Cell. Mol. Life Sci. (2014) 71:3627–3631 DOI 10.1007/s00018-014-1668-2
Cellular and Molecular Life Sciences
Visions and reflections
Possible mechanisms and function of nuclear trafficking of the colony-stimulating factor-1 receptor Elisabetta Rovida · Persio Dello Sbarba
Received: 20 February 2014 / Revised: 28 May 2014 / Accepted: 11 June 2014 / Published online: 28 June 2014 © Springer Basel 2014
Abstract Receptor tyrosine kinases (RTK) have long being studied with respect to the “canonical” signaling. This includes ligand-induced activation of a receptor tyrosine kinase at the cell surface that leads to receptor dimerization, followed by its phosphorylation in the intracellular domain and activation. The activated receptor then recruits cytoplasmic signaling molecules including other kinases. Activation of the downstream signaling cascade frequently leads to changes in gene expression following nuclear translocation of downstream targets. However, RTK themselves may localize within the nucleus, as either full-length molecules or cleaved fragments, with or without their ligands. Significant differences in this mechanism have been reported depending on the individual RTK, cellular context or disease. Accumulating evidences indicate that the colony-stimulating factor-1 receptor (CSF-1R) may localize within the nucleus. To date, however, little is known about the mechanism of CSF-1R nuclear shuttling, as well as the functional role of nuclear CSF-1R. Keywords c-Fms · Nuclear receptor tyrosine kinase · Trafficking · Signaling · Nuclear translocation
Regulated intramembrane proteolysis of CSF‑1R preludes to its shuttling from cell surface to intracellular compartments Regulated intramembrane proteolysis (RIP), has been reported for several receptor tyrosine kinases (RTK), E. Rovida (*) · P. Dello Sbarba Dipartimento di Scienze Biomediche Sperimentali e Cliniche “Mario Serio”, Sezione di Patologia e Oncologia Sperimentali, Istituto Toscano Tumori, Università degli Studi di Firenze, Viale G.B. Morgagni 50, 50134 Florence, Italy e-mail: [email protected]
including CSF-1R . This process is a well-characterized system for the nuclear trafficking of transmembrane receptors [2, 3]. Indeed, two major paradigmatic pathways have emerged. One is involved in the proteolytic cleavage of receptor-like tyrosine kinase such as Ryk that lacks kinase activity in the intracellular domain but is able to elicit wntdependent signaling. Specifically, proteases cleave Ryk in a ligand-independent manner, thus releasing the extracellular domain (ECD). Subsequently, the γ-secretase complex generates Ryk intracellular domain (ICD) that remains in the cytoplasm. Subsequent Ryk ICD translocation to the nucleus is Wnt ligand dependent. Notch RIPing is the prototype of the second pathway. Indeed, Notch activation by cognate ligands such as Delta triggers ECD shedding by the metalloprotease a disintegrin and metalloprotease (ADAM) 10 or ADAM17 (also referred to as tumor necrosis factor-α converting enzyme, TACE), which triggers subsequent Notch ICD release following cleavage by the γ-secretase complex. Notch ICD then enters the nucleus and activates target genes without a need of further Notch activation. CSF-1R down-modulation by ectodomain shedding is induced by the binding of its ligand, or following PKC activation by toll-like receptor-activating molecules, including LPS [4–10], or by Interleukins (specifically IL-4 and IL-2) [11–13]. The shedding of human CSF-1R ectodomain generates an about 100-kDa ECD, corresponding to the ligandbinding domain of CSF-1R (Fig. 1) . More than 10 years ago, we identified TACE/ADAM17 as the protease responsible for CSF-1R ectodomain shedding . The site where TACE cleaves CSF-1R is still unknown. Although no consensus amino acid sequences have been identified for the cleavage site in TACE substrates, it is known that TACE cleaves its target molecules ~12–16 amino acids from the transmembrane domain. Moreover, the cleavage site is
E. Rovida, P. D. Sbarba
Fig. 1 Human CSF-1R protein. Schematic representation of CSF-1R. TM transmembrane domain; ECD extracellular domain, ICD intracellular domain; white rectangle, plasma membrane; Section sign data from UniProt (http://www.uniprot.org/uniprot/P07333); asterisk data from . Amino-acidic sequences adjacent to and within the
transmembrane domain of human and murine CSF-1R are shown. Hydrophobic amino acids are highlighted in gray; asterisks indicate aliphatic residues in the region where possible TACE cleavage sites are located
