-cell

lymphomas when introduced into mouse bone marrowstems cells (Pear et al., 1996) and can transform rat

kidney cells in co-operation with the adenovirusE1A oncogene (Capobianco et al., 1997; Gallahan andCallahan, 1997). No genetic lesions of the Notchlocus have been described in human tumors with theexception of a rare translocation in T cell malignancies(Ellisen et al., 1991).

SEL1L shares similarities with the yeast Hrd3pprotein, a member of endoplasmic-reticulum (ER)-associated protein degradation (ERAD)-specific ubiqui-tin ligase complex that functions in the detection,targeting, and degradation of malfolded and normalendoplasmic reticulum (ER) resident proteins (Hamp-ton et al., 1996; Hampton, 2002). The human and plantSEL1L genes belong to the so-called unfolded proteinresponse genes which are induced in response to theaccumulation of unfolded and misfolded proteins in theER stress (Kaneko and Nomura, 2003; Kamauchi et al.,2005).ER stresshas been implicated in thepathogenesisof a variety of human disorders, including neural

degenerative diseases, diabetes, viral pathogenesis,and cancer (Kaufman, 2002; Dissemond et al., 2004). Ithas been postulated that in cancer, misfolding of keyproteins can cause the cells to lose tumor suppression

2005 WILEY-LISS, INC.

Monica Cattaneo and Rosaria Orlandi have equal merit for thework.

*Correspondence to: Dr. Sylvie Menard, Department of Experi-mental Oncology, Molecular Targeting Unit, Istituto Nazionaleper lo Studio e la Cura dei Tumori, Via Venezian 1, Milan, Italy.E-mail: [email protected]

Received 13 October 2005; Accepted 20 October 2005

DOI: 10.1002/jcp.20574


2/17

functions (Qu and Koromilas, 2004; Stavridi andHalazonetis, 2004), and this is a novel and veryinteresting aspect of tumor biology (Stefani, 2004).

A survey of the expression of SEL1L mRNA aswell as its encoded protein, on a series of cancerousand pre-neoplastic lesions, revealed, at least partially,the role of SEL1L in cancer progression (Granelli et al.,2004; Barberis et al., 2005; Ferrero et al., 2005,submitted); furthermore, its expression in breast cancercorrelated with patientssurvival (Orlandi et al., 2002a).Takinga step over, in vitro studies indicatedthat SEL1Lprotein affects those pathways which regulate signaling(cellcell and/or cellmatrix) interactions. Presently,there is not much literature on this gene, which resultsin the difficulty in generating a precise theory onthe function of the protein. However, considering theavailable data derived from several organisms, anhypothesis can be forwarded in that SEL1L mayfunction in the protein degradation processes throughubiquitin-proteosome system and perhaps in regulatingimportant pathways, such as Notch and TGF-beta.

Since the protein structure consists of at least threeblocks of highly conserved domains (SEL-1-like repeatsand Hrd3 motif) and of two modules acquired laterduring taxonomic evolution (fibronectin type II domainand proline-rich region), the complexity of its specificityis highlighted.

SEL1L is not a member of a vast family of proteins butthe several described isoforms (over four) give theappearance of belonging to a multifamily of moleculeshaving perhaps redundant functions.

The architectural complexity of the SEL1L gene andprotein is an example of its flexibility and reflects how insome genes, different combinations of exons can becomeactive at different times and each combination can yielda different protein with a different sub-cellular location,

this phenomenon allows the control of gene expressionin an exquisite manner.

SEL1L GENE AND PROTEIN STRUCTUREGenomic structure

The SEL1L gene consists of 21 exons spanning62.24 Kb of genomic DNA (Fig. 1A), the coding regionconsists of 2,385 Kb (GenBank accession no.NM_005065) encoding for a protein of 794 amino acids

with a calculated molecular weight of 88,750 KDa.Upstream of the starting methionine, a genomicsequence of 1,271 Kb (GenBank accession no. AF157516), contains the essential features of thepromoter region. The C-terminal tail consists ofover 5.0 Kb untranslated sequences likely containingfundamental regulatory elements. Two polymorphic(CA)n repeats are located in intron 2 and intron 20,useful in linkage studies (Biunno et al., 2000).

Chromosomal location

SEL1L resides on chromosome 14q24.3-q31 from81069886 to 81007646 (NCBI build 35.1, August 2005),on the reverse strand (Fig. 1A), in the same region

as IDDM11, a locus showing evidence for linkage to type1 diabetes (Pociot et al., 2001). However, no associationwas found between SEL1L microsatellite markers andtype 1 diabetes nor with the Graves disease alsoresiding in the same chromosomal band (Ban et al.,2001; Pociot et al., 2001). Extensive computer aidedanalysis of this chromosomal interval indicates that the7.3 Mb of DNA stretch harboring SEL1L, is ratherscarce in disease causing genes. Indeed, only fiveproteincoding genes (neurexin 3, deiodinaseiodothyronine typeII, thyroid stimulating hormone receptor, generaltranscription factor IIA, and fibronectin leucine-richtransmembrane protein 2), four hypothetical ESTand ten pseudogenes surround the SEL1L gene(Fig. 1B). The chromosomal interval between SEL1L

and fibronectin leucine-rich transmembrane protein2 genes belongs to the so-called Gene Desert area or

TABLE 1. SEL1L gene card

Human SEL1L

Official symbol SEL1LGene description Sel-1 suppressor of lin-12 (C. Elegans)-likeGene alias IBD2, SEL1-LIKEGeneID 6400SwissProt Q9UBV2 Sel-1 homolog precursor

Chromosomal location 14q24.3-31

Transcripts Three alternative splicing forms experimentally confirmedIsoform coding the complete SEL1L gene is isoform A

Ref. Seq. mRNA NM_005065Protein 794 amino acids; 88755 Da

Ref. seq. protein NP_005056Domains and motifs Fibronectin II domain

Sel1-like repeats (three clusters)Hrd3-like motifProline-rich motif

Subcellular localization Cytoplasmic and nuclearProtein expression

Fetal tissues Ubiquitously expressedNormal tissues Highly expressed in pancreatic cells and other protein secreting cells,

including B cellsTumor tissues Differentially expressed in pancreatic, gastric, breast, and lung cancer

Function . May have a role in the ER-associated protein degradation (ERAD)system (similarity with Hrd3)

. Negative regulator of the Notch pathway in C. elegans

. May play a role in TGF-beta signaling. In breast and pancreatic tumor decrease tumor growth andaggressiveness, possibly involving cellmatrix interactions

Regulation . Upmodulated in ER-stress. Upmodulated in early stages of cancer progression. Downmodulated in comparison to normal tissues in the poor

prognosis breast tumors

24 BIUNNO ET AL.

Journal of Cellular Physiology DOI 10.1002/jcp


3/17

Fig. 1. A: SEL1L genomic structure: SEL1L gene resides on chro-mosome 14q24.3-q31 from 81069886 to 81007646 on the reversestrand (NCBI build 35.1, August 2005) spanning 62,240 Kb of genomicDNA. The coding region consists of 2,385 Kb and contains 21 exons(NM_005065). B: Simplistic representation of chromosome 14 (mod-ified from Genomic Cartography, http://acg.media.mit.edu/people/fry/genocarto.html). This figure engenders a very compact, insightfulvisual for chromosome 14, the DNA stretch harboring SEL1L isenlarged in the botton. The empty gray (yellow) wireframe boxessignify a gap between genes, an area of unused data. The dark floor(blue) areas represent parts where genes exist. The dark floor (blue)wireframe boxes are proportional in size to the start and end point ofthe gene within the sequence of letters, usually a few thousand lettersapiece. The solid (blue)boxes within these framesrepresent the codingportions of a gene, shown in proportion to the start and end of the geneitself. By using cubes to represent the data, it allows the map to be re-weighted, such that the much smaller percentage of coding informa-tion still remains relevant in scale. Similar to a logarithmic scaling, itscales using a cube root in the linear direction, resulting in a morebalanced visualization. NRXN3, neurexin 3; DIO2, deiodinaseiodothyronine type II; TSHR, thyroid stimulating hormone receptor;GTF2A1, general transcription factor IIA; FLRT2, fibronectin leucine-rich transmembrane protein 2; GALC, galactosylceramide. C: SEL1Lpromoter: nucleotide positions are numbered to the left with respect to

the translation initiation site (ATG is in bold and red characters).Potential cis-acting elements are predicted with MatInspectorprogram (Jan, 2005), blue and yellow boxes correspond to regulatorsinvolved in ER stress and tissue-specific factors, respectively. TheCAAT and the four GC boxes are boxed in celestial and pink,respectively. The multiple transcription start sites are blue under-lined. The lowercase letters indicate part of the first intron sequence.The two polymorphisms (366 and 354) are in bold and blackunderlined. The CpG island located between nucleotide 500 and thestart codon is highlight in yellow. D: SEL1L protein structure: SEL1Lis a multimodular protein containing several structural and functionaldomains as well as signal sequences. The signal peptide (from 1 to 22amino acid residues) and the Pest sequence (from 80 to 102 amino acidresidues) are represented by red and pink rectangles. The fibronectintype II domain (from 120to 168residues) is symbolized by the hexagon(FN2), the SEL-1-like repeats are represented by rhombi and aredistributed in tandem along the central portion of the protein in threelarge clusters (I cluster: 183326, II cluster: 373554, and III cluster:664675 residues). The Hrd3 like motif is located within the last SEL-1-like repeat (664675 residues) and is represented by a circle. Thetransmembrane region (TM) (739761 residues) and the proline-richtail (770793 residues) are symbolized by a blue and pink rectangle.The N-linked glycosilation is also underlined.

SEL1L AND ITS STRUCTURAL AND FUNCTIONAL COMPLEXITY 25



4/17

Genome Deserts (http.www.chr7.org/March2003/GeneDeserts.php: table S7d) since consists in no genesin>500 Kb genomic stretch. The functional importanceof these areas is undetermined, it hasbeen reported thatmegabace deletions of gene deserts result in viablemice, supporting the hypothesis of the existence ofdisposable DNA in mammaliangenomes (Nobrega et al.,2004). A genome-wide analysis of those chromosomalfeatures able to repress the transcription of the human

immunodeficiency virus identified the integration of theviral particle in gene deserts areas. This suggeststhe presence of a negative transcription regulatoryelements able to maintain the integrated virus in alatency state (Lewinski et al., 2005). Residing in thisdesert area, SEL1L may protect itself from the commonrecombination events that occur in the immunoglobinheavy (IGH) chain locus, a gene cluster of immunologi-cal importance located in the telomere of chromosome14 and frequently involved in chromosomal transloca-tions in lymphocyte malignancies. This isolated locationmay ensure SEL1L integrity.

