Elsevier

Cellular Signalling

Volume 21, Issue 2, February 2009, Pages 196-211
Cellular Signalling

Review
Controlling cell surface dynamics and signaling: How CD82/KAI1 suppresses metastasis

https://doi.org/10.1016/j.cellsig.2008.08.023Get rights and content

Abstract

The recent identification of metastasis suppressor genes, uniquely responsible for negatively controlling cancer metastasis, are providing inroads into the molecular machinery involved in metastasis. While the normal function of a few of these genes is known; the molecular events associated with their loss that promotes tumor metastasis is largely not understood. KAI1/CD82, whose loss is associated with a wide variety of metastatic cancers, belongs to the tetraspanin family. Despite intense scrutiny, many aspects of how CD82 specifically functions as a metastasis suppressor and its role in normal biology remain to be determined. This review will focus on the molecular events associated with CD82 loss, the potential impact on signaling pathways that regulate cellular processes associated with metastasis, and its relationship with other metastasis suppressor genes.

Introduction

Metastatic cancer remains an incurable disease. The successful identification and analysis of tumor suppressor genes responsible for the initiation of primary tumors, and identification of the genes that they suppress, has been key to the successful development of new cancer therapies. It is logical to propose that a similar approach, i.e. identification of metastasis suppressor genes, would be equally beneficial to curing metastatic disease. Indeed, such genes exist; over 20 metastasis suppressor genes have been identified based on their specific ability to suppress metastasis, but not primary tumor growth, in xenograft models [1], [2]. Many follow-up studies in human cancer tissues support their role as metastasis suppressors, as loss of expression is observed almost exclusively in metastatic disease. The normal biological function of several, but not all, metastasis genes are known; however, few have been characterized with respect to how their loss promotes tumor metastasis. It is not even known if loss of any of these genes in vivo is sufficient or absolutely required for metastasis. This review will focus on what is known about the metastasis suppressor gene, KAI1/CD82. Despite intense scrutiny, many aspects of how CD82 specifically functions as a metastasis suppressor and its role in normal biology remain to be determined.

The metastasis suppressor function of KAI1/CD82 was first detected in a genetic screen using the metastatic rat AT6.1 prostate cancer cell line expressing fragments of chromosome 11, on which CD82 is located [3]. Re-expression of CD82 in AT6.1 cells and subsequent subcutaneous injection into nude mice significantly reduced metastases without affecting primary tumor growth. Numerous xenograft studies using other metastatic cell lines, including MDA-MB-435, breast LCC6, liver MHCC97-H, lung LLC, HT1080 sarcoma, and prostate LNCaP, further confirmed the metastasis suppressor function of CD82 [4], [5], [6], [7], [8], [9]. The route of injection, subcutaneous, orthotopic, or tail vein, did not impact the outcome and no major effects on primary tumor growth were seen in any model.

That CD82 is a valid metastasis suppressor gene is further supported by numerous clinical studies. Loss of CD82, both protein and mRNA, is strongly correlated with poor prognosis in many malignancies (reviewed in [10]), including prostate, colon, lung, pancreatic, breast, ovarian, and several others. While the direct association between CD82 and metastasis per se is not always straight forward, the number of reports citing CD82 loss out-numbers five to one those reporting an increase or no change. In two prostate cancer mouse models, down regulation of CD82 expression was observed in metastatic disease (CKM unpublished data)[11]. Thus CD82 loss is highly correlative with metastatic disease, CD82 loss occurs at the transcriptional level, and based on xenograft animal models CD82 suppresses metastasis. Furthermore, CD82 loss is likely to be a universal event in the development of metastasis since it is observed in many types of cancer.

CD82 was first cloned based on its affinity for several monoclonal antibodies (R2, IA4, C33, 4F9) that recognize lymphocytic surface antigens [12], [13], [14], [15], [16]. CD82 was subsequently assigned membership in the Cluster of Differentiation antigens [17]. CD82 is a member of the 4-span transmembrane super family (TM4SF) of type III membrane proteins, specifically of the tetraspanin subgroup (Tspan). There are 33 tetraspanins in the human genome (Table 1). Tetraspanin proteins are not present in yeast or bacteria, but are present in fungi and all multicellular organisms [18]. Direct functional comparisons between divergent species are difficult because of low conserved DNA sequence homology.

CD82 and several other tetraspanins are fairly ubiquitously expressed. Northern analysis of human tissues reveals high expression of CD82 in the spleen, thymus, prostate, ovary, small intestine, colon, placenta, lung, liver, kidney, and pancreas. Significantly, lower expression is seen in the heart, brain, muscle, and testis [3]. Mouse CD82 mRNA was highest in the spleen, kidney, lung, and liver [19], with a similar distribution seen in the rat [20]. Immunostaining of mouse tissues reveals distinct subtissue distributions. Mouse CD82 protein is primarily localized to spleen lymphoid tissue, medullary collecting ducts and distal convoluted tubules of the kidney, arteriolar smooth muscle of the lung, hepatocytes and sinusoidal lining of the liver, islet cells, and is found in many epithelial cells including epidydimus, prostate; colon; bladder, ureter, urethra, uterus, ovary, oviducts, testes, and seminal vesicles. CD82 was also present in most vascular endothelium except arterioles and the brain [21].

There is little evidence for gene mutation, loss of heterozygosity, promoter mutation, or hypermethylation to explain the loss of CD82 expression in clinical isolates of metastatic cancers [22], [23], [24], [25], [26], [27], [28]. Altered transcription or splice variants remain as possible mechanisms for loss of CD82 mRNA. One splice variant in which exon 7 is deleted has been reported [29]. Spliced KAI1 mRNA was detected in metastatic and invasive tissues of gastric and bladder cancers as well as several cell lines [29], [30]. However, the level of spliced transcript was present at several-fold lower levels than full length mRNA, and did not correlate well with invasiveness or metastasis. Thus its significance in the etiology of metastasis remains unknown.

The human CD82 promoter is G-C rich and Tata-less, and contains an array of potential promoter elements including Sp1, AP-2, GATA-1, PEA3, NF-IL6, MEP1, Myb, TCF-1, HNF3, NF-1, zeste, and Ets binding sites [31], [32]. The mouse promoter contains many of the same putative promoter elements [19]. Several extracellular stimuli have been reported to enhance CD82 expression and include cytokines (IL-1β, IL-4, IL-6, IL-13, IFN-γ, TNF-α), growth factors (NGF), phorbol esters (PMA), drugs (Genistein, etoposide), and 8-bromo-cAMP [11], [33], [34], [35], [36], [37], [38], [39]. The mechanisms by which these stimuli regulate CD82 transcription are virtually unknown, except that NF-κB is known to mediate some of the effects of the cytokines.

Promoter deletion analysis initially identified three transcriptional regions in the CD82 promoter; an enhancer region (− 922 to − 846); a negative regulatory region (− 735 to − 197); and the minimal promoter (− 197 to + 351) [10], [32], [40]. A p53-like regulatory element, responsible for etoposide induction of CD82, is located at − 860 in the enhancer region and was initially intriguing given that loss of p53 in prostate cancer is a late event correlating with progression to metastasis. Despite the fact that over expressed p53 can bind to the promoter and enhance CD82 expression in transfected cells, the correlation between p53 loss and CD82 loss in clinical samples does not stand up, arguing against a strict one-to-one relationship [36], [41], [42]. Other reports suggest that various combinations of p53, AP2, or JunB binding at the extended AP2-p53-AP1 element are responsible for regulating full CD82 expression and binding of these factors may be differentially altered during metastasis [40], [43].

Cytokine induced CD82 expression in immune cells is mediated primarily via NF-κB [44]. In two p53 mutant epithelial cell lines, TNFα-induced CD82 expression was also dependent on NF-κB [45]. Subsequently, ChIP analysis revealed that NF-κB p50, but not p65, is present on the CD82 promoter. The CD82 promoter recruits NF-κB p50, Bcl3 (functionally related to IκBα) and the N-CoR/TAB2/HDAC3 corepressor complex, which results in transcriptional inactivation. IL-1β stimulation transiently recruits Tip60 to p50 bound at the CD82 promoter, which is coincident with loss of the N-CoR complex, increased acetylation and phosphorylation of histones, and recruitment of Pol II. Over expression of a Tip60/Fe65/APP complex was sufficient to displace the N-Cor complex [46]. The Fe65 transcription activation domain binds to the nucleosome assembly factor SET, which is required for Fe65-mediated transactivation. ChIP experiments demonstrated that a complex including Fe65/APP/Tip60 and SET is associated with the CD82 promoter. SET is required for full levels of CD82 transcription.

However, the Tip60 coactivator complex was not recruited to the CD82 promoter in metastatic prostate cancer cells due to low levels of Tip60 expression in these cells [8]. IL-1β-induced Tip60 expression and recruitment could be restored by inhibiting β-catenin expression. A reptin/β-catenin complex was detected at the CD82 promoter that was present only in metastatic cells. The reptin/β-catenin complex could be displaced after IL-1β stimulation by over expressing Tip60. Modulation of Tip60 or β-catenin levels in metastatic cells or normal cells respectively alters IL-β1-mediated matrigel invasion. It was proposed that high levels of the β-catenin–reptin complex, due to Wnt activation, and simultaneous down regulation of Tip60 act together to inhibit CD82 expression and drive metastasis. Subsequently it was shown that the reptin repressive function requires its sumoylation at Lys456 by SENP1/SUSP1 [47]. Blocking reptin sumoylation in metastatic prostate cells restored CD82 mRNA and decreased matrigel invasion.

A recent study in metastatic breast cancer cells suggests CD82 transcription can also be regulated at the level of genomic organization. SATB1 regulates gene expression by recruiting chromatin remodeling enzymes and transcription factors, and tethers multiple genomic loci via specialized DNA sequences to globally control transcription. RNAi-mediated knockdown of SATB1 in metastatic MDA-MB-231 breast cancer cells restored acinar polarity and inhibited tumor growth and metastasis in vivo. Intriguingly, CD82, in addition to several other known metastasis suppressor genes, including nm23, KiSS1, BRMS1, claudin 1, and E-cadherin, were coordinately up-regulated in the SATB1 deleted tumor cells [48]. Thus, loss of p53, enhanced Wnt/β-catenin signaling, stress activated Jnk, and increased expression of SATB1 could all work together or in various combinations to promote the loss of CD82 expression and metastasis.

The incomplete correlation between CD82 mRNA or protein loss and metastasis in some clinical specimens may reflect additional mechanisms for removing CD82 function. Inhibition of E3 ubiquitin ligase gp78 expression in highly metastatic HT1080 sarcoma cells suppressed metastasis, but had no effect on primary tumor growth [9]; an effect that mimics metastasis suppressor genes. CD82 was identified as a primary substrate of gp78. The E3 ligase activity of gp78 was required for its metastatic effects. Loss of gp78 resulted in increased CD82 expression and over expression of gp78 increased CD82 degradation. An inverse relationship between gp78 and CD82 expression was detected in human sarcoma tissue samples and inhibition of CD82 expression in gp78 negative cells restored metastasis. Thus, another possible mechanism for CD82 loss in metastatic tumors is its enhanced degradation or turnover.

Whether other post-transcriptional or translational mechanisms can account for loss of CD82 function in metastatic tumors remains to be addressed. At least two types of post-translational modifications, glycosylation and palmitoylation, are known to occur on CD82. Inhibition of these modifications affects CD82 function [49], [50]. Whether the enzymes responsible for CD82 modification are altered in metastatic tumors is not known.

Section snippets

CD82 function

Tetraspanin proteins function in many aspects of cell physiology. Several excellent reviews provide extensive information on the genetics, structure, and function of tetraspanins [51], [52], [53], [54]. Tetraspanins contain no intrinsic catalytic activity; therefore, current hypotheses favor a model whereby tetraspanins serve as master regulators of membrane organization, through interactions with surface molecules and each other. Through these interactions tetraspanins regulate a variety of

Acknowledgments

The author is supported by funding from the American Cancer Society (RSG CSM-109378) and the Department of Defense Prostate Cancer Research Program of the Office of Congressionally Directed Medical Research Programs (W81XWH-08-1-0053). Additional support was also provided by the generous gifts of the Van Andel Institute.

References (235)

  • L.J. Stafford et al.

    Int. J. Biochem. Cell Biol.

    (2008)
  • J.H. Kim et al.

    Nature

    (2005)
  • H. Tonoli et al.

    Trends Mol. Med.

    (2005)
  • L.H. El Touny et al.

    Biochem. Biophys. Res. Commun.

    (2007)
  • Y. Nojima et al.

    Cell. Immunol.

    (1993)
  • S. Huang et al.

    Genomics

    (2005)
  • M. Nagira et al.

    Cell. Immunol.

    (1994)
  • F.S. Liu et al.

    Gynecol. Oncol.

    (2000)
  • P. Jackson et al.

    Cancer Lett.

    (2000)
  • J.T. Dong et al.

    Genomics

    (1997)
  • B. Gellersen et al.

    Am. J. Pathol.

    (2007)
  • H. Akita et al.

    Cancer Lett.

    (2000)
  • T. Mashimo et al.

    Biochem. Biophys. Res. Commun.

    (2000)
  • S. Lebel-Binay et al.

    Cell Immunol.

    (1994)
  • A. Marreiros et al.

    Gene

    (2003)
  • P. Jackson et al.

    Urol. Oncol.

    (2002)
  • J. Li et al.

    J. Biol. Chem.

    (2001)
  • S.H. Baek et al.

    Cell

    (2002)
  • M. Ono et al.

    Biochem. Biophys. Res. Commun.

    (2000)
  • C.S. Stipp et al.

    Trends Biochem. Sci.

    (2003)
  • M. Seigneuret et al.

    J. Biol. Chem.

    (2001)
  • M. Seigneuret

    Biophys. J.

    (2006)
  • C. Lagaudriere-Gesbert et al.

    Cell. Immunol.

    (1997)
  • G. Horvath et al.

    J. Biol. Chem.

    (1998)
  • E. Rubinstein et al.

    Dev. Biol.

    (2006)
  • S. Kelic et al.

    Mol. Cell. Neurosci.

    (2001)
  • M.J. Heikens et al.

    Blood

    (2007)
  • L.M. Lau et al.

    Blood

    (2004)
  • M.W. Goschnick et al.

    Blood

    (2006)
  • A.J. Cowin et al.

    J. Invest. Dermatol.

    (2006)
  • Y. Takeda et al.

    Blood

    (2007)
  • B. He et al.

    J. Biol. Chem.

    (2005)
  • X.A. Zhang et al.

    J. Biol. Chem.

    (2003)
  • C.W. Rinker-Schaeffer et al.

    Clin. Cancer Res.

    (2006)
  • J.T. Dong et al.

    Science

    (1995)
  • K.K. Phillips et al.

    Cancer Res.

    (1996)
  • X. Yang et al.

    Cancer Res.

    (2001)
  • J.M. Yang et al.

    Liver Int.

    (2008)
  • T. Takeda et al.

    Cancer Res.

    (2007)
  • Y.C. Tsai et al.

    Nat. Med.

    (2007)
  • H.W. Gaugitsch et al.

    Eur. J. Immunol.

    (1991)
  • K.I. Virtaneva et al.

    Immunogenetics

    (1993)
  • M.L. Gil et al.

    J. Immunol.

    (1992)
  • T. Imai et al.

    J. Immunol.

    (1992)
  • P. Engel et al.

    Leuk. Lymphoma

    (1994)
  • H. Suzuki et al.

    Prostate

    (1998)
  • M.C. Custer et al.

    Prostate

    (2006)
  • J.T. Dong et al.

    Cancer Res.

    (1996)
  • K. Tagawa et al.

    Jpn. J. Cancer Res.

    (1999)
  • T. Miyazaki et al.

    Cancer

    (2000)
  • Cited by (88)

    • KAI1 reverses the epithelial-mesenchymal transition in human pancreatic cancer cells

      2019, Hepatobiliary and Pancreatic Diseases International
      Citation Excerpt :

      During EMT, a number of molecular alterations occur, including upregulation of mesenchymal markers (e.g., N-cadherin and vimentin) and downregulation of epithelial markers (e.g., E-cadherin) [9]. KAI1 is a member of the transmembrane 4 super family which regulates many critical biological events, including cell signaling, differentiation, motility and fertilization [10–12]. Previous research [10–12] demonstrated that KAI1 is an important suppressor of metastasis in PC and that down-regulated expression of KAI1 is closely associated with PC metastasis.

    View all citing articles on Scopus
    View full text