TACC3: a multi-functional protein promoting cancer cell survival and aggressiveness

ABSTRACT

TACC3 is the most oncogenic member of the transforming acidic coiled-coil domain-containing protein (TACC) family. It is one of the major recruitment factors of distinct multi-protein complexes. TACC3 is localized to spindles, centrosomes, and nucleus, and regulates key oncogenic processes, including cell proliferation, migration, invasion, and stemness. Recently, TACC3 inhibition has been identified as a vulnerability in highly aggressive cancers, such as cancers with centrosome amplification (CA). TACC3 has spatiotemporal functions throughout the cell cycle; therefore, targeting TACC3 causes cell death in mitosis and interphase in cancer cells with CA. In the clinics, TACC3 is highly expressed and associated with worse survival in multiple cancers. Furthermore, TACC3 is a part of one of the most common fusions of FGFR, FGFR3-TACC3 and is important for the oncogenicity of the fusion. A detailed understanding of the regulation of TACC3 expression, its key partners, and molecular functions in cancer cells is vital for uncovering the most vulnerable tumors and maximizing the therapeutic potential of targeting this highly oncogenic protein. In this review, we summarize the established and emerging interactors and spatiotemporal functions of TACC3 in cancer cells, discuss the potential of TACC3 as a biomarker in cancer, and therapeutic potential of its inhibition.

Introduction

Members of the transforming acidic coiled-coil domain-containing protein (TACC) family are highly acidic structural proteins that are present in different organisms, from yeast to mammals [1]. They are vital for the regulation of microtubule stability and dynamics, and centrosome integrity. They are characterized by a conserved coiled-coil domain at the C-terminal, known as the TACC domain that is approximately 200 amino acids long and is necessary for binding to microtubules and centrosomes [2]. The TACC domain comprises the majority of the TACC proteins in C. elegans (i.e. two of the three exons), while in higher organisms, from D. melanogaster to human, the TACC domain is encoded by the final exons of the gene [1]. Vertebrates express three TACC genes: TACC1, TACC2, and TACC3. The interaction of TACC proteins with microtubules and centrosomes is mediated solely by the TACC domain, hence making the microtubule and spindle functions of TACC proteins the ancient, conserved functions of the TACC family. However, there are also major differences among family members as they show relatively little homology outside the TACC domain [1]. For instance, while TACC1 is a substrate of Aurora C [3], TACC3 is phosphorylated by Aurora A [4]. On the other hand, TACC2 is phosphorylated by TTK [5] but not by Aurora A or C, although it binds to Aurora C [6].

TACC1 and TACC2 proteins have serine/proline-rich (SPAZ) motif at their central region [7] (Figure 1). The SPAZ motif was suggested to be potentially important for kinase recognition and is possibly a new member of the immunoglobulin superfamily (IgSF), providing an interface for protein-binding. One of the interacting partners of TACC1, TRAP was found to bind to the region of TACC1 containing the SPAZ motif [8]; however, the functional significance of the SPAZ motif has been mostly unknown. The greatest variability in size and sequence among the vertebrate TACC proteins is observed for TACC3 at its central region, causing a total protein size range of 599 to 942 amino acids from rat to zebrafish [1]. Human TACC3 protein consists of 838 amino acids with an N-terminal region conserved in higher eukaryotes. The TACC1 and TACC2 proteins contain nuclear localization signal (NLS), while TACC3 does not [7] (Figure 1). Although there is still no evidence showing direct binding of TACC proteins to DNA, all members of the family can be localized both to cytosol and nucleus [2,9] depending on the cell type or the cell cycle phase. The lack of NLS in TACC3 protein despite a nuclear localization suggests that TACC3 localization to nucleus is likely mediated by its interaction with other proteins containing an NLS sequence. Indeed, TACC3 is known to interact with several nuclear proteins [10,11], as described below that could facilitate its nuclear import in various conditions. Furthermore, TACC3 localization to nucleus might also be in part mediated by its interaction with a nuclear envelope protein, TSC2 [12,13]. Thus, the lack of NLS or the SPAZ motif at the central region might necessitate TACC3 to acquire unique interactors/functions during evolution, such as its interaction with the nuclear HIF-1β (ARNT) or the NuRD chromatin remodeling complex members. Although most of these interactions occur at the conserved TACC domain, the mouse and human TACC3 are known to harbor two functionally distinct modules, CC1 and CC2 [14], compared to TACC1 and TACC2 proteins. These modules uniquely control the interaction of TACC3 with its partners as discussed below in detail. Overall, it is becoming apparent that the TACC proteins, especially TACC3 might have numerous non-canonical functions, in addition to their ancestral spindle/centrosome functions, such as chromatin remodeling and transcription regulation via acting as adaptors, i.e. bridging proteins important for the formation of different protein complexes [1].