SEL1L does not appear to bea memberof a vastfamilyof proteins; a SEL1L-like protein has been annotated byNCBI; the gene encoding for this protein is located on

chromosome 20p12.1 and contains structural featuressimilar to SEL1L (Cattaneo et al., 2004).

Promoter characteristics

SEL1L contains a TATA-less promoter with multipletranscription initiation start sites (Fig. 1C). The basalcore of the promoter consists of four SP1 binding sitesand a CAAT box, the sequence of 900 bp extendingupstream, harbors several putative binding sites forknown transcription factors including tissue-specificregulators such as hepatocyte nuclear factor-3 beta

(HNF3b

), homeobox Nkx2-5, hepatic nuclear factor1 (HNF1), GATA binding factor 1, prostate-specifichomeodomain protein Nkx3.1, Pdx1 (IDX1/IPF1) pan-creatic and intestinal homeodomain TF, myelin tran-scription factor 1 like, Myt 1 zinc finger involved inprimary neurogenesis as well as factors involved in theER stress such as CHOP and C/EBP alpha hetero-dimers, and the hypoxia inducible factor bHLH/PAS(HIFF/HIF1).

Computer-assisted analysis of the entire promoterregion predicted a CpG island located between nucleo-tide 500 and the start codon, suggesting that SEL1Lmay be transcriptionally regulated through changes inthe methylation status.

The SEL1L promoter results highly active in the

pancreatic beta and in the embryonic kidney cells(Cattaneo et al., 2001b). The tissue-specific regulation

Fig. 1. (Continued)

26 BIUNNO ET AL.



5/17

of SEL1L promoter is also underlined by Odomand his collaborators who demonstrated, by chromatinimmunoprecipitation (ChIP) assay and genomic micro-array hybridization, that SEL1L promoter interactswith two transcriptional regulators, the HNF1a HNF4ain human pancreatic islet cells as well as in hepatocytes(Odom et al., 2004). Both regulators are required forthe normal function of liver and pancreatic islets.

Mutations in these factors have been associatedwith the type 3 form of maturity-onset diabetes ofthe young (MODY3 and MODY1). The characteristics ofthe SEL1L promoter make it a useful tool to efficientlydrive transcription of a therapeutic gene; indeedtwo laboratories reported on the use of the SEL1Lminimal promoter to drive the expression of the cytosinedeaminase suicide gene for gastric cancer genetherapy (Mathlouthi et al., 2003; Aberle et al., 2004). Although these findings are very interesting andpromising, they require further in vitro studies inaddition to tumor transplants prior to any clinicalapplication.

Protein structure

SEL1L is a multimodular protein consisting of severaldomains and signal sequences that confer the multi-faceted specificities to the molecule.

The main domains are schematized in Figure 1D.

(i) The fibronectin type II domain is a small compacttwo-disulphide-bond module of about40 aminoacidresidues located in the amino-terminal region (from120to 168residues). It wasinitially identifiedin thecollagen-binding region of the matrix proteinfibronectin, and in some seminal fluid proteinssuch as bovine seminal plasma proteins PDC-109and BSP-A3. Related domains are found in theextracytoplasmic parts of membrane-associatedproteins, such as members of the mannose recep-

tor-phospholipase A2 receptor family, mannose-6-phosphate receptors. FNII modules are also pre-sent in matrix metalloproteinases MMP-2 andMMP-9, as well as serine protease factor XIIand hepatocyte growth factor activator. There isevidence that high affinity interaction with col-lagen is stabilized by the presence of multiple FNIIdomains in fibronectin and MMP2 proteins (Pick-ford et al., 2001). Despite their widespread occur-rence in diverse vertebrate proteins, FNII modulesare absent in invertebrates, including C. elegansand D. melanogaster. This domain is missing fromthe SEL1L invertebrate orthologs arising only inthe chordate lineage, representing an interestingaspect of the evolutionary history of SEL1L andindicating a specific lineage function. This suggeststhat in addition to the acquisition of an ulteriorfunction in higher organisms, SEL1L can exert itsfunction independently of the fibronectin domain,and its presence may redirect or increase itsspecificity.

It is worth mentioning that among the chordates,only the mouse and rat SEL1L proteins show avariant form consisting in the absence of theFNII domain, likely due to alternative splicing(NP_035474 and NP_808794, respectively).

(ii) The SEL-1-like repeats represent a subtype ofthe tetratricopeptide repeat (TPR) sequence dis-tributed in tandem along the central portion of the

protein and assembled in three large clusters(see figure) (from 183 to 326, from 373 to 554,

from 627 to 699 residues). The TPR repeats areproteinprotein interaction modules found in mul-tiple copies in a wide variety of proteins havingdifferent biological functions, the majority of themare involved in cell cycle control, transcription andsplicing events, protein transport and import,regulatory phosphate turnover, and protein folding(Blatch and Lassle, 1999). They are often found in

protein complexes and may form amphipathic a-helices that mediate protein protein interactionsas well as the assemblage of multiprotein com-plexes. The repeat consists of a degenerate34 amino-acid motif that forms two amphipathica-helice subdomains (Das et al., 1998). SEL-1-likerepeats differ from standard TPRs in possessing avariable number of amino acids between the two a-helices (Ponting, 2000). They are highly conservedduring SEL1L taxonomic evolution reinforcingtheir functional and structural importance. TheSMARTs nrdb database annotates 2258 SEL-1 likedomains in 500 proteins from bacteria to eukar- yotes. Seventeen of the annotated proteins arehuman SEL-1 like multi repeat proteins; the

majority of them do not have as yet a precisefunction, with the exception of the Tg737gene andthe Megalin binding protein that harbor three andfive repeats, respectively. The Tg737 is a candidatepolycystic kidney disease gene with tumor suppres-sion activity in liver cancer (Isfort et al., 1997;Richards et al., 1997). The Megalin is an adapterprotein with receptor signaling complex scaffoldactivity involved in the signal transduction cascadeand cell communication. This protein probably,through the SEL-1-like repeats, binds the receptortail, the transcriptional regulators, and compo-nents of signal transduction cascades (Petersenet al., 2003).

In the Gram-negative human pathogen Helico-bacter pylori, the secreted cysteine-rich proteins(HcpC) belong to the family of Sel1-like proteins(Luthy et al., 2004). Hcp structure has beenproposed as a template for modeling SEL1-likeproteins because the conserved repeat geometry ishighlighted by the consensus sequences. If thestructural homology with the SEL1L protein needto be proved, itis interesting to note that the overallfold of Hcp proteins support the hypothesis ofpossible multiple protein protein interactionamong the TPR proteins and confirmed the peptidebinding site seen in eukaryotic TPR protein, suchas Hsp70/Hsp90 and PEX5 (Richter and Buchner,2001). In this regard, Ponting also predicted

that the sel-1 repeats in Hrd3p ofS. cerevisae

maypossess a peptide-binding function similar to that ofTPRs in the Hop protein, a co-chaperone bindingheat-shock cognate protein 70 (Hsc70) (Ponting,2000). In yeast, the Ydl203c protein, a substratefor the E3 ubiquitin protein ligase Rsp5, containsthe PY motif, a potential target for the E3ligase and seven terminal SEL1-like repeats (Kuset al., 2005).

(iii) The Hrd3-like motif was originally identified in theyeast Hrd3p protein, a component of ER membraneassociated ubiquitin ligase complex (Hamptonet al., 1996). It is the most highly conserved motifconsisting of 12 aa residues (from 664 to 675residues) being shared by several divergent species

including mammals, insects, plants, and yeast. Itmaps within a SEL1L region that plays a crucial




6/17

role in the transmission of anti-proliferative signals(Cattaneo et al., 2004).

(iv) The proline-rich motif is an adopter systemthat brings together proteins and mediates pro-teinprotein interactions (Kay et al., 2000),it is located in the COOH-terminal end of theSEL1L protein (from 770 to 793 residues). Thesesequences are found in situations requiring the

rapid recruitment or interchange of several pro-teins, such as during initiation of transcription,signaling cascades, and cytoskeletal rearrange-ments. Thus, its role is not to provide a structurallydefined complex but rather to bring proteinstogether in such a waythat subsequent interactionsare more probable. The best-known examples ofmodules that bind proline-rich region are repre-sented by SH3, WW domains, and several newprotein interactions. The SEL1L proline stretchmay be better recognized by SH3 domains whoseoptimal ligand preference varies around the PxxPcore, where x denotes any amino acids. Thisterminal tail is an evolutionary acquisition of theSEL1L protein since it is present only in thechordates with the except of chicken and zebrafish.

SEL1L protein harbors several additional signalsequences:

(i) A cleavable signal peptide consisting in 20 hydro-phobic amino acids present at the Nterminus andfollowed by a cleavage site indicating that signalpeptide may not be retained.

(ii) An XXRR-like ER membrane retention signal(RVRI) located in the first five aa residues.

(iii) A PEST sequence present (from aa 80 to 102residues) known to target proteins for rapiddegradation.

(iv) A consensus sequences for nuclear localization(from aa 107 to 113 residues) and export signals(at position 9 18 and 747 755). These nuclearsignals mediate the binding to karyophein/importinand karyophein/exportinproteins allowing for largeprotein to shuffle in and out of the nucleus, and areconsidered extremely important regulators of theproteins sub-cellular location having a strongimpact on important processes related to cancer,cell cycle, cell differentiation, and other fundamen-tal aspects of cell viability (Kressel and Schmucker,2002; Qu and Koromilas, 2004).

Using a novel strategy for quantitative glycoproteinprofiling, Zhang and his collaborators, reported that

SEL1L is a N-linked glycoprotein since SEL1L-N-glycosylated peptide was identified in the crude mem-brane fraction from the LNCaP prostate cancer epithe-lial cell line (Zhanget al., 2003).The SEL1L glycosylatedpeptide contains the motif NSSQ at position 272, threeadditional glycosylated motifs are predicted at the N-terminus: (NGSN), (NHTK), and (NETY). N-linkedglycosylation is prevalent in proteins destined forextracellular environments, including extracellularside of the plasma membrane, secreted proteins, andproteins contained in body fluids, such as blood serum,cerebrospinal fluid, urine, breast milk, saliva, lunglavage, fluid, or pancreatic juice, and changes in theextent of glycosylation and the carbohydrate structureof these proteins have been shown to correlate with

cancer and other diseases, highlighting the clinicalimportance of this modification as an indicator or

effector of pathologic mechanisms. Although there isno evidence indicating extracellular location of SEL1L,computational, and experimental data documentedthe existence of two protein isoforms, one of eachis predicted to be secreted, the other is indeed foundto be secreted by Clark et al. (2003); see Section2.1), highlighting their relevant biological and clinicalimplications.