Fig. 2 Possible pathways for CSF-1R nuclear trafficking. a CSF1R intracellular domain signaling. Following activation of CSF-1R by its ligand or other stimuli, a two-step proteolytic cleavage leads to the release of an extracellular (ECD) and an intracellular domain (ICD) of CSF-1R in the extracellular environment or into the cytosol, respectively. Whether ICD is phosphorylated or interacts with other molecules otherwise is not known. For the indicated enzymes, as well as the putative nuclear localization of CSF-1R ICD, see the text. b
Signaling by nuclear CSF-1R holoreceptor. Whole-length CSF-1R, possibly following interaction with its ligand CSF-1, translocates to the nucleus. Whether this occurs via the endocytic internalization of the receptor or via other mechanisms has not been clarified. It is unknown whether nuclear shuttling involves the monomeric or dimerized CSF-1R. Whether the mechanisms a and b have distinct roles and/or are cell- or disease-associated is also not known. NLS nuclear localization signal
likely to be located in proximity of an aliphatic hydrophobic residue [15, 16] (Fig. 1). A later study identified the mechanism of intramembrane proteolysis of CSF-1R operated by γ-secretase, leading to the release of an about 50 kDa (in the case of human CSF-1R or 45 kDa in the case of murine CSF-1R)
CSF-1R ICD . In that paper and in subsequent work, the authors speculated on the possible nuclear translocation of CSF-1R ICD [9, 10]. Indeed, the above-mentioned paper  showed the appearance of “some nuclear staining” when using antibodies against CSF-1R in immunofluorescence experiments performed in p388D1 macrophages.
Trafficking of CSF-1R to the nucleus
The authors themselves were cautious about saying that the results indicated definitively a nuclear location of CSF-1R. Moreover, there was no evidence that the immuno-positivity observed was related to the CSF-1R ICD. In fact, similar results could have been obtained for the whole-length receptor. Therefore, the only conclusion that could be driven from that study was that further experiments were required to confirm the presence of CSF-1R ICD in the nucleus. Of course, this would have not been a surprise on the basis of research papers published for RTK other than CSF-1R [1, 19, 20]. In a later paper, Glenn and van der Geer stated that CSF-1R ICD is apparently an unstable molecule, difficult to detect, and they mentioned unpublished preliminary results, which showed that a stable version of the ICD localizes to the nucleus . However, to date, no data have been published which undoubtedly demonstrate the presence of CSF-1R ICD in the nucleus (Fig. 2a). Mechanism of translocation to the nucleus Definitive evidence of nuclear CSF-1R ICD is still missing. Dual protease cleavage is necessary to release ICD, but the mechanism of its possible translocation from the cytosol to the nucleus is unknown (Fig. 2a). Functional role No functional role has been identified. Therapeutic targeting TACE and/or γ-secretase inhibitors should be used in case nuclear RIPed CSF-1R would be found and demonstrated to be linked to undesired biological actions.
forms that can enter the nucleus have been reported . For example, escaping from plasma membrane has been postulated for other RTK following changes in the hydrophobicity of transmembrane domain. We sequenced CSF1R mRNA from the SKBR3 breast cancer cell line, exhibiting nuclear CSF-1R, to identify possible nucleotide changes that could justify changes in the properties of CSF-1R protein. In these cells, CSF1R mRNA, when compared to the wild-type CSF-1R sequence, exhibits two single nucleotide changes (A83G; G726A) that are conservative. This excludes the presence of a splice variant lacking the transmembrane domain, or mutations responsible for changes in hydrophobicity. Moreover, no mutations have been reported in the transmembrane region of CSF-1R protein (http://www.uniprot.org/uniprot/P07333). The only variant reported consists of a polymorphism (L536V) that does not change the physicochemical property of CSF-1R domain, since both residues are medium size and hydrophobic. Other membrane-escaping mechanisms have been reported for other RTK [20, 23], but no information is available for CSF-1R. In our study, nuclear CSF-1R was undetectable in other CSF-1R-expressing cells, such as macrophages, where, by contrast, CSF-1R localization at the nuclear envelope had been reported . Indeed, full-length CSF-1R was detected at the nuclear envelope of bone marrow-derived murine macrophages as well as breast, ovarian and cervical cancer cell lines . Moreover, in the same paper, nuclear envelope-associated CSF-1R was reported to be phosphorylated upon CSF-1 treatment and to co-localize with CSF-1 in this cellular district. Perinuclear CSF-1R was also found able to transduce signals inside the nucleus (i.e., nuclear AKT phosphorylation). However, it should be noted that, at least in the case of epidermal growth factor receptor, the perinuclear compartment may serve as a reservoir for nuclear RTK transport .
