CCN3 as a moonlighting signaling protein: a potential model for dual nuclear and extracellular functions across the CCN family

Abstract

Originally identified in the 1990s as an extracellular matrix (ECM) protein, Cellular Communication Network factor 3 (CCN3) was recognized as a member of a structurally related family of six proteins sharing a conserved tetramodular organization and broad regulatory functions in biological processes such as cell proliferation, attachment, migration, differentiation, wound healing, and angiogenesis, as well as in pathological conditions including fibrosis and tumorigenesis. Among these proteins, CCN3 rapidly emerged as the first tumor-suppressive member capable of negatively regulating cell growth in both normal and pathological contexts. A conceptual shift emerged following observations that, in addition to full-length CCN3 detected in the extracellular matrix, amino-truncated CCN3 proteoforms are targeted to the nucleus. In addition to physically interacting with the Rpb7 subunit of RNA polymerase II (Pol II), CCN3 binds to the promoter region of PAI-2 and exhibits transcriptional regulatory activity both in vitro and ex vivo. The dual localization and functional versatility of CCN3 led to the proposal that this protein may function as a “moonlighting” factor capable of integrating extracellular cues with nuclear transcriptional regulation. This conceptual framework, however, has remained relatively underexplored. The nuclear localization of other family members, together with the more recent observation that nuclear CCN6 physically interacts with RNA polymerase II, may extend this concept to the entire CCN protein family. By revisiting and integrating a substantial body of published data that has remained insufficiently acknowledged, this review aims to underscore the biological significance of CCN proteoforms in the nuclei of tumor cells. The demonstration of physical interactions between CCN3 and components of the central transcriptional machinery invites reconsideration of the prevailing ECM-centered view of CCN proteins and opens new perspectives on their biological roles. 

Graphical Abstract

Panel A: Dual biological effects of CCN3. Expression of full-length CCN3 via pRSV-recombinant vectors in chicken embryo fibroblasts (CEF) leads to CCN3 secretion and marked inhibition of cell growth. The transforming potential of amino-truncated CCN3 species, originally observed in nephroblastoma, was reproduced ex vivo by infecting CEF with a pRSV vector carrying truncated CCN3 DNA cloned from the same tumor. Panel B: Combinatorial events of CCN3 modules. The secreted CCN3 proteins exhibit combinatorial biological activities through their constitutive modules. Each module can perform specific functions individually, or together they can generate new functions present in the full-length protein, highlighting the modular “whole in one” concept. Panel C: Pre-initiation Pol II complex. The 12-subunit RNA Pol II pre-initiation complex (PIC) includes a protruding rpb4/rpb7 heterodimer that stabilizes the complex and extends its capacity to interact with transcription factors. Two-hybrid assays suggest CCN3 physically contacts Pol II at this level. Binding of the CCN3 C-terminal (CT) domain to the PAI-2 promoter supports a role for CCN3 in regulating PIC progression and its interaction with TFIID/TBP family regulators. Panel D: CCN5 as a dominant regulator. Some transcription factors require dimerization to regulate target genes. CCN proteins, with CT and VWC domains, can dimerize or multimerize to modulate transcription positively or negatively. CCN5, which lacks the CT domain, may still dimerize thanks to the VWC domain, and could act as a dominant negative regulator when CT-mediated dimerization is essential for promoter binding and transcriptional regulation.

Introduction

The CCN family of proteins comprises six homologous regulatory molecules involved in fundamental biological processes, including cell proliferation, adhesion, migration, differentiation, wound repair, angiogenesis, and tissue remodeling, as well as in pathological conditions such as fibrosis and cancer [1, 2]. The acronym “CCN” was originally proposed to unify the nomenclature of the founding members CYR61, CTGF, and NOV [5] and was later extended to include WISP1, WISP2, and WISP3 [1]. In 2018, the HUGO Gene Nomenclature Committee adopted the standardized terminology CCN1–CCN6 originally proposed at the first International Workshop on the CCN Family of Genes [1, 3] and recommended the use of CCN for Cellular Communication Network factors 1 through 6 [4].