SEL1L cross-species conservation

Comparative sequence analysis across differentregna, including metazoa, fungi, viridiplantae, andbacteria, revealed the remarkable conservation ofits primary sequence, although the gene structuralcomplexity increased in evolution, suggesting differen-tiation of its function (Table 2) (Biunno et al., 2002).However, several components remain perfectly con-served suggesting that the SEL1L protein exerts a veryimportant biological function and may belong to thatclass of proteins considered to be essential. Theincreasing availability of genome sequence data from avariety of model organisms as well as highervertebrates

strengthens thestrong constrainto keep SEL1L. Amongmammals, SEL1L shares strict amino acid identity withchimpanzee (99%), dog (97%), hamster (92%), mouse(93%), andrat (92%).It also shows a good similarity withthe model organisms such as xenopus (82%), chicken(83%), zebrafish (73%),Drosophila melanogaster (51%),and C. elegans(46%) (Table2).Arabidopsisthaliana andSaccharomyces cerevisiae display lower similarity (34%and28%, respectively) with thehigh amino acid identityin the C-terminal region containing and flanking theHrd3 motif.

SEL1L ALTERNATIVE ISOFORMS ANDEXPRESSION PATTERNS

SEL1L alternative splicing eventsBeside the 7.5 kb SEL1L mRNA species, several

smaller alternated transcripts have been reported onlyin pancreatic cells (Donoviel et al., 1998; Harada et al.,1999). The Aceview program (August 2005) (http://www.ncbi.nlm.nih.gov/AceView) supports the presenceof splice variants and predicts at least four differenttranscripts putatively encoding for four protein isoforms(Fig. 2A,B) and five shed variants with exon contact, butnot shared intron boundary; they are ignored in theassessment of alternative features of the gene sincepartial or unspliced variants. Thetwo isoforms (B andC:genebank accession no BM312853 and AY358651,respectively) are a truncated version of SEL1L, due to

splicing events that occur within the ninth and eighthexons, respectively, differing dramatically in the extentof their 30 terminal ends and in the transcriptioninitiation start sites. These were experimentallydetected using specific primers by RT-PCR analysis(Cattaneo et al., personal observations). Both isoformsloose close to half of theSEL1L protein at theCOOH-endcontaining the middle and the final clusters of SEL-1like repeats, the Hrd3 like motif, the transmembraneregion, and the proline tail. Although Aceview programpredicts a deletion in the first 64 nucleotides of thefourth exon splicing out the beginning of the type IIfibronectin domain, sequence analysis performed ontranscript B extracted from the pancreatic neoplasticcell line Suit-2 does not confirm this deletion (Cattaneo

et al., personal observations), suggesting the existenceof other tissue-specific variants. The splicing event

28 BIUNNO ET AL.



7/17

occurring in the eight exon of isoform C generates aperoxisomal domain (SRL) at position 299301.

Isoform D (genebank accession no. BE000519) isreported to be truncated at the 3 0 end, posses threeexons (18, 19, and 20) encoding for a 72 amino acidprotein containing the functional Hrd3-like domain andan ER membrane signal.

Supports on the presence of several SEL1L variantsarise also from recent reports and additional computer-assisted programs.

Clark and collaborators through the secreted proteindiscovery initiative strategy (SPDI), a large-scale

method that utilizes both biological and computationalmethods, identified the SEL1L C variant (Genebank AY358651) having the Type II fibronectin domain, atransmembrane region, and secreted subcellular loca-tion (Clark et al., 2003). This variant is also predicted bySwiss-prot program (Q9UBV2-2) (Fig. 2A).

Recently,a newisoform derived fromhighthroughputcloning of full length human cDNAs from placentalibraries optimized for large and rare transcripts, isannotated in the Genbank databases (DR005068.1).This isoform shares with SEL1L the common exons 1 6(Fig. 2A).

Overall all these data provide evidences that SEL1Lgene does not encode for one unique expressed proteinbut rather for several proteins, increasing the complex-

ity of its biology. Actually, the functional role of all theseisoforms is yet unknown, but it can be hypothesized

that they may contribute to creation of the diversityaround one single gene in order to allow tissue andtemporal-dependent combinatorial patterns of proteinexpression. The role of isoforms B and C may berestricted to the proteinprotein interaction and/or binding since they retain only the first cluster ofSEL-1-like repeats as well as the fibronectin type IIdomain; moreover the predicted extracellular location ofthese two isoforms highlights their clinical importanceas useful biomarkers of body fluids, such as blood serumand urine. The incomplete isoform D may represent theonly variant retaining the functional Hrd3 motif.

Multimodal approaches are essential for the fullcharacterization of these multiple isoforms in relationto the cell physiology and SEL1L, since their functionmay be restricted to a specific tissue, sub-cellularcompartments, or temporal conditions performingrelated roles in various cellular contexts, such asstimulations, phase of cell cycle, levels of interactingproteins.

SEL1L expression in healthy adult human tissues

Previously works generated an expression archive ofSEL1L which determined the ubiquitous expression ofthe gene only in fetal and neoplastic tissues (Cattaneoet al., 2000). Northern blot analysis revealed that the7.5 kb SEL1L transcript is almost always under-

represented in normal and healthy adult cells with theexception of the pancreas (Biunno et al., 1997). SAGE

TABLE 2. SEL1L cross-species conservation

Phylum Speciesa Identities aa Human and subject region Protein length (aa) Acc. number

Chordata SEL1L 100% 1794 794 NP_005065Homo sapiens 794/794 1794SEL1L 99% 17794 817 XP_510102

Pan troglodytes 772/778 40817SEL1L 97% 1769 794 XP_537530Canis familiaris 753/769 1751

SEL1L 92% 1794 794 BAB12403Mesocricetus auratus 736/794 1794SEL1L 93% 19794 790 BAB23750

Mus musculus 726/776 20790SEL1L 92% 19794 794 Q80Z70

Rat rattus 721/776 19794SEL1L 83% 24780 798 XP_421303Gallus gallus 642/766 25783SEL1L 73% 13767 776 XP_697044

Danio rerio 558/755 16750Unknown protein 82% 76767 822 AAH95916

Xenopus laevis 575/701 90787 Arthropoda ENSANGP 57% 182761 596 XP_310466

Anopheles gambiae 333/581 4579ENSANG 58% 166761 659 XP_392802

Apis mellifera 354/601 52647CG10221 51% 169763 819 NP_651179

Drosophila melanogaster 311/606 125721

GA10167 51% 182793 771 EAL27774Drosophila pseudoobscura 315/614 137739Nematode Hyp. protein CBG23527 47% 182768 685 CAE75512

Caenorhabditis briggsae 285/604 92685SEL1L 46% 182768 685 NP_506144Caenorhabditis elegans 281/604 92685

Ascomycota Hyp. Protein FG00909.1 34% 206705 821 XP_381085Gibberella zeae 192/563 126681Hyp. Protein 32% 206712 1307 XP_331768

Neurospora crassa 185/569 142705Hyp. Protein SPAC1B3 24% 202692 700 NP_594794Schizosaccharomyces pombe 140/568 86629Hrd3p 28% 413726 833 NP_013308Saccharomyces cerevisiae 97/335 373690

Streptophyta SEL1L 35% 236788 670 BAE02648Glycine max 206/575 102666SEL1L 34% 212712 678 NP_564049

Arabidopsis thaliana 178/517 83588

aThe species with relevant biological importance and sharing the high identity with the human SEL1L protein are elencated in the table.




8/17

analysis using a complete series of tissue sections fromnormal and neoplastic cells (www.ncbi.nlm.nih.gov/projects/SAGE/index.cgi?cmdaccsearch tags: TAA-GTTGAGT and TAAGTTGAGTGGAATGT), and theexpression profile documented in the mouse tissues(Donovielet al., 1998) confirm our previous observationswhich strongly suggest that the function of the encodedprotein needs to be found within the pancreas. Atprotein level, SEL1L immunohistochemical expressionis restricted to protein secreting cells with a typicallydiffuse granular cytoplasmic positivity. These included

normal breast, pancreas, prostate, salivary glands, andserous/mucinous glands associated with the gastroin-testinal and ovary. In the pancreas, the protein isabundantly present in the acini and in the islet ofLangherans, but completely absent in ductal cells(Cattaneo et al., 2003). Among endocrine organs,thyroid, adrenal, and insulae of Langherans are posi-tive. Plasma cells andmegacaryocytes also show intensestaining. Representative examples of immunostainingof SEL1L are shown in Figure 3.

Immunohistochemical analysis on human tissuesreported in this review has been performed using themurine monoclonal antibody MSel1 that hasbeen raisedagainst the N-terminus of human recombinant SEL1Lprotein (Orlandi et al., 2002b). The growth arresting

function exerted by the C-terminal region of SEL1L onbacterial cells prevented the possibility to produce the

entire recombinant SEL1L protein and select mono-clonal antibodies directed against the Hrd3 motifcontaining C-terminus domains of SEL1L protein.MSel1 then recognizes all the possible SEL1L isoformslisted in Figure 2A and the studies performed with thisantibody could not address the question of differenttissue expression or cellular sub-localization of theSEL1L isoforms. The MSel1 expression pattern in fetaland adult tissues is concordant with the expressionpattern reported in C. elegans (Grant and Greenwald,1997) and mouse (Donoviel et al., 1998) using anti sel1-l

conventional polyclonal antisera.

SEL1L expression in human fetal tissues

SEL1L immunoreactivity varies in different fetaltissues and at different gestational ages. Fetal tissues,from 6 to 24 weeks (w) of gestational age, obtainedfrom 12 therapeutic or spontaneous abortions, weresubmitted for routine histological analysis.In summary,SEL1L shows a nucleo-cytoplasmic pattern of reactivityin the early embryo and in most immature tissues;at later ages of development the pattern of reactivityis restricted and present also in glandular structureswith a pattern almost exclusively cytoplasmic. Repre-

sentative examples of immunostaining of SEL1L areshown in Figure 3.