Nuclear localization of CSF‑1R holoreceptor
Mechanisms of translocation to the nucleus
Holoreceptor nuclear localization has been reported for RTK, but is less characterized with respect to RIPing . We recently reported that whole-length CSF-1R localizes within the nucleus of breast cancer cell lines and tissues of different subtypes . We also showed that, once in the nucleus, CSF-1R may localize in the nucleolus. Our results provided the first evidence of CSF-1R localization within the nucleus. What we observed seems to involve the wholelength CSF-1R, thus excluding a RIP-dependent nuclear translocation. One of the most critical questions in nuclear translocation of holoreceptors is how transmembrane proteins could escape the plasma membrane and become soluble in the cytosol, so that they can be translocated to the nucleus. In this respect, RTK splice variants or mutated
The nuclear translocation of whole-length CSF-1R seems to be CSF-1-dependent, as nuclear CSF-1R increased following CSF-1 administration to serum-starved SKBR3 breast cancer cells. This is well in keeping with a previous study indicating that CSF-1 is produced by breast cancer cell lines and may be located in the nucleus in breast cancer tissues . Accordingly, we found for CSF-1 a relatively high nuclear localization signal (NLS) score, allowing to speculate that CSF-1R could translocate into the nucleus together with CSF-1. This hypothesis was strengthened via the demonstration by confocal immunofluorescence of the nuclear colocalization of CSF-1 and CSF-1R . We could not find an appreciable NLS score for interleukin-34, a recently discovered CSF-1R ligand . However, the
CSF-1-dependent, NLS-mediated mechanism may be only one of those driving nuclear translocations of CSF-1R. For instance, Fms-interacting protein (FMIP), a nucleocytoplasmic shuttling protein endowed with NLS, binds transiently to the cytoplasmic domain of, and is phosphorylated on tyrosine by, activated CSF-1R . However, whether FMIP is involved in CSF-1R nuclear localization has not been addressed. On the other hand, a possible interaction of CSF-1R with chaperone molecules, including heat shock proteins, involved in nuclear translocation has not been reported . Alternatively, an endocytic pathway that drives receptor internalization following, for example, ligand binding could be involved. This is indeed the mechanism which initiates the process of nuclear translocation of many RTK . On the whole, further experiments are needed to elucidate the mechanisms of whole-length CSF1R entry into the nucleus (Fig. 2b). Functional role We showed that nuclear CSF-1R is located in the chromatin-bound subcellular compartment. Moreover, chromatin immunoprecipitation experiments revealed that nuclear CSF-1R binds to the promoters of proliferation-related genes, such as CCND1, c-JUN and c-MYC, as well as to the promoter of its ligand, CSF-1. Of note, the expression of CCND1, c-JUN and c-MYC is inhibited following CSF-1R silencing . Further experiments, however, are required to address if and how nuclear CSF-1R works as a transcriptional regulator. In this respect, it should be noted that transcriptional activity has been found for other RTK . In conclusion, our data highlighted a novel aspect of CSF-1R function. Nuclear CSF-1R could work in parallel to, and synergizes with, the canonical CSF-1R signaling . Therapeutic targeting Further investigations have to be directed to determine whether nuclear CSF-1R is a druggable target and/or is suitable as a prognostic or predictive factor in breast cancer. Once the mechanism of holoreceptor CSF-1R nuclear translocation will have been elucidated, it will be possible to design strategies interfering with this mechanism.