Five of the six CCN proteins share a conserved tetramodular organization composed of structural domains commonly found in regulatory and signaling proteins [1, 3]. This modular architecture has long supported the view that CCN proteins primarily function as extracellular matrix-associated signaling molecules coordinating cell–cell and cell–matrix communication [1, 5, 6, 7, 8, 9, 10, 11]. CCN5 represents a structural exception, as it lacks the conserved C-terminal domain present in other family members [8], yet retains distinctive biological activities and has been reported to localize within the nucleus in certain cellular contexts [12].

Studies performed in CCN3-transfected insect or mouse cells demonstrated that both full-length and truncated proteoforms were detectable in cytoplasmic and extracellular fractions and were subjected to post-translational glycosylation [13, 14]. Microsequencing showed that the amino-truncated CCN3 corresponded to proteolytically processed molecules containing only the two carboxy-proximal domains (14). Similarly, biologically active truncated CCN2 proteoforms identified in pig uterine luminal flushings were composed of different combinations of carboxy-proximal domains [15].

Both full-length and amino-truncated CCN3 proteoforms were also detected in the cytoplasm and conditioned medium of cancer cells derived from H295 human adrenocarcinoma [16, 17], G59 glioblastoma cells [17, 18], melanomas [19], and choriocarcinoma [20].

CCN3 has attracted particular attention because of observations suggesting functions extending beyond extracellular signaling. Early reports describing the presence of CCN3 in the nuclei of HeLa cells were initially met with skepticism and attributed to potential antibody cross-reactivity [21]. However, subsequent studies using multiple experimental approaches confirmed the specificity of these findings [22] and established that CCN3 can indeed be detected within the nuclear compartment of other cancer cells derived from melanomas and choriocarcinoma [19, 20].

These observations suggest that CCN proteins, previously regarded primarily as extracellular regulators, may also participate in intracellular processes linked to transcriptional regulation. The observation that CCN3 can physically interact with the Rpb7 subunit of RNA polymerase II [22] provided a molecular rationale for this possibility and raised the hypothesis that CCN3 could be associated with nuclear regulatory complexes.

Since the initial reports published more than two decades ago, a growing body of experimental evidence has accumulated in support of nuclear localization and transcription-related activities of CCN3 [23, 24]. Nevertheless, most functional studies of CCN proteins remain centered on extracellular matrix-associated roles [6, 7, 25–28] and the potential contribution of nuclear proteoforms has been less frequently integrated into the overall conceptual framework of CCN biology.

The possibility that CCN3 exerts dual roles, in extracellular signaling and in nuclear regulatory processes, raises important questions regarding the mechanisms governing its subcellular distribution, the generation of distinct proteoforms, and their functional significance. Observations reported for other CCN family members, including CCN6 and CCN2, further suggest that nuclear localization and transcription activities may represent a broader, potentially conserved property of this protein family [29–32].

In this context, the present review revisits a collection of data produced since the late 1990 s, integrating earlier experimental observations with subsequent molecular findings. By assembling these results into a coherent framework, we aim to clarify how CCN3 and other CCN proteforms may contribute to both extracellular signaling pathways and intracellular regulatory mechanisms associated with gene expression.

Experimental observations

The CCN3 protein exists as multiple proteoforms with distinct subcellular localizations and biological activities. Analyses of MAV1-induced nephroblastomas [1, 5, 33, 34] revealed that viral integration disrupted CCN3, producing a truncated mRNA lacking sequences encoding the amino-terminal signal peptide [5] (Fig. 1). The resulting amino-truncated protein was associated with anchorage-independent growth, whereas full-length CCN3 showed little or no transforming activity [5] Fig. 1.

Fig. 1.