Fig. 2. A: A graphical representation of SEL1L isorforms. The black numbered rectangles correspond tothe exons, while the white rectangles correspond to the intronic sequence which is retained in the

alternative isoforms. The SEL1L domains are indicated on the top of the isoforms.B

: A schematicsummary of SEL1L isoforms features.

30 BIUNNO ET AL.



9/17

Early embryos (6 7 w), show SEL1L immunostainingin the majority of cells in all fetal layers, with a moreintense reactivity in the neuroepithelial structures; thepattern of staining is both nuclear and cytoplasmic. At8 12 weeks, SEL1L expression is still widely present inseveral tissues. In gastrointestinal tract, strong nuclearand cytoplasmic staining is present in the mucosa; non-epithelial cells of the muscular layer are also labeled,although with weaker intensity. Ductal pancreaticstructures are identified in a 10 weeks fetus and show aweak SEL1L staining with scattered cells more inten-sely stained. Hepatocytes are intensely stained butreactivity is only cytoplasmic; hematopoietic cells in theliver are negative with the exception of megakariocytes.

In the respiratory tract, intense staining is observed inbronchial epithelium showing a nucleo-cytoplasmic

pattern of reaction. In the urogenital tract, epithelialstructures are weakly stained with glomerular epithe-lial cells showing strong perinuclear reactivity. Gonadsare diffusely stained but germinal cells are negative.Neural structures display SEL1L intense cytoplasmicstaining witha nucleo-cytoplasmic pattern of reaction inmore immature cells. Epithelial cells of the choroidplexus are positive. Most mesenchymal cells are immu-nostained. Nucleo-cytoplasmic reactivity is observed incartilage, skeletal, smooth muscle, and endothelial cells;myocardial cells are only weakly positive. In moremature fetal tissues (1324 weeks), strong SEL1Limmunoreactivity is confined to selected tissues includ-ing gastrointestinal mucosa and pancreas, while liver

cell positivity is very weak. Immunostaining is alsolocalized in newly formed structures, such as skin

Fig. 3. Representative examples of SEL1L immunostaining in adult and fetal tissues.




10/17

glands, salivary and mucous mucosal glands, thyroid,endometrium, thymic squamous epithelium, renaltubular epithelium, some lymphoid element, butusuallyonly with a cytoplasmic pattern of reaction. In thepancreas acini and Langherans cells are intenselystained whereas ductal cells are only occasionallypositive.

In the placenta, intense staining is observed in the

syncitiotrophoblast, whereas cytotrophoblast is weaklystained or negative; decidual cells and endometrialglands are positive.

In agreement, RNA in situ analysis of developingmouse embryos indicated that mSEL-1L was moder-ately expressed throughout the neural tube and dorsalroot ganglia; particular high levels were observed in thefloor plate of the neural tube beginning at E10.5 andincreased at E11.5. High levels of expression wereobserved at E14.5 and E17.5 in the acini of the pancreasand moderate in the epithelial cells of the gut villi(Donoviel et al., 1998).

In summary, both in fetal and adult tissues,SEL1L immunohistochemical expression is restrictedto protein secreting cells indicating that SEL1L

may play a crucial role in secretory/trafficking/qualitycontrol checkpoint processes occurring in proliferativecells.

SEL1L POSSIBLE FUNCTION(S)

Although the real function of the SEL1L protein inmammalian cells is still unknown, much can be learnedand deduced from the available nucleotide and proteinsequence, as well as from functional data obtained inhomologous proteins, such as sel-1 in C. elegans andHrd3p in S. cerevisiae.

SEL1L and the endoplasmic reticulum (ER)

Human SEL1L was indicated to be the humancandidate homolog of HRD3 (Kaneko and Nomura,2003) suggesting that it may have a role in the ER-associated protein degradation (ERAD) system.

ER facilitates the synthesis and processing ofalmost all the secretory and membrane proteins,and is susceptible to various forms of stress thatprovoke the accumulation of unfolded or misfoldedproteins (Kaufman, 2002). ER stress can be induced bya variety of pathophysiological conditions includingexpression of mutant proteins, inhibition of asparagine(N-linked glycosylation), overload of viral proteinsduring virus replication (Kaufman, 2002).

Alterations in the homeostasis of the ER by variousforms of stress lead to the accumulation of unfoldedproteins and protein aggregates that are detrimental to

cell survival. Eukaryotic cells can adopt to ER stress byactivating specific signaling pathways and mechanisms(upregulation of unfolded protein response [UPR] genesespecially those involved in the ERAD processes), whoseprimary purpose is to limit theaccumulationof unfoldedproteins in the ER. UPR entails the coordinatedtranscriptional induction of genes encoding for ERresident proteins (glucose regulated proteins Grps),ER chaperones (Kaufman, 2002), and Hrd (Hrd1p and3p) proteins (Kikkert et al., 2004). In order to decreasethe protein overload in ER, another pathway developedby the cells is to inhibit protein synthesis (Ron, 2002).The final pathway undertaken is to eliminate theunfolded or misfolded proteins by proteasome-depen-dent proteolysis (Kostova and Wolf, 2003). If cell

adaptation is not sufficient, then the stressed cells areeliminated by apoptosis by activating the JNK pathway

and caspases 7, 12, or 3 (Kaufman, 2002). The UPRand ERAD genes have well been defined in yeast and inC.elegans, these under normalgrowthconditions arenotessential for cell viability but are of paramountimportance during ER stress or when UPR is blocked.One main function of Hrd3p in the ERAD system isto affect the stabilization of Hrd1p, preventing thecytosolic RING-H2 domain from programming Hrd1p

degradation possibly through auto-ubiquitination. Stu-dies with truncated alleles and overexpression ofHrd3p indicated that it is also required to modulateHrd1p RING-H2 ubiquitin ligase activity by physicallyinteracting with the Hrd1p transmembrane domain.In the absence of subsrate, Hrd3p stabilizes Hrd1pby binding to its transmembrane domain thus suppres-sing the cytosolic RING-H2 domain ubiquitin ligaseactivity. When the substrate is present, Hrd3psenses the requirement for Hrd1p transmembranedomain to activate Hrd1p function in a correct temporaland spatial manner. It remains however to be estab-lished if the human SEL1L interacts directly withHRD1 to confer stability to the complex. Kikkertshowed that both endogenous and transfected human

HRD1 are relatively stable hence it is unlikely thatSEL1L ensures the stability of HRD1 in a mannersimilar to that observed in yeast (Kikkert et al., 2004).Soon, at least a partial map of all human proteinprotein interactions, interactome, will be availablemaking possible the understanding of how parts worktogether.

Actually, studies performed on several organisms,documented the SEL1L involvement in the UPR/ERADpathway.

(i) In C. elegans, sel-1 mediated interference (RNAi)resulted in marked upregulation of the ER stressindicator hsp-4::gfp (BIP) in a xbp-1 dependentmanner. Inactivation of sel-1 had no impact on the

viability of wild-type animals and only modestlyreduced the viability of xbp-1 mutants, but whencombined with theinactivation of abu-1 it increasedthe lethality of xbp-1 mutants (Urano et al., 2002).

(ii) In pancreatic islet cells from Akita mouse, sub-jected to ER-stress by the production of misfoldedinsulin, was found an upregulation of the residentmolecular chaperone Bip as well as Hrd1 and Sel1lproteins (Allen et al., 2004).

(iii) InArabidopsis thaliana, by functional DNA micro-array and real time PCR, it was observed anincreased expression of the At-SEL1L trancript intunicamycin-treated plants (Kamauchi et al.,2005).

(iv) Studies on human embryonic kidney 293 cells,demonstrated that treating the cells with ER-stressinducing agents such as thapsigargin or tunicamy-cin, resulted in an increase of HRD1 and SEL1LmRNA levels, and both upregulation contribute toprotect the cells from ER-stress induced death bydegradating unfolded proteins accumulated in theER (Kaneko and Nomura, 2003).

(v) In humans, a possible role of SEL1L in ER-stressprocesses, may be found in the proteomic andmicroarray results obtained while studying thebreast cancer cell line MCF-7 containing the entireSEL1L gene stably transfected and deliberatelyinduced to be expressed. The ectopic expression ofSEL1L revealed changes in the level of proteins and

transcripts involved in protein folding, and cellstress response such as protein disulfide isomerase

32 BIUNNO ET AL.



11/17

A3 (a major protein in the ER lumen, multifunc-tional folding catalyst, and molecular chaperone),proteasome b6 subunit, heat shock protein60, and peroxiredoxin 1 (antioxodant enzymethat regulates the cellular redox state). Indeedseven transcriptof proteasomasubunits were foundupmodulated in SEL1L overexpressing cells (Bian-chi et al., 2005).

(vi) In a genome-wide analysis study of UPR response,in fibroblasts from congenital disorders of glycosy-lation type-1 patients, it was recently documentedthat SEL1L expression is markedly increased alongthe PERK inhibitor DNAJC3/P58 (IPK) genes(Lecca et al., 2005).

SEL1L AND CANCER

Several studies have focused on the role of SEL1L inmany aspects of malignant transformation and tumori-genic processes,providing significant in vitro andin vivoevidences to link its increased expression to a decreasein tumor aggressiveness.

Chromosome 14q24.3-31 deletionsand amplifications

Changes in DNA copy number contributes to cancerpathogenesis, but alterations of the chromosomal region14q24.3-q31 in cancer is not a very common event.Recent studies reported that loss of heterozygosity(LOH) on chromosome 14q is common in astrocytomas,oligodendrogliomas, glioblastomas, and meningiomas(Tse et al., 1997; Hu et al., 2002; Dichamp et al., 2004).Two distinct minimal regions 14q22.3-14q24.3 (betweenmarkers D14S276 and D14S74) and 14q31.3-14q32.1(between markers D14S74 and D14S280) have beenreported to be deleted in two subgroups of gliomas(astrocytomas, oligodendrogliomas, oligoastrocytomas)and glioblastomas. The last region spans 13.5 Mb (from

77,728,449 to 91,272,543 ) contains, among 23 knowngenes, SEL1L gene, 14 pseudogenes and17 hypotheticalgenes, and open reading frames. To our knowledge, notumor suppressor genes associated to this chromosomalregion have as yet been identified in these neoplasms.Studies are in progress to determine whether SEL1L islost in oligodendroglial astrocytic tumors (Dichampet al., 2004). Allelic imbalance and loss of chromosome14q were also found in head and neck squamous cellcarcinomas (HNSCCs), in primary colorectal carcino-mas with metastatic ability as well as metastases, in therenal cell carcinomas associated with neuroblastoma, inthe Korean ovarian carcinomas, and in the oralsquamous cell carcinomas (Lee et al., 1997; Medeiroset al., 1999; Thorstensen et al., 2001; Bruder and Moch,2004).

Mutation analysis and promoter polymorphisms

Mutations in the SEL1L gene were searched inhuman normal and lung, pancreatic, and insulinomaneoplastic tissues by direct sequencing but neithercausative nor functional mutations were found exceptfor thepresence of twobase substitutions in theminimalpromoter region in two well differentiated lung adeno-carcinoma that led to a significant increase in thetranscription of thegene (Cattaneoet al., 2001) (Fig. 1C). A polymorphic base substitution was reported in thefibronectin type II domain of the gene in childrensaffected by persistent hyperinsulinemic hypoglycemia

(insulinoma) of infancy, which induces a major changein the amino acid composition (Saltini et al., 2004). The

absence of SEL1L mutations in cancer and normal cellssuggests that SEL1L must retain its integrity.

SEL1L differential expression in cancer

SEL1L is differentially expressed in cancer tissuescompared to normal counterpart. Analysis of a series ofhuman primary breast carcinomas and pancreatic

adenocarcinoma revealed downmodulation or absenceof SEL1L protein in about two-thirds of the tumors ascompared to normal epithelial cells (Orlandi et al.,2002a; Cattaneo et al., 2003). The overall survivalanalysis of breast carcinoma patients indicated astatistically significant correlation between SEL1Ldownmodulation and poor prognosis (Orlandi et al.,2002a). All in situ breast tumors show very strongSEL1L positivity (Orlandi, personal communication). Ithas been documented that during the transition fromhyperplasia to in situ breast carcinoma, cancer cellsshow higher levels of genomic instability and a highernumber of aberrations (Chin et al., 2004). If the genomicinstability is paralleled with the maximum productionof mutated proteins, then it canbe hypothesizedthat the

increase in SEL1L transcription and translation isrequired because the increase of intracellular proteindegradation.

In agreement, immunohistochemical analysis per-formed on oesophageal, prostate, and non-small celllung cancer documented that SEL1L protein becomesconsistently expressed in the initial stages of cancer-ogenesis and persists in neoplastic cells, making it auseful biomarker of cell transformation with relevantbiological and clinical implications (Granelli et al., 2004;Barberis et al., 2005; Ferrero et al., submitted). It isworth noting the dual sub-cellular localization (nucleus/cytoplasm/nucleo-cytoplasmic) of the SEL1L protein inthe majority of non-small cell lung carcinomas where itis associated with tumor histotype, indeed the cytoplas-mic immunostaning is preferentially observed in squa-mous cell carcinomas while nuclear expression issignificant associated to adenocarcinomas (Ferreroet al., submitted). Nucleocytoplasmic transfer of theSEL1L protein may result from specific alternativesplicing exons, in response to several forms of genotoxicinsults. The differential subcellular location is evidentnot only in neoplastic cells but also in fetal tissues,reflecting a common regulatory mechanism and com-mon SEL1L variant activation. However in healthyadult tissues, the cytoplasmic location predominatesover the nucleus, indicating that different isoforms areactivated to elicit perhaps the same function in differentcell compartments. There are well known examples of

protein isoforms, i.e., the E3-ligase BRCA1, showingcompletelydifferent localization in the cytoplasmversusnucleus between tumors and normal tissue, due todifferent relative expression levels of isoforms (Wilsonet al., 1997). Overall, tissue integrity may be dependenton the balanced expression level of different isoforms intissue and development-specific pattern.

Since no causative nor functional alterations werefound in the SEL1L genomic gene, the differentialrepresentation of the gene product in differentcells and in particular in cancer may be due to bothpost-transcription (methylation) and post-translation(glycosilation or others) regulations. Significant is theincreasing levels of SEL1L protein in those cells under-going transition from a benign to tumors stage as

reported for breast, lung, esophagus, and prostatecancers (Orlandi et al., 2002a; Ferrero et al., submitted;




12/17

Granelli et al., 2004; Barberis et al., 2005). Unpublishedresults show different methylation status of the singleCpG island located in the minimal promoter region ofthe gene (Cattaneo, personal observations) among asmall series of pancreatic cancer cell lines.

SEL1L possible function in cancer

Two biological systems were used to initially evaluate

the role of SEL1L in cancer: the human breast MCF-7and pancreatic Suit-2 cancer cell lines both transfectedwith the entire SEL1L cDNA driven by an induciblepromoter (Orlandi et al., 2002a; Cattaneo et al., 2003). Adrastic reduction in anchorage-dependent growthand colony formation in soft agar was observed inboth biological systems and this involved cell-matrixcommunication, in pancreatic cancer cells was alsodocumented a decreased invasive ability to penetrate amatrigel coated-filter, an inhibition of tumor growth inimmunodeficient mice andan alteration in the cell-cycleprogression (Cattaneo et al., in press).

The reversion of the pancreatic phenotype correlateswith the modulated expression of mediators involved inthe remodeling of the extra-cellular matrix, such as the

matrix metalloproteinases MMP1, MMP7, the inhibi-tors of MMPs TIMP1, TIMP2 as well as the tumorsuppressor gene PTEN and members of TGF-b signal-ing, such as Smad4, Activin A, and Activin receptor IIgenes (Cattaneo et al., in press).

The proteomic approach and global expression screen-ing performed on breast system identified in response toSEL1L signaling, the modulation of the expression ofseveral proteins, and transcripts operating in differentsignaling pathways (TGF beta and Notch pathway), incytoskeletal reorganization as well as tumor associatedproteins (Bianchi et al., 2005).

A deeper description of the Notch and TGF-b path-ways in relation to SEL1L is outlined in Sections 4.4.1and 4.4.2.

SEL1L and the NOTCH1 pathway. The Notchpathway, although originally identified in fruit flies, isnow among the most heavily studied in mammalianbiology; it is best characterized as a mediator of the cell-cell signaling between adjacent cells to generate cellularheterogeneity. Considerable effort has been placed intrying to dissect out the role of the proteins in signalingpathways mediated by Notch. Genetic screens basedon phenotype or suppressor/enhancer activity haveidentified a number of genes that influence lin-12/Notchactivity both in C. elegans and in Drosophila (Green-wald, 1998). Sel-1 (suppressor-enhancer of Lin12/Notch) is considered to be the C. elegans ortholog ofthe human SEL1L. It was identified as an extragenic

suppressor of mutations that reduces lin-12, it behavesas a negative regulator of the receptor activityby controlling its turnover (Grant and Greenwald,1996).

Notch activity and signaling is a highly conservedevolutionary pathway known for decades to develop-mental biologists. It is used not only by metazoans tocontrol their cell fate, but also by humans to influencedifferentiation, proliferation, and apoptotic eventsthroughout all the developmental stages (Miele andOsborne, 1999; Iso et al., 2003). Indeed there are onlyfew embryonic tissues that are not influenced by Notchsignaling. Four NOTCH genes are known in rodents andhumans differing in the number EGF-like repeats andthe length of the intracellular domain, however a

comprehensive discussion of Notch signaling is beyondthe scope of this review.

In the mammalian central nervous system Notchsignaling is implicated in processes ranging from neuralstem cell regulation (to maintain the stem cells in aprogenitor state) to learning and memory (Yoon andGaiano, 2005). During the organogenesis of the pan-creas, Notch signaling locks the pancreatic epithelialcells in an undifferentiated progenitor like state whilemaintaining their proliferative capacity through

Fgf10 signaling, and this appears to be critical for thedecision between the progenitor and endocrine fates.Hart and collaborators reported that the persistentexpression of Fgf10 in the embryonic pancreas oftransgenic mice decreases sel-1 expression, suggestingthat reduced sel-1 expression may result in an atypicalmaintenance of activated Notch thus impairing celldifferentiation (Hart et al., 2003).

Over the last 12 years, evidence has steadily accumu-lated in relating deregulation of Notch signaling inseveral malignancies, and it is being considered as apotential target for therapeutic intervention (Nickoloffet al., 2003). The first link was the 9:7 chromosomaltranslocation associated with about 10% of T-celllymphoblastic leukemias, the translocation produces a

truncated Notch-1, resulting in the constitutive activa-tion of the gene. The truncated version of the Notchisoforms have transforming activity both in vitro and invarious animal model systems. Several solid tumors andhematological malignancies, a subset of acute myeloidleukemias and B-cell chronic lymphoid leukemias havederegulated expression of wild-type Notch receptors,ligands, and targets.

The role of human SEL1L in Notch signaling isstill largely unexplored, the only published data arerestricted to twoexpression studies. Oneof these studiesexamined the transcriptional levels of Notch1, HES1,and SEL1L genes in leukemia and lymphoma cell lines,while the other reported the upmodulation of the jagged2 and Notch 3 transcripts in breast mammary cancercells MCF-7 expressing ectopic SEL1L (Chiaramonteet al., 2002; Bianchi et al., 2005). The first work reportedthat SEL1L does not exert a negative regulatoryinfluence on Notch signaling since was not found aninverse relationship between SEL1L expression and thestatus of Notch signaling (Chiaramonte et al., 2002).Despite this preliminary evidence, the possible func-tional link between SEL1L and Notch activity may bedeeper investigated in vitro using more sophisticatedbiological models (gain/loss of function genetic screens)and in vivo during sequential stages of carcinogenesissince SEL1L expression reflects the different phases ofcancer progression.

The results obtained from the breast model system

prompt to investigate the SEL1L involvement in Notchsignaling in breast tumor. Studies of mammary tumor-igenesis induced by Notch in mouse and in vitro modelsprovide evidence that Notch activation is a causalfactor in human breast cancer (Politi et al., 2004). Itwas reported that the subversion of Numb/Notchbiological antagonism through specific Numb ubiquiti-nation and proteasomal degradation contributes tohuman mammary carcinogenesis (Pece et al., 2004). Itis becoming increasing clear that the activity of Notchand several receptor systems, is mediated by thecapacity of a cell to clear itself from the surfacereceptors, their ligands and/or other extracellularsoluble molecules. Endocytosis plays an importantrole and is required in both the signal-generating and

signal-receiving cell for appropriate Notch activation(Polo et al., 2004). The ubiquitin-proteosome pathway

34 BIUNNO ET AL.



13/17

appears to be a crucial role in the degradation of suchproteins and is therefore becoming a very importanttherapeutic target for cancer (Orlowski and Dees, 2003;Rajkumar et al., 2005).

It is then very important to explore the function ofSEL1L in the NOTCH pathway, as far as its capacity tomodulate Notch activity.

SEL1L and transforming growth factor-beta

(TGF-b) pathway. Studies performed on mammalsdocumented the involvement of SEL1L in the TGF-bpathway. The first line of evidence was reportedby Furue et al. (2001) while working with the ratRSMG-1 cells (a sub-mandibular gland epithelial cellline); in this system SEL1L mRNA expression wasinduced by Activin A (Furue et al., 2001). Second, it wasreported that SEL1L ability to reduce the aggressivebehavior of human neoplastic cells correlates with thetranscriptional modulation of genes belonging to theTGF-b pathway: the inducible expression of SEL1L inthe pancreatic cancer cell line Suit-2 upmodulates thetranscriptional levels of endogenous Activin A andSmad4 (Cattaneoet al., 2003), while SEL1L inactivationby RNAi resulted in the downmodulation of Activin A

and Activin receptor II transcripts (Cattaneo et al., inpress).

Similar results were obtained in the MCF-7 cells inwhich the ectopic expression of SEL1L resulted in themodulation in the transcription of Activin A receptor IBand the four signal transducers SMADs (Bianchi et al.,2005).

The TGF-b superfamily consists of a disparate groupof polypeptide cytokines that regulate a plethora ofbiological processes. TGF-b signaling proceeds from thecell membrane to the nucleus through the cooperation oftype I and II serine/threonine kinase receptors, andtheir downstream SMAD effectors. When TGF-b relaysteam of signalers fails to function properly, theresultingimbalance can lead to a variety of diseases, includingcancer, heart disease, and asthma. Too much of TGF-bactually turns the tumor suppressor pathway into atumor-promoting pathway by: (i) promoting angiogen-esis and (ii) suppressing T cells and other components ofthe immune system. As favor to suppressor pathway, Activin A exerts its effect on a wide range of cellulartargets by modulating cell differentiation and prolifera-tion (Chen et al., 2002) appearing as a potent cell growthinhibitor in the human breast cancer cell lines MCF-7and T47D (Liu et al., 1996; Cocolakis et al., 2001).Moreover, disruption or mutations in the TGF-bcomponents are involved in a high percentage of colon,breast, and pancreatic cancers (Lei et al., 1996; Schutteet al., 1996), and SMAD4 is inactivated by mutation or

homozygous deletion in about one-half of all pancreaticcancers (Rozenblum et al., 1997). Furthermore, DPC4/Smad4 re-expression in pancreatic carcinoma cell lineinduced the cell to shift from potently angiogenic toan antiangiogenic phenotype in vitro and in vivo bymodulating regulators of angiogenesis (Schwarte-Waldhoff et al., 2000).

The intensity and duration of the TGF-b-Smadresponse are important determinants for signalingspecificity, thus the activity of receptors and Smadsare carefully regulated. One mechanism of regulation isendocytic clearance.

Recent work has shown that TGF-b signaling path-ways can be compartmentalized through differentinternalization routes of the TGF-b receptors. Whereas

clathrin-dependent TGF-b receptor internalizationinto SARA-containing early endosomes promotes Smad

signaling, internalization via lipid raft-caveolar com-partments containing receptor bound to Smad7-Smurf2results in accelerated receptor turnover by promotingpoly-ubiquitination (Di Guglielmo et al., 2003).

Considering the pancreatic biological system, SEL1Lmay play a crucial role in the regulation of Smad7 and/orSmuf2 levels, since Smad7-Smurf2 are involved in theTGF-b pathway termination, SEL1L might be impor-

tant as an amplifier of TGF-b signaling.

Conclusion and prospective

This review has posed several questions concerningthe complexity of SEL1L in terms of its structure,physiology, and function in normal developmentalprocesses as well as in cancer biology. Considering itsstructure, we know it has a multimodular architectureraising the important issue such as: (i) which and howmany proteins can interact with SEL1L. Are the threemain proteinprotein interacting domains (fibronectintype II, sel-1-like repeats, proline-rich motif) adeptto bind-specific substrates? These questions can beaddressed by assaying the entire protein as well as eachdomain in yeast two-hybrid system, co-immunoprecipi-

tation, and pull down experiments. (ii) Are the nuclearsignals sufficient to shuffle the protein in and out of thenucleus? By tagging the SEL1L nuclear signals with thegreen fluorescent protein or peptide epitope tags willhelp to dissect out the mechanisms that regulate thenuclear/cytoplasmic transport.

Asfor theseveral computer predictedprotein isoformsas well as those experimentally isolated, it will beimportant to understand their biological function inrelation to the complete SEL1L. Are these isoformsrestricted to specific sub cellular compartments? Arethey regulated by specific stimuli or temporal cellconditions (developmental stages vs. adults; normal vs.tumor)? Is there one or more substrates/templates perisoform? Functional studies based on the use of fusionprotein isoforms with peptide epitope tags will aid toclarify these biochemical and functional aspects.

Considering the overall results published on SEL1Lby various investigators working in different organisms,it can, perhaps, safely be deduced that this gene plays afundamental role in eukaryotic intracellular proteindegradation. However, it still remains to be establishedin which step of the ubiquitin proteosoma pathwaySEL1L plays part as well as which domain retains thisfunction. Since the C-terminal region contains thehrd3 motif (the most conserved motif)than it is plausibleto hypothesize that it preserves the biological function.Both yeast and C. elegans complementation experi-ments have the potential to address this issue.

Arguably, the greatest difficulty is to understand howhuman SEL1L works in the protein degradationprocesses. The control of protein degradation is highlysophisticated mechanism which controls the organismswell being. Since it is a highly delicate and preciseprocess, the cell engages different kind of proteins (E3-ligase, chaperones and so on) to work in the samepathway finalized in the protein degradation. The typeof multi-protein complex that gets formed depends onthe kind of polypeptide, which needs to be degraded.Thus, does SEL1L interact with Hrd1p or with the otherseveral E3 ligases described? Does the increased proteincomplexity confer to SEL1L the multi-features toparticipate to several chaperone/E3 ligase pathways?Which are the classes of proteins that SEL1L target for

degradation (Notch, TGF-b pathways)? Is SEL1L regu-lated by ubiquitination?




14/17

Once these questions will be answered (or at least afew of these) than we will be able to tackle the issue of itsimplication(s) in cancer. The fundamental questionraised by the observation that it gets upmodulatedduring the early steps of tumor transformation is ofparamount importance for early diagnosis. Currently, itis only possible to hypothesize that the increased SEL1Llevels are required in order to meet the advent of genetic

and/or genomic structural alterations acquired duringcancer initiation or to influence intra-cellular signaling.Its presence may be important in protecting cellularhomeostasis from genetic mutations. This may beexplained by previous studies which revealed thatSEL1L specific knock-down by RNA interferenceresulted in severe perturbation of bTC-3 growth andmetabolic activity (Diaferia et al., 2004), and by theunsuccess in obtaining a SEL1L knock-out mouse(Biunno personal communication), thus highlightingthe notion that the constitutive expression of SEL1Lmay be essential for achieving and maintaining homeo-static balance.

Malignant cells exhibit an exceptional ability tomaintain homeostasis in an hostile environment,

upregulationof proteins involvedin the stress responseresults in an increase in the apoptosis resistance, andin diminishing cell senescence or permanent arrest ofcell division. Tumors are often subjected to hypoxia oracidosis, conditions in which stress-response proteins(UPR, PI3K/Akt, MEK1/ERK) enhance cell survival bymaintaining the normal protein-folding environment. Acute activation of endogenous PI3K/Akt governscell survival by directly counteracting ER stressinduced cell death (Hu et al., 2004). By modulatingPTEN expression, a tumor suppressor which inhibitsPI3K-dependent signaling, SEL1L may elicit its anti-tumor response by affecting survival mechanisms. Inaddition to facilitating cell survival in stressfullenvironmental challenges, UPR proteins also allowtumor cells to tolerate alterations from within, includ-ing mutation of critical signaling molecules that wouldotherwise be lethal. Thus, UPR proteins can serve asbiochemical buffers for the genetic instability that ischaracteristic of many human cancers, facilitatingand maintaining the transformed phenotype.The increased expression of chaperone proteins(Hps70,Hps90, Hps27) over the level seen in normaltissues seems to be a common finding in many humancancers, both solid and hematologic malignancies(Conroy et al., 1998; Vargas-Roig et al., 1998; Cioccaand Calderwood, 2005).

Protein degradation is becoming a central theme incancer biology and recently therapeutic approachesthat

use inhibitors of proteins belonging to ubiquitin-proteo-soma pathway have been developed in solid tumors andhematological diseases (Orlowski and Dees, 2003;Bagatell and Whitesell, 2004; Rajkumar et al., 2005).SEL1L is a new entry in this process and although itsprecise allocation within this pathway is still largelyunknown, of paramount importance is to identify thenetworks controlled by SEL1L as well as thesignals thatregulate its activity in order to develop specific ther-apeutic approaches both in cancer as well as otherpathological disorders.

LITERATURE CITED

Aberle S, Schug N, Mathlouthi R, Seitz G, Kupper JH, Schroder K, Blin N.

2004. Promoter selection for the cytosine deaminase suicide gene constructs ingastric cancer. Eur J Gastroenterol Hepatol 16:6367.

Allen JR, Nguyen LX, Sargent KE, Lipson KL, Hackett A, Urano F. 2004.High ER stress in beta-cells stimulates intracellular degradation of misfoldedinsulin. Biochem Biophys Res Commun 324:166170.

Artavanis-Tsakonas S, Rand MD, Lake RJ. 1999. Notch signaling: Cellfate control and signal integration in development. Science 284:770776.Review.

Bagatell R, Whitesell L. 2004. Altered Hsp90 function in cancer: A uniquetherapeutic opportunity. Mol Cancer Ther 3:10211030. Review.

Ban Y, Taniyama M, Tozaki T, Yanagawa T, Tomita M, Ban Y. 2001. SEL1Lmicrosatellite polymorphism in Japanese patients with autoimmune thyroiddiseases. Thyroid 11:335338.

Barberis M, Roz E, Biunno I. 2005. SEL1L expression in prostatic intraepithelialneoplasia and adenocarcinomas: An immunohistochemical study. Histopatho-logy (in press).

Bianchi L, Canton C, Bini L, Orlandi R, Menard S, Armini A, Cattaneo M, PalliniV, Bernardi LR, Biunno I. 2005. Protein profile changes in the human breastcancer cell line MCF-7 reponse to SEL1L gene induction. Proteomics 5:24332442.

Biunno I, Appierto V, Cattaneo M, Leone BE, Balzano G, Socci C, Saccone S,Letizia A, Della Valle G, Sgaramella V. 1997. Isolation of a pancreas-specificgene located on human chromosome 14q31: Expression analysis in humanpancreatic ductal carcinomas. Genomics 46:284286.

Biunno I, Bernard L, Dear P, Cattaneo M, Volorio S, Zannini L, Bankier A, ZolloM. 2000. SEL1L, the human homolog of C. elegans sel-1: Refined physicalmapping, gene structure and identification of polymorphic markers. HumGenet 106:227235.

Biunno I, Castiglioni B, Rogozin IB, DeBellis G, Malferrari G, Cattaneo M. 2002.Cross-species conservation of SEL1L, a human pancreas-specific expressinggene. OMICS 6:187198.

Blatch GL, Lassle M. 1999. The tetratricopeptide repeat: A structuralmotif mediating proteinprotein interactions. Bioessays 21:932939. Review.

Bruder E, Moch H. 2004. Pediatric renal cell carcinoma. Pathologe 25:324

327.CapobiancoAJ, ZagourasP, Blaumueller CM, Artavanis-Tsakonas S, Bishop JM.1997. Neoplastic transformationby truncatedalleles of human NOTCH1/TAN1and NOTCH2. Mol Cell Biol 17:62656273.

Cattaneo M, Orlandi R, Ronchini C, Granelli P, Malferrari G, Menard S, BiunnoI. 2000. The expressionof SEL1L and TAN-1 in normaland neoplastic cells. Int

J Biol Markers 15:2632.Cattaneo M, Zollo M, Malferrari G, Orlandi R, DAngelo A, Menard S, Biunno I.

2001a. Allelic polymorphisms in the transcriptional regulatory region ofhuman SEL1L. Mutat Res 458:7176.

Cattaneo M, Sorio C, Malferrari G, Rogozin IB, Bernard L, Scarpa A, Zollo M,Biunno I. 2001b. Cloning and functional analysis of SEL1L promoter region, apancreas-specific gene. DNA Cell Biol 20:19.

Cattaneo M, Orlandini S, Beghelli S, Moore PS, Sorio C, Bonora A, Bassi C,Talamini G, Zamboni G, Orlandi R, Menard S, Bernardi LR, Biunno I, Scarpa

A. 2003. SEL1L expression in pancreatic adenocarcinoma parallels SMAD4expression and delays tumor growth in vitro and in vivo. Oncogene 22:63596368.

Cattaneo M, Canton C, Albertini A, Biunno I. 2004. Identification of a regionwithin SEL1L protein required for tumour growth inhibition. Gene 326:149156.

Cattaneo M, Fontanella E, Canton C, Delia D, Biunno I. SEL1L affects humanpancreatic cancer cell cycle and invasiveness through modulation of PTEN andgenes related to cellmatrix interactions. Neoplasia (in press).

Chen YG, LuiHM,Lin SL,LeeJM, Ying SY.2002.Regulationof cell proliferation,apoptosis, and carcinogenesis by activin. Exp Biol Med (Maywood) 227:7587.Review.

Chiaramonte R, Calzavara E, Basile A, Comi P, Sherbet GV. 2002. Notch signaltransduction is not regulated by SEL1L in leukaemia and lymphoma cells inculture. Anticancer Res 22:42114214.

Chin K, de Solorzano CO, Knowles D, Jones A, Chou W, Rodriguez EG, Kuo WL,Ljung BM, Chew K, Myambo K, Miranda M, Krig S, Garbe J, Stampfer M,

Yaswen P, Gray JW, Lockett SJ. 2004. In situ analyses of genome instability inbreast cancer. Nat Genet 36:984988.

Ciocca DR, Calderwood SK. 2005. Heat shock proteins in cancer: Diagnostic,prognostic, predictive, and treatment implications. Cell Stress Chaperones10:86103.

Clark HF, Gurney AL, Abaya E, Baker K, Baldwin D, Brush J, Chen J, Chow B,Chui C, Crowley C, Currell B, Deuel B, Dowd P, Eaton D, Foster J, Grimaldi C,Gu Q, Hass PE, Heldens S, Huang A, Kim HS, Klimowski L, Jin Y, Johnson S,Lee J, Lewis L, Liao D, Mark M, Robbie E, Sanchez C, Schoenfeld J, SeshagiriS, Simmons L, Singh J, Smith V, Stinson J, Vagts A, Vandlen R, Watanabe C,

Wieand D, Woods K, Xie MH, Yansura D, Yi S, Yu G, Yuan J, Zhang M, Zhang

Z, Goddard A, Wood WI, Godowski P, Gray A. 2003. The secreted proteindiscoveryinitiative(SPDI), a large-scale effort to identifynovel human secretedand transmembrane proteins: A bioinformatics assessment. Genome Res13:22652270.

Cocolakis E, Lemay S, Ali S, Lebrun JJ. 2001. The p38 MAPK pathway isrequired for cell growth inhibition of human breast cancer cells in response toactivin. J Biol Chem 276:1843018446.

Conroy SE, Sasieni PD, Fentiman I, Latchman DS. 1998. Autoantibodies to the90 kDa heat shock protein and poor survival in breast cancer patients. Eur JCancer 34:942943.

Das AK, Cohen PW, Barford D. 1998. The structure of the tetratricopeptiderepeats of protein phosphatase 5: Implications for TPR-mediated proteinprotein interactions. EMBO J 17:11921199.

Di Guglielmo GM, Le Roy C, Goodfellow AF, Wrana JL. 2003. Distinct endocyticpathways regulate TGF-beta receptor signalling and turnover. Nat Cell Biol5:410421.

Diaferia G, Cattaneo M, Saltini G, Proverbio MC, Monferini E, Malferrari G, Albertini A, Biunno I. 2004. RNA-mediated interference indicates thatSEL1L plays a role in pancreatic beta-cell growth. DNA Cell Biol 23:510518.

Dichamp C, Taillibert S, Aguirre-Cruz L, Lejeune J, Marie Y, Kujas M, Delattre

JY, Hoang-Xuan K, Sanson M. 2004. Loss of 14q chromosome in oligoden-droglial and astrocytic tumors. J Neurooncol 67:281285.

36 BIUNNO ET AL.



15/17

Dissemond J, Busch M, Kothen T, Mors J, Weimann TK, Lindeke A, Goos M, Wagner SN. 2004. Differential downregulation of endoplasmic reticulum-residing chaperones calnexin and calreticulin in human metastatic melanoma.Cancer Lett 203:225231.

Donoviel DB, Donoviel MS, Fan E, Hadjantonakis A, Bernstein A. 1998. Cloningand characterization of Sel-1l, a murine homolog of the C. elegans sel-1 gene.Mech Dev 78:203207.

Ellisen LW, Bird J, West DC, Soreng AL, Reynolds TC, Smith SD, Sklar J. 1991.TAN-1, the human homolog of the Drosophila notch gene, is broken bychromosomal translocations in T lymphoblastic neoplasms. Cell 66:649661.

Furue M, Zhang Y, Okamoto T, Hata RI, Asashima M. 2001. Activin A inducesexpression of rat Sel-1l mRNA, a negative regulator of notch signaling, in ratsalivary gland-derived epithelial cells. Biochem Biophys Res Commun282:745749.

Gallahan D, CallahanR. 1997. The mouse mammary tumorassociated geneINT3is a unique member of the NOTCH genefamily (NOTCH4). Oncogene 14:18831890.

Granelli P, Cattaneo M, Ferrero S, Bottiglieri L, Bosari S, Fichera G, Biunno I.2004. SEL1L and squamous cell carcinoma of the esophagus. Clinical CancerRes 10:58575861.

Grant B, Greenwald I. 1996. The Caenorhabditis elegans sel-1 gene, a negativeregulator of lin-12 and glp-1, encodes a predicted extracellular protein.Genetics 143:237247.

Grant B, Greenwald I. 1997. Structure, function, and expression of SEL-1, anegative regulator of LIN-12 and GLP-1 in C. elegans. Development 124:637644.

GreenwaldI. 1998. LIN-12/Notchsignaling: Lessons from wormsand flies. GenesDev 12:17511762. Review.

Hampton RY. 2002. ER-associated degradation in protein quality control andcellular regulation. Curr Opinion Cell Biol 14:476482.

Hampton RY, Gardner RG, Rine J. 1996. Role of 26S proteasome and HRD

genes in the degradation of 3-hydroxy-3-methylglutaryl-CoA reductase,an integral endoplasmic reticulum membrane protein. Mol Biol Cell 7:20292044.

Harada Y, Ozaki K, Suzuki M, Fujiwara T, Takahashi E, Nakamura Y, Tanigami A. 1999. Complete cDNA sequence and genomic organization of a humanpancreas-specific gene homologous to Caenorhabditis elegans sel-1. J HumGenet 44:330336.

Hart A, Papadopoulou S, Edlund H. 2003. Fgf10 maintains notch activation,stimulates proliferation, and blocks differentiation of pancreatic epithelialcells. Dev Dyn 228:185193.

Hu J, Pang JC, Tong CY, Lau B, Yin XL, Poon WS, Jiang CC, Zhou LF, Ng HK.2002. High-resolution genome-wide allelotype analysis identifies loss ofchromosome 14q as a recurrent genetic alteration in astrocytic tumours. Br JCancer 87:218224.

Hu P, Han Z, Couvillon AD, Exton J. 2004. Critical role of endogenous Akt/IAPsand MEK1/ERK pathways in counteracting endoplasmic reticulum stress-induced cell death. J Biol Chem 279:4942049429.

IsfortRJ, Cody DB,DoersenCJ, RichardsWG, Yoder BK,Wilkinson JE, Kier LD,Jirtle RL, Isenberg JS, Klounig JE, Woychik RP. 1997. The tetratricopeptiderepeat containing Tg737 gene is a liver neoplasia tumor suppressor gene.Oncogene 15:17971803.

Iso T, Kedes L, Hamamori Y. 2003. HES and HERP families: Multiple effectors ofthe notch signaling pathway. J Cell Physiol 194:237255.

KamauchiS, Nakatani H, Nakano C, Urade R. 2005. Geneexpression in responseto endoplasmic reticulum stress in Arabidopsis thaliana. FASEB J 272:34613476.

Kaneko M, Nomura Y. 2003. ER signaling in unfolded protein response. Life Sci74:199205.

Kaufman RJ. 2002. Orchestrating the unfolded protein response in health anddisease. J Clin Invest 110:13891398.

Kay BK, Williamson MP, Sudol M. 2000. The importance of being proline: Theinteraction of proline-rich motifs in signaling proteins with their cognatedomains. FASEB J 14:231241. Review.

Kikkert M, Doolman R, Dai M, Avner R, Hassink G, van Voorden S, Thanedar S,Roitelman J, Chau V, Wiertz E. 2004. Human HRD1 is an E3 ubiquitin ligaseinvolved in degradation of proteins from the endoplasmic reticulum. J BiolChem 279:35253534.

Kostova Z, Wolf DH. 2003. For whom the bell tolls: Protein quality control of theendoplasmic reticulum and the ubiquitin-proteasome connection. EMBO J22:23092317. Review.

Kressel M, Schmucker B. 2002. Nucleocytoplasmic transfer of the NF2 tumorsuppressor protein merlin is regulated by exon 2 and a CRM1-dependentnuclear export signal in exon 15. Hum Mol Genet 11:22692278.

KusB, GajadharA, Stanger K, Cho R, SunW, Rouleau N,Lee T, ChanD, WoltingC, Edwards A, Bosse R, Rotin D. 2005. A high throughput screen toidentify substrates for the ubiquitin ligase Rsp5. J Biol Chem 280:2947029478.

Larsen ZM, Angelo AD, Cattaneo M, Nerup J, Biunno I, Zollo M, Pociot F. 2001.Complete mutation scanning of the human SEL 1L gene: A candidate gene fortype 1 diabetes. Acta Diabetol 38:191192.

Lecca MR, Wagner U, Patrignani A, Berger EG, Hennet T. 2005. Genome-wideanalysis of the unfolded protein response in fibroblasts from congenitaldisorders of glycosylation type-I patients. FASEB J 19:240242.

LeeDJ, Koch WM,Yoo G, Lango M, Reed A,CalifanoJ, BrennanJA, WestraWH,Zahurak M, Sidransky D. 1997. Impact of chromosome 14q loss on survival inprimary head and neck squamous cell carcinoma. Clin Cancer Res 3:501505.

Lei J, Zou TT, Shi YQ, Zhou X, Smolinski KN, Yin J, Souza RF, Appel R, WangS, Cymes K, Chan O, Abraham JM, Harpaz N, Meltzer SJ. 1996. InfrequentDPC4gene mutationin esophageal cancer, gastric cancer andulcerative colitis-associated neoplasms. Oncogene 13:2459 2462.

Lewinski MK, Bisgrove D, Shinn P, Chen H, Hoffmann C, HannenhalliS, VerdinE, Berry CC, Ecker JR, Bushman FD. 2005. Genome-wide analysis ofchromosomal features repressing human immunodeficiency virus transcrip-

tion. J Virol 79:66106619.

Liu QY, Niranjan B, Gomes P, Gomm JJ, Davies D, Coombes RC, BuluwelaL. 1996. Inhibitory effects of activin on the growth and morpholgenesis ofprimary and transformed mammary epithelial cells. Cancer Res 56:11551163.

Luthy L, Grutter MG, Mittl PR. 2004. The crystal structure of helicobactercysteine-richproteinC at 2.0 A resolution: Similar peptide-bindingsites in TPRand SEL1-like repeat proteins. J Mol Biol 340:829841.

Mathlouthi R, Aberle S, Schug N, Kupper JH, Schroder K, Seitz G, Blin N. 2003. Assessing optimal promoter activity for constructs in gastrointestinal genetherapy. Anticancer Res 23:40114015.

Medeiros LJ, PalmedoG, Krigman HR, Kovacs G, Beckwith JB. 1999. Oncocytoid

renal cell carcinoma after neuroblastoma: A report of four cases of a distinctclinicopathologic entity. Am J Surg Pathol 23:772780.Miele L, Osborne B. 1999. Arbiter of differentiation and death: Notch signaling

meets apoptosis. J Cell Physiol 181:393409. Review.Nickoloff BJ, Osborne BA, Miele L. 2003. Notch signaling as a therapeutic target

in cancer: A new approach to the development of cell fate modifying agents.Oncogene 22:65986608. Review.

Nobrega MA, Zhu Y, Plajzer-Frick I, Afzal V, Rubin EM. 2004. Megabasedeletions of gene deserts result in viable mice. Nature 431:988993.

Odom DT, Zizlsperger N, Gordon DB, Bell GW, Rinaldi NJ, Murray HL, VolkertTL, Schreiber J, Rolfe PA, Gifford DK, Fraenkel E, Bell GI, Young RA. 2004.Control of pancreas and liver gene expression by HNF transcription factors.Science 303:13781381.

Orlandi R, Cattaneo M, Troglio F, Casalini P, Ronchini C, Menard S, Biunno I.2002a. SEL1L expression decreases breast tumor cell aggressiveness in vivoand in vitro. Cancer Res 62:567574.

Orlandi R, Cattaneo M, Troglio F, Campiglio M, Biunno I, Menard S. 2002b.Production of a monoclonal antibody directed against the recombinant SEL1Lprotein. Int J Biol Markers 17:104111.

Orlowski RZ, Dees EC. 2003. The role of the ubiquitination-proteasomepathway in breast cancer: Applying drugs that affect the ubiquitin-

proteasome pathway to the therapy of breast cancer. Breast Cancer Res5:17. Review.Pear WS, Aster JC, Scott ML, Hasserjian RP, Soffer B, Sklar J, Baltimore D.

1996. Exclusive development of T cell neoplasms in mice transplantedwith bone marrow expressing activated Notch alleles. J Exp Med 183:22832291.

Pece S, Serresi M, Santolini E, Capra M, Hulleman E, Galimberti V, Zurrida S,Maisonneuve P, Viale G, Di Fiore PP. 2004. Loss of negative regulation byNumb over Notch is relevant to human breast carcinogenesis. J Cell Biol167:215221.

Petersen HH, Hilpert J, Militz D, Zandler V, Jacobsen C, Roebroek AJ, WillnowTE. 2003. Functional interaction of megalin with the megalinbinding protein(MegBP), a novel tetratrico peptide repeat-containing adaptor molecule. J CellSci 116:453461.

Pickford AR, Smith SP, Staunton D, Boyd J, Campbell ID. 2001. The hairpinstructure of the (6)F1(1)F2(2)F2 fragment from human fibronectin enhancesgelatin binding. EMBO J 20:15191529.

Pociot F, Larsen ZM, Zavattari P, Deidda E, Nerup J, Cattaneo M, ChiaramonteR, Comi P, Sabbadini M, Zollo M, Biunno I, Cucca F. 2001. No evidence forSEL1L as a candidate gene for IDDM11-conferred susceptibility. DiabetesMetab Res Rev 17:292295.

Politi K, Feirt N, Kitajewski J. 2004. Notch in mammary gland development andbreast cancer. Semin Cancer Biol 14:341347. Review.

Polo S, Pece S, Di Fiore PP. 2004. Endocytosis and cancer. Curr Opin Cell Biol16:156161. Review.

Ponting CP. 2000. Proteins of the endoplasmic-reticulum-associated degradationpathway: Domain detection and function prediction. Biochem J 351:527535.

Qu L, Koromilas AE. 2004. Control of tumor suppressor p53 function byendoplasmic reticulum stress. Cell Cycle 3:567570.

Rajkumar SV, Richardson PG, Hideshima T, Anderson KC. 2005. Proteasomeinhibition as a novel therapeutic target in human cancer. J Clin Oncol 23:630639. Review.

Richards WG, Yoder BK, Isfort RJ, Detilleux PG, Foster C, Neilsen N, WoychikRP, Wilkinson JE. 1997. Isolation and characterization of liver epithelial celllines from wild-type and mutant TgN737Rpw mice. Am J Pathol 150:11891197.

Richter K, Buchner J. 2001. Hsp90: Chaperoning signal transduction. J CellPhysiol 188:281290.

Ron D. 2002. Translational control in the endoplasmic reticulum stress response.J Clin Invest 110:13831388. Review.

Rozenblum E, Schutte M, Goggins M, Hahn SA, Panzer S, Zahurak M, GoodmanSN, Sohn TA, Hruban RH, Yeo CJ, Kern SE. 1997. Tumor-suppressivepathways in pancreatic carcinoma. Cancer Res 57:17311734.

Saltini G, Proverbio MC, Malferrari G, Biagiotti L, Boettcher P, Dominici R,Monferini E, Lorenzini E, Cattaneo M, Antonello D, Moore PS, Zamproni I,

Viscardi M, Chiumello G, Biunno I. 2004. Identification of a novel polymorph-ismin thefibronectintypeII domain of theSEL1L gene andpossible relationtothe persistent hyperinsulinemic hypoglycemia of infancy. Mutat Res 554:159163.

Schutte M, Hruban RH, Hedrick L, Cho KR, Nadasdy GM, Weinstein CL, BovaGS, Isaacs WB, Cairns P, Nawroz H, Sidransky D, Casero RA, Jr., Meltzer PS,Hahn SA, Kern SE. 1996. DPC4 gene in various tumor types. Cancer Res56:25272530.

Schwarte-Waldhoff I, Volpert OV, Bouck NP, Sipos B, Hahn SA, Klein-Scory S,Luttges J, Kloppel G, Graeven U, Eilert-Micus C, Hintelmann A, Schmiegel W.2000. Smad4/DPC4-mediated tumor suppression through suppression ofangiogenesis. Proc Natl Acad Sci USA 97:96249629.

StavridiES, HalazonetisTD. 2004.p53and stressin theER.GenesDev18:241244.Stefani M. 2004. Protein misfolding and aggregation: New examples in medicine

andbiologyof thedarksideof theproteinworld. BiochimBiophysActa 1739:525. Review.

Thorstensen L, Qvist H, Nesland JM, Giercksky KE, Lothe RA. 2001.Identification of two potential suppressor gene regions on chromosome arm14q that are commonly lost in advanced colorectal carcinomas. Scand J

Gastroenterol 36:13271331.




16/17

sel proliferation

Documents