Figure 1.

Domains of TACC family proteins in human and mouse. The domains are color coded, and their names are provided at the bottom. This figure is generated in Biorender.com.

Among the vertebrate TACC orthologs, TACC1 is expressed from the centromeric region of the breast cancer amplicon on chromosome 8p11 [15], while TACC2 has been mapped to 10q26 [15,16], also the site of another breast cancer amplicon [17]. TACC3 gene localizes to 4p16, a region disrupted in various cancers, such as bladder [18] and multiple myeloma [16]. TACC proteins are widely expressed in several tissues throughout the developmental stages. Among the three members, TACC3 is the member that is expressed the most in highly proliferating and differentiating cell types and tissues, such as the testis, spleen, thymus, peripheral blood leukocytes, and hematopoietic stem cells [19,20]. It has key roles during cell division, thus making TACC3-deficient mice embryonically lethal with reduced cell number, increased apoptosis, and mitotic defects [19].

TACC family members are frequently deregulated in cancer. TACC1 protein was initially found to be downregulated in around one-third of breast tumors, but no correlation with the clinical parameters was observed [21]. Furthermore, the number of normal breast samples was very limited in this study with lack of quantitative data that could explain the observed downregulation in tumors despite the localization of TACC1 gene to the breast cancer amplicon on chromosome 8p11. Indeed, TACC1 was later shown to accelerate mammary tumor formation in mice via activating Ras and PI3K pathways [22], suggesting an oncogenic function. TACC2-deficient mice are shown to be viable and develop normally, and TACC2 deficiency did not lead to an increased incidence of tumor development [23] suggesting that TACC2 is not a tumor suppressor in contrast to what has been proposed initially [8,24]. Furthermore, subsequent studies showed that increased TACC2 expression correlates with poor prognosis in patients with breast cancer and promotes the growth of the castration-resistant prostate cancer [25,26]. TACC3 is the most widely studied and best-characterized oncogenic family member. It has been shown to be overexpressed and have prognostic value in different cancer types, including breast [27], prostate [28], colorectal [29], and gastric cancers [30]. It is also the key player in driving the survival of the highly aggressive cancer cells, such as cancer cells with centrosome amplification (CA) [11]. It is a multifunctional protein that evolutionary gained unique functions in addition to its ancestral microtubule/centrosome-associated roles by interacting with unique proteins/protein complexes. This broad spectrum of interactors offers TACC3 pivotal roles in many aspects of cancer biology, including enhanced mitotic progression, cell proliferation, stemness, and migration/invasion. Thus, a detailed understanding of how TACC3 expression is regulated in cancer, its key partners and molecular functions of TACC3-centered protein complexes in a cell cycle- and cellular localization-dependent manner would allow us better understand how TACC3 drives these oncogenic processes that would further unleash novel treatment strategies. In this review, we discuss the (de-) regulation of TACC3 expression in cancer, its canonical and non-canonical oncogenic partners, and the spatiotemporal functions of each of the TACC3-bearing complexes in different cancer-associated processes. We summarize the types of cancers where TACC3 is overexpressed and associated with worse clinical outcome, thus likely has a biomarker potential, and discuss the potential therapeutic benefits of targeting TACC3.