Oncogenic activation of CCN3 by MAV1 insertion. Panel A Modular structure of the chicken ccn3 gene. The wild-type chicken ccn3 gene is composed of five exons spanning approximately 10 kb of DNA. Exons 2–5 respectively, the four domains (IGFBP, VWC, TSP1 and CT) that are constitutive of the CCN3 protein. Panel B MAV-1 insertional mutagenesis of ccn3 sequences in chicken nephroblastoma. The integration of MAV-1 sequences into the second cellular ccn3 exon disrupts the coding sequence and leads to the production of a chimeric MAV-ccn3 mRNA, whose elevated expression is likely driven by the MAV LTR U3 promoter. Panel C Organization of the CCN3 protein encoded by the disrupted ccn3 gene The chimeric MAV-ccn3 mRNA expressed after MAV-1 insertion encodes an amino-terminally truncated CCN3 protein lacking the signal peptide and CCN3 domain 1. The first ATG initiation codon of the truncated polypeptide is located within domain 2 (VWC). Panel D The amino-terminally truncated CCN3 is oncogenic. Infection of chicken embryo fibroblasts with pRSV recombinant vectors (see text for details) demonstrates that expression of the full-length cellular CCN3 leads to cell growth inhibition, whereas expression of the amino-terminally truncated CCN3 produced by MAV-1 integration induces morphological cell transformation. Cells expressing the amino-terminally truncated CCN3 isoform also acquire the capacity to grow in an anchorage-independent manner and form large colonies when seeded in soft agar (see text)

Immunodetection studies in human tumor cell lines identified two CCN3 species: a 51 kDa full-length protein and a 30 kDa truncated variant [1, 18]. Domain-specific antibodies established that the 30 kDa form contains only the carboxy-proximal modules [35], indicating that post-transcriptional or post-translational events generate structurally distinct proteoforms.

A nuclear CCN3 species, corresponding in size to the amino-truncated variant, was consistently detected in HeLa, osteosarcoma 143, and other tumor cell lines, whereas full-length CCN3 was primarily cytoplasmic or secreted. Confocal immunostaining showed that nuclear CCN3 colocalizes with the ICP4 transcription factor [36] in HSV1-infected cells [22], but not with the replication factor ICP8 [37]. Ultrastructural immunogold analysis in H295R cells confirmed nuclear localization, with labeling detected near nuclear pores [23].

Yeast two-hybrid screening using full-length CCN3 as bait identified several interacting proteins, including integrins and regulatory partners such as the Rpb7 subunit of RNA polymerase II [22, 38, 39]. Pull-down assays supported the physical association between CCN3 and Rpb7, suggesting a possible link between CCN3 and transcriptional regulatory complexes [14].

Mechanistic studies indicated that the C-terminal (CT) domain is necessary and sufficient for nuclear targeting [24]. Recombinant CCN3 constructs containing the CT domain but lacking the signal peptide localized to the nucleus, whereas constructs lacking the CT domain did not [24]. A putative nuclear localization signal within the CT region [PTDKKGKKCLRTKKSLKAIHLQFK] directed GFP fusion proteins to the nucleus, supporting a role for this domain in nuclear import.

Functional assays further suggested that the CT domain can modulate transcriptional activity [24]. Gal4 fusion experiments in BHK21 cells showed that constructs containing the CT region reduced luciferase reporter activity, whereas constructs lacking this domain had no detectable effect. CCN3-CT was also reported to bind the PAI-2 promoter [40], supporting the possibility of direct involvement in gene regulatory processes.

In tumor samples and tumor-derived cell lines, nuclear CCN3 was associated with proliferative phenotypes. In melanoma and choriocarcinoma models, amino-truncated nuclear species predominated in highly proliferative cells, whereas full-length CCN3 was more frequently associated with less aggressive phenotypes [19, 20]. Induction of connexin 43 in Jeg3 cells shifted CCN3 localization toward the cell periphery and was accompanied by reduced proliferation, suggesting that cellular context can influence the balance between secreted and nuclear CCN3 forms [20].

Taken together, these observations are consistent with a model in which CCN3 exists as distinct proteoforms with localization-dependent activities. Secreted full-length CCN3 is generally associated with growth-inhibitory contexts, whereas nuclear truncated variants are linked to transcription functions and increased proliferative capacity. The CT module, which is known to be essential for the antiproliferative activity of CCN3 [41], appears central to nuclear targeting and regulatory activity. The interaction with RNA polymerase II components suggests a possible association of CCN3 with the transcriptional environment.