Concluding remarks Few other papers reported nuclear localization of CSF-1R, but with information insufficient to discriminate whether this localization was due to RIPed or holoreceptor CSF1R translocation. Nuclear CSF-1R was indeed identified by immuno-histochemistry in cervical pre-neoplastic tissues . Whether this was the whole-length CSF-1R or
E. Rovida, P. D. Sbarba
a fragment of it was not addressed. However, since the antibody used was raised against the external domain of the feline v-Fms (the gene encoding for the viral oncogenic homolog of CSF1R gene; amino acids 399-411), it is straightforward to exclude that nuclear staining was due to RIPed CSF-1R. In another paper, Li and coworkers  found an increased amount of anti-CSF-1R-stained bands in total cell lysates following immortalization (by way of retroviral introduction of human telomerase reverse transcriptase) of normal ovarian surface epithelial primary cells. In particular, they detected a 108 and a 52 kDa CSF1R fragments within the immortalized cells. Whether the 108 kDa molecule is an immature form of CSF-1R or a truncated one was not addressed in the paper. Moreover, the authors detected in the nucleus both the 108 and the 52 kDa (two molecules with a molecular weight consistent with the CSF-1R ECD or ICD, respectively, see Fig. 1) fragments. This fact, apparently exclude RIPing since the 108 kDa molecule should have been shed in the extracellular environment before RIPing occurred. Current knowledge about the nuclear transport of either CSF-1R ICD or the whole-length CSF-1R is summarized in Fig. 2. The downstream mechanisms driving nuclear shuttling in particular awaits elucidation. With respect to a potential physiological role of nuclear CSF1R, it should be kept in mind that the effects of nuclear CSF-1R are necessarily flanked by those driven via the cytosolic signaling of membrane-associated CSF-1R. Nuclear CSF-1R has been indeed observed so far in neoplastic cells only. It is not known at the moment whether nuclear CSF-1R is associated with a worse prognosis, as it happens for other nuclear RTK . Nevertheless, interference with nuclear CSF-1R activity and/or transport should be taken into consideration whenever it is necessary to inhibit CSF-1/CSF-1R signaling for therapeutic purposes. Acknowledgments This work was supported by Associazione Italiana per la Ricerca sul Cancro, Istituto Toscano Tumori, Ministero della Salute, Regione Toscana, Fondazione Cassa di Risparmio di Volterra, Fondazione Oretta Bartolomei-Corsi. Conflict of interest The authors have no conflict of interest to disclose.
References 1. Ancot F, Foveau B, Lefebvre J, Leroy C, Tulasne D (2009) Proteolytic cleavages give receptor tyrosine kinases the gift of ubiquity. Oncogene 28:2185–2195 2. Zhong W (2008) Going nuclear is again a winning (Wnt) strategy. Dev Cell 15:635–636. doi:10.1016/j.devcel.2008.10.012 3. Bray SJ (2006) Notch signalling: a simple pathway becomes complex. Nat Rev Mol Cell Biol 7(9):678–689
Trafficking of CSF-1R to the nucleus 4. Chen BD, Lin HS, Hsu S (1983) Lipopolysaccharide inhibits the binding of colony-stimulating factor (CSF-1) to murine peritoneal exudate macrophages. J Immunol 130:2256–2260 5. Chen BD, Lin HS, Hsu S (1983) Tumor-promoting phorbol esters inhibit the binding of colony-stimulating factor (CSF-1) to murine peritoneal exudate macrophages. J Cell Physiol 116:207–212 6. Guilbert LJ, Stanley ER (1984) Modulation of receptors for the colony-stimulating factor, CSF-1, by bacterial lipopolysaccharide and CSF-1. J Immunol Methods 73:17–28 7. Downing JR, Roussel MF, Sherr CJ (1989) Ligand and protein kinase C down-modulate the colony-stimulating factor 1 receptor by independent mechanisms. Mol Cell Biol 9:2890–2896 8. Baccarini M, Dello Sbarba P, Buscher D, Bartocci A, Stanley ER (1992) IFN-gamma/lipopolysaccharide activation of macrophages is associated with protein kinase C-dependent downmodulation of the colony-stimulating factor-1 receptor. J Immunol 149:2656–2661 9. Glenn G, van der Geer P (2007) CSF-1 and TPA stimulate independent pathways leading to lysosomal degradation or regulated intramembrane proteolysis of the CSF-1 receptor. FEBS Lett 581:5377–5381 10. Glenn G, van der Geer P (2008) Toll-like receptors stimulate regulated intramembrane proteolysis of the CSF-1 receptor through Erk activation. FEBS Lett 582:911–915 11. Dello Sbarba P, Rovida E, Caciagli B, Nencioni L, Labardi D, Paccagnini A, Savini L, Cipolleschi MG (1996) Interleukin-4 rapidly down-modulates the macrophage colony-stimulating factor receptor in murine macrophages. J Leukoc Biol 60:644–650 12. Dello Sbarba P, Nencioni L, Labardi D, Rovida E, Caciagli B, Cipolleschi MG (1996) Interleukin 2 down-modulates the macrophage colony-stimulating factor receptor in murine macrophages. Cytokine 8:488–494 13. Dello Sbarba P, Rovida E (2002) Transmodulation of cell surface regulatory molecules via ectodomain shedding. Biol Chem 383:69–83 14. Rovida E, Paccagnini A, Del Rosso M, Peschon J, Dello Sbarba P (2001) TNF-alpha-converting enzyme cleaves the macrophage colony-stimulating factor receptor in macrophages undergoing activation. J Immunol 166:1583–1589 15. Caescu CI, Jeschke GR, Turk BE (2009) Active-site determinants of substrate recognition by the metalloproteinases TACE and ADAM10. Biochem J 424(1):79–88. doi:10.1042/BJ20090549 16. Horiuchi K (2013) A brief history of tumor necrosis factor α-converting enzyme: an overview of ectodomain shedding. Keio J Med 62(1):29–36 17. Sherr CJ (1990) Colony-stimulating factor-1 receptor. Blood 75(1):1–12 18. Wilhelmsen K, van der Geer P (2004) Phorbol 12-myristate 13-acetate-induced release of the colony-stimulating factor 1 receptor cytoplasmic domain into the cytosol involves two separate cleavage events. Mol Cell Biol 24:454–464
3631 19. Carpenter G (2003) Nuclear localization and possible functions of receptor tyrosine kinases. Curr Opin Cell Biol 15(2):143–148 20. Carpenter G, Liao HJ (2013) Receptor tyrosine kinases in the nucleus. Cold Spring Harb Perspect Biol 5:a008979 21. Barbetti V, Morandi A, Tusa I, Digiacomo G, Riverso M, Marzi I, Cipolleschi MG, Bessi S, Giannini A, Di Leo A, Dello Sbarba P, Rovida E (2013) Chromatin-associated CSF-1R binds to the promoter of proliferation-related genes in breast cancer cells. Oncogene. doi:10.1038/onc.2013.542 22. Wells A, Marti U (2002) Signalling shortcuts: cell-surface receptors in the nucleus? Nat Rev Mol Cell Biol 3(9):697–702 23. Bryant DM, Stow JL (2005) Nuclear translocation of cell-surface receptors: lessons from fibroblast growth factor. Traffic 6:947–954 24. Zwaenepoel O, Tzenaki N, Vergetaki A, Makrigiannakis A, Vanhaesebroeck B, Papakonstanti EA (2012) Functional CSF-1 receptors are located at the nuclear envelope and activated via the p110δ isoform of PI 3-kinase. FASEB J 26:691–706 25. Wanner G, Mayer C, Kehlbach R, Rodemann HP, Dittmann K (2008) Activation of protein kinase Cepsilon stimulates DNArepair via epidermal growth factor receptor nuclear accumulation. Radiother Oncol 86:383–390 26. Scholl SM, Pallud C, Beuvon F, Hacene K, Stanley ER, Rohrschneider L, Tang R, Pouillart P, Lidereau R (1994) Anti-colonystimulating factor-1 antibody staining in primary breast adenocarcinomas correlates with marked inflammatory cell infiltrates and prognosis. J Natl Cancer Inst 86(2):120–126 27. Lin H, Lee E, Hestir K, Leo C, Huang M, Bosch E et al (2008) Discovery of a cytokine and its receptor by functional screening of the extracellular proteome. Science 320:807–811 28. Mancini A, Koch A, Whetton AD, Tamura T (2004) The M-CSF receptor substrate and interacting protein FMIP is governed in its subcellular localization by protein kinase C-mediated phosphorylation, and thereby potentiates M-CSF-mediated differentiation. Oncogene 23:6581–6589 29. Morandi A, Barbetti V, Riverso M, Dello Sbarba P, Rovida E (2011) The colony-stimulating factor-1 (CSF-1) receptor sustains ERK1/2 activation and proliferation in breast cancer cell lines. PLoS One 6:e27450 30. Hammes LS, Tekmal RR, Naud P, Edelweiss MI, Kirma N, Valente PT, Syrjänen KJ, Cunha-Filho JS (2008) Up-regulation of VEGF, c-fms and COX-2 expression correlates with severity of cervical cancer precursor (CIN) lesions and invasive disease. Gynecol Oncol 110(3):445–451. doi:10.1016/j.ygyno.2008.04.038 31. Li NF, Kocher HM, Salako MA, Obermueller E, Sandle J, Balkwill F (2009) A novel function of colony-stimulating factor 1 receptor in hTERT immortalization of human epithelial cells. Oncogene 28:773–780 32. Dittmann K, Mayer C, Rodemann HP (2010) Nuclear EGFR as novel therapeutic target: insights into nuclear translocation and function. Strahlenther Onkol 186(1):1–6. doi:10.1007/ s00066-009-2026-4