Regulation of TACC3 expression in cancer

TACC3 expression can be regulated both at the transcriptional and post-transcriptional levels in cancer. In colorectal cancer tissues, TACC3 was found to be correlated with E2F1 at the mRNA and protein levels [31]. E2F1 knockdown reduces TACC3 mRNA and direct binding of E2F1 to TACC3 promoter region was demonstrated by chromatin immunoprecipitation (ChIP) assay [31]. Interestingly, TACC3 was shown to be essential for the transcriptional activity of E2F1 via binding to the E2F1 promoter region in a feed-forward loop (Figure 2). We have recently demonstrated that TACC3 is overexpressed in tumors with p53 loss/mutations or with CA, the Achilles’ heel of cancer [11,32]. We showed that the p53-repressed transcription factor FOXM1 drives TACC3 transcription by directly binding to TACC3 promoter in breast cancer cells with p53 loss/mutations and/or CA [11]. TACC3 was also shown to be stabilized at mRNA and protein levels upon activation of the hypoxia-regulated transcription factor, HIF1-α; however, no direct binding of HIF1-α to TACC3 promoter has yet been reported [33]. TACC3 is also in a feedforward loop with HIF-1 complex via interacting with HIF-1β (ARNT) as described in detail below. At the epigenetic level, a histone methyltransferase, WHSC1 has been shown to transcriptionally activate TACC3 expression by binding to TACC3 promoter and inducing H3K36me2 modification [34]. In addition to regulation at the transcriptional level, TACC3 expression can also be regulated at the post-transcriptional level. In one of the first studies showing the role of the m⁶A demethylase, ALKBH5 in self-renewal, TACC3 was identified as a functionally important, direct target of ALKBH5 in acute myeloid leukemia (AML) cells [35]. ALKBH5 knockdown increases the m⁶A abundance on TACC3 mRNA, causing reduced mRNA stability and loss of expression. The functional importance of TACC3 as the downstream target of ALKBH5 was demonstrated by reduced colony formation, cell growth and increased apoptosis upon decreased MYC and increased p21 levels following TACC3 inhibition in AML cells [35]. Another post-transcriptional regulation of TACC3 involves microRNAs (miRNAs), which are small non-coding RNAs repressing the expression of their target genes [36]. miR-425 was found to be downregulated in multiple myeloma, leading to upregulation of its target, TACC3 [37]. Furthermore, global network analysis of the TACC3-associated miRNAs in glioma revealed negative correlation of 10 miRNAs with the levels of the TACC3 mRNA, which has predicted binding sites for these miRNAs at its 3’-UTR [38]. Importantly, some of these miRNAs were shown to be tumor suppressors in various cancers, suggesting that TACC3 expression might be upregulated at the post-transcriptional level in tumors upon downregulation of these tumor suppressor miRNAs. The lack of the 3’-UTR region of FGFR3 due to truncation of the last exon in the FGFR3-TACC3 fusion, which is one of the most common fusions of FGFR in cancer [39] leads to stronger expression of the fusion gene in glioma, since the FGFR3-targeting miRNA, miR-99a can no longer negatively regulate the fused FGFR3 [40]. Overall, the expression of TACC3 or TACC3-bearing fusions are regulated both at transcriptional and post-transcriptional levels in cancer cells (Figure 2).

Figure 2.

Multiple ways of regulation of TACC3 expression in cancer cells. E2F1 transcribes TACC3, and TACC3 is in a feed-forward loop with E2F1 via promoting its activity (i). In cancer cells with centrosome amplification and/or p53 loss of mutations, the p53-suppressed FOXM1 is activated and transcribes TACC3 (ii). HIF-1α drives TACC3 expression under hypoxia, and TACC3 in turn facilitates HIF-1-mediated transcription (iii). WHSC1 transcribes TACC3 via inducing H3K36me2 modification (iv). In addition to transcriptional regulation, the m⁶A demethylase, ALKBH5 stabilizes TACC3 mRNA via inhibiting m⁶A methylation of TACC3 mRNA (v). Also, miR-425 degrades TACC3 mRNA, while FGFR3 within the FGFR3-TACC3 fusion mRNA escapes from miR-99a-mediated degradation (vi). This figure is generated in Biorender.com.