Current summary, discussion and perspectives

A coherent body of experimental observations supports the existence of distinct CCN3 proteoforms with different intracellular localizations and biological activities. The amino-truncated nuclear CCN3 species initially identified in virally induced tumors lacks the signal peptide required for secretion and is associated with proliferative and transforming properties. In contrast, full-length CCN3 is secreted or membrane-associated and has been repeatedly linked to growth-inhibitory and tumor-suppressive contexts.

Several complementary lines of evidence support the presence of CCN3 within the nuclear compartment. Nuclear localization of truncated CCN3 species has been demonstrated by immunodetection and ultrastructural analyses, including immunogold labeling near nuclear pores [23]. Full-length CCN3 was shown to interact with the Rpb7 subunit of RNA polymerase II, and nuclear CCN3 colocalizes with transcription-associated proteins in infected cells [22]. In addition, the C-terminal domain contains a functional nuclear localization signal and appears necessary and sufficient to direct amino-truncated CCN3 to the nucleus.

Functional assays indicate that this C-terminal region can influence transcriptional activity and can associate with promoter regions such as that of PAI-2 [40]. Together, these observations suggest that CCN3 proteoforms may participate, directly or indirectly, in processes associated with transcriptional regulation. While the precise mechanisms remain to be clarified, the available data support the concept that CCN3 is not restricted to extracellular signaling roles but may also act within intracellular regulatory environments.

Several important questions remain open. The mechanisms responsible for generating bioactive CCN3 proteoforms in vivo are still incompletely understood. Both proteolytic processing and alternative splicing have been proposed as potential contributors, but the exact pathways leading to the production of nuclear truncated species have not yet been established. Likewise, the precise domains mediating interactions between CCN3 and components of the transcriptional machinery, including RNA polymerase II, remain to be mapped in detail.

The identification of transcriptional targets of nuclear CCN3 represents another major area for future investigation. Binding of the CT domain to the PAI-2 promoter provides an initial indication that CCN3 may influence gene expression directly. Broader genomic and proteomic approaches will be required to determine whether this represents a general mechanism and to identify additional target genes.

The presence of nuclear forms is not restricted to CCN3. Nuclear localization and transcription functions have also been reported for other members of the CCN family. In particular, CCN6 has been shown to associate with RNA polymerase II in chondrocytes and during development, and to influence transcription of genes involved in mitochondrial function [29]. Nuclear localization of CCN2 has also been described in certain cellular contexts, where it may influence cell-cycle regulation and transcriptional responses [31, 32].

Taken together, these observations support a broader model in which CCN proteins can operate in distinct cellular compartments through structurally diverse proteoforms. Among the several MMP proteases (MMP-1, −3, −7, −8, −9) reported to cleave CCN proteins, MMP14 processes CCN3 in vitro at the level of the variable hinge region separating the amino-proximal IGFBP-VWC and carboxy-proximal TSP-CT domains [42]. Further proteomic studies should establish whether in vivo post-translational proteolysis of CCN proteins contributes to controlling the balance between secreted and nuclear variants. Disruption of this balance, as observed in tumor cells expressing amino-truncated nuclear forms, may alter transcriptional programs and contribute to disease progression via the regulation of proliferation, differentiation, and cellular homeostasis.

Despite the progress achieved over the past decades, several fundamental issues remain unresolved. The biochemical mechanisms governing nuclear targeting, the identity of protein partners involved in these processes, and the full spectrum of transcriptional targets are still largely unknown. Understanding the physical basis of CCN protein interactions with RNA polymerase II remains a key open question. Addressing these questions will require the integration of modern approaches, including high-resolution proteomics, proximity-based interaction mapping, and global transcriptomic analyses.

In this context, CCN3 provides a useful model for exploring how structurally distinct CCN3 proteoforms can operate in multiple cellular compartments and participate in distinct signaling and regulatory pathways. A better understanding of these mechanisms should help clarify the biological significance of nuclear CCN proteins and their potential roles in development and disease.

Author’s contributions

B. Perbal wrote the manuscript. No conflict of interest.

Funding

None.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.