Spatiotemporal partners of TACC3 and its oncogenic functions

Spindle functions of TACC3

TACC3 recruitment on to the spindle

TACC3 localizes to mitotic spindles in a complex with the microtubule polymerase, chTOG, and the vesicle-coating protein, clathrin. The complex is formed on spindle microtubules between the poles and chromosomes (i.e. the kinetochore-fibers (K-fibers)) [2,41,42], and its major function is to crosslink the K-fibers which provide mechanical strength and stability to the fibers, thus facilitating mitotic progression. It was demonstrated that enhanced loading of this complex which can be achieved by overexpressing TACC3 has led to increased numbers of microtubules per fiber and the cross-sectional area occupied by these microtubules, and enhanced interconnectivity, suggesting tighter local packing [43]. Similarly, loss of TACC3 was shown to reduce the number of inter-microtubule bridges between adjacent microtubules within the K-fibers in cancer cells [44].

There are two different hypotheses on the formation and spindle recruitment of the TACC3/clathrin/chTOG complex. One hypothesis that was put forward by three independent research groups in 2010 supports a clathrin-centric model where clathrin recruits TACC3 to mitotic spindles which then recruits chTOG [45–47]. The hypothesis is based on the observation that depleting clathrin reduces the spindle recruitment of TACC3 and chTOG while depleting TACC3 inhibits the spindle association of only chTOG but not clathrin, suggesting that clathrin is potentially responsible for spindle recruitment of the complex. It was later supported by additional studies demonstrating similar findings [48]. The second model argues that TACC3 is the primary recruitment factor, since clathrin does not directly bind to microtubules but it is tethered to spindles via auxiliary proteins, e.g. B-Myb [49] and GAK [50]; unlike TACC3, which is predominantly localized to spindles. This hypothesis was supported by the observation that depleting either TACC3 or chTOG reduced clathrin recruitment to the spindle while clathrin or chTOG depletion caused only a minor reduction in TACC3 spindle recruitment [44].

Later on, a composite microtubule-binding interface between TACC3 and clathrin has been identified [51], suggesting that the TACC3-clathrin subcomplex has the strongest affinity to spindles than either of the two proteins in their free form. Likewise, chTOG itself has a high binding affinity for tubulin [52], yet it still requires TACC3 and possibly clathrin to accumulate onto spindles. Therefore, it seems likely that the formation and spindle recruitment of this complex could be even more complex than what was proposed by the two models. Instead, it is likely that the formation of the subcomplexes depends on multiple regulatory factors and potentially multiple domains within the TACC3 protein that orchestrate the localization and stability of the complex in a harmony as discussed below.

TACC3-mediated K-fiber crosslinking and spindle stability

Two functionally distinct subdomains have been identified within the TACC domain of human and mouse TACC3, CC1 (amino acid (aa) 414–530 for mouse) and CC2 (aa 530–630 for mouse). Upon interaction of the C terminus of chTOG (aa 1806–2032 for mouse) with the CC1 domain, the intradomain interaction of CC2 with the central repeat region is disrupted, giving rise to the effector-bound state (i.e. unmasked state) of TACC3 that allows subsequent binding of other interactors. The binding of TACC3 to clathrin occurs at the clathrin-interaction domain (CID, aa 522–577 for human) of TACC3 and is dependent on the Aurora A-mediated phosphorylation at S558 and possibly at S552 [45,47,53], while interaction with chTOG is independent of TACC3 phosphorylation [14]. Inhibition of Aurora A kinase or overexpressing a truncated version of TACC3 with deleted CID causes loss of TACC3 phosphorylation and reduces its spindle localization [53]. Nevertheless, there were still significant levels of TACC3 on microtubules even in the absence of the CID domain, most likely through the TACC domain [53]. A class II PI3 kinase, PI3K-C2a was found to be an integral part of the TACC3/chTOG/clathrin complex at metaphase, interacting with both TACC3 and clathrin and crosslinking K-fibers in breast cancer cells independent of its kinase activity [54]. Another ancillary member of the TACC3/chTOG/clathrin complex in cancer cells is GTSE1, a microtubule plus-end tracking protein [55], which has been reported to play important roles in cancer, e.g. cell proliferation, migration [56] and drug resistance [57]. Formation of the GTSE1-bearing complex on the spindles is dependent on Aurora A-mediated TACC3 phosphorylation and is mutually exclusive with the PI3K-C2a-containing complex [46]. Sorting Nexin 9 (SNX9) was identified as a novel partner of clathrin, and its depletion reduced clathrin binding to spindles during metaphase which further led to reduced TACC3 recruitment in cancer cells [48]. However, a potential interaction between TACC3 and SNX9 has not yet been shown.

Overall, TACC3 is localized on to the spindles in complex with several partners and mediates K-fiber crosslinking and spindle stability (Figure 3). Nevertheless, further research is needed to get a more comprehensive view of the nature of this complex, the main and the axillary elements, their recruitment to spindles, and the functions of each member on the formation, stability, and crosslinking function of the complex. Furthermore, even if the Aurora A-mediated phosphorylation is critical for clathrin and GTSE1 binding and the efficient spindle localization of the complex, the formation of the Aurora A-independent TACC3-chTOG subcomplex could be as important as the former, as it acts as a priming factor, unmasking the TACC domain allowing for binding to other interaction partners.

Figure 3.

The functions of TACC3 and its interaction partners on the spindles. Along the microtubules, TACC3, which is downstream of Aurora A, and in complex with clathrin (CLTC)/chTOG/PI3K-C2a (i); CLTC/chTOG/GTSE1 (ii), or CLTC/SNX9 (iii) regulates K-fiber crosslinking and spindle stability. At the extreme plus ends of microtubules, TACC3, in complex with chTOG promotes microtubule polymerization. EB1 is localized at the microtubule ends, upstream of TACC3/chTOG (iv). This figure is generated in Biorender.com.

TACC3-mediated microtubule polymerization

In addition to localizing along the spindles to mediate K-fiber crosslinking and stability, TACC3 also localizes to the plus ends of the microtubules to promote microtubule polymerization (i.e. elongation). In embryonic neuronal cells, TACC3 depletion reduced both axonal number and axonal length, suggesting that TACC3 is required for microtubule polymerization and proper axonal outgrowth during development [58]. Further supporting the spatiotemporal functions of TACC3, a study demonstrated that during interphase, TACC3 localizes at the extreme microtubule plus ends in HeLa cancer cells and mediates microtubule polymerization in complex with chTOG [59]. Unlike the plus end binding protein EB1, which binds solely to growing microtubule plus ends, TACC3 is able to bind to paused and even shrinking microtubules, indicating that TACC3 recognizes a structural conformation that is present at both growing and shrinking plus ends [59]. Strikingly, in the same study, the plus end activity of TACC3 was shown to be potentially involved in cell migration in interphase cells. Overall, TACC3 has several key functions on the spindles by promoting K-fiber crosslinking, thus increasing microtubule stability, and also by triggering microtubule elongation/polymerization to facilitate mitotic progression (Figure 3).

Centrosomal functions of TACC3

TACC3 recruitment to the centrosomes

In addition to its roles in K-fiber stability and crosslinking, TACC3 that is in complex with chTOG and clathrin is also important for K-fiber formation through nucleation of centrosomal or acentrosomal microtubules at the onset of mitosis [60]. Although TACC3 is essential for the spindle localization of chTOG, depletion of TACC3 or overexpression of a defective mutant lacking chTOG binding domain only moderately affects centrosomal chTOG levels [61] but rather causes a diffuse pattern of chTOG around the vicinity of the mitotic centrosomes [62]. That could be in part due to compensation of other TACC family members [61]. Likewise, chTOG is also not required for the localization of TACC3 to centrosomes [61,62].

The Aurora A-mediated TACC3 phosphorylation was found to be critical for centrosome recruitment of TACC3 as inhibiting the kinase activity of Aurora A or overexpressing a non-phosphorylatable mutant of TACC3 results in less centrosomal TACC3 [4,63,64]. Depletion of clathrin whose interaction with TACC3 depends on Aurora A-mediated phosphorylation leads to loss of centrosomal TACC3 in Xenopus. However, the centrosome recruitment of human TACC3 is not affected by clathrin inhibition, similar to chTOG inhibition. This suggests that TACC3 could be the initiator for the localization of the complex on centrosomes in human cells [65].

Functions of TACC3 at the centrosomes

It was demonstrated that overexpressing TACC3 mutants defective of chTOG-binding stimulates recruitment of the γ-tubulin ring complex (γ-TuRC), a multi-protein complex mediating microtubule nucleation to centrosomes, leading to abnormally dense spindle microtubules [62]. However, depletion of the entire TACC3 protein reduces γ-TuRC proteins and increases the levels of γ-tubulin small complex (γ-TuSC), the building block of γ-TuRC, suggesting that TACC3 is potentially required for the assembly of γ-TuRC from γ-TuSC [65]. Depletion of TACC3 reduces the assembly of astral microtubules and impairs the microtubule nucleation at the centrosomes in both mitotic and interphase cancer cells [66]. In line with these findings, we have also demonstrated reduced TACC3 localization on centrosomes of both mitotic and interphase breast cancer cells upon treatment with the TACC3 inhibitor, BO-264 [67]. Likewise, mutating the S558 site of TACC3 to alanine (S558A) disrupts the localization of the γ-TuRC proteins at the spindle poles and reduces the astral microtubules in HeLa cells. As astral microtubules are essential for positioning the spindle poles during mitosis, cells expressing the S558A mutant TACC3 fail to position the spindle poles, thus exhibiting delays in metaphase-to-anaphase transition [68]. NDEL1, a serine oligopeptidase, was identified as another interactor of TACC3. NDEL1 phosphorylation by Aurora A precedes TACC3 phosphorylation and knocking out NDEL1 in mouse embryonic fibroblasts (MEFs) impairs TACC3 S347 levels, corresponding to S558 site in humans, as well as its centrosomal loading [69] (Figure 4).

Figure 4.

The functions of TACC3 and its interaction partners on the centrosomes. TACC3 controls spindle pole organization by interacting with clathrin (CLTC) and chTOG (i), and it also mediates spindle pole position (ii) under the control of Aurora A. TACC3 in complex with NDEL1 regulates centrosome maturation (iii), and TACC3 triggers microtubule nucleation by controlling the assembly of γ-TuRC from γ-TuSC (iv) under the control of Aurora A. In cancer cells with CA, TACC3 is critical for clustering extra centrosomes via interacting with KIFC1 and ILK (v). This figure is generated in Biorender.com.

Centrosomes are required to build bipolar spindles that is necessary for proper cell division. Centrosome amplification (CA) is one of the hallmarks of aggressive cancers [32] characterized by supernumerary (extra) centrosomes [70] and is strongly associated with tumor progression and worse prognosis in a variety of different cancers, e.g. breast, prostate, ovarian and lung [32]. CA can be induced by various causes, including centrosome overduplication by PLK4 overexpression, cytokinesis failure or genomic aberrations, such as p53 mutations [71] that are commonly observed in aggressive tumors, e.g. triple negative breast cancer (TNBC) [72]. Clustering of extra centrosomes is crucial for the survival of cancer cells with CA since it ensures proper cell division [73]. We have recently identified KIFC1, a nonessential kinesin motor protein as a novel interactor of TACC3 at the centrosomes in cancer cells bearing CA, mediating centrosome clustering [11]. Targeting TACC3 causes loss of interaction between TACC3 and KIFC1, leading to centrosome de-clustering and mitotic cell death in these highly aggressive cancer cells [11]. The integrin-linked kinase (ILK) was also found to make a complex with TACC3 and chTOG at the centrosomes of cancer cells with CA [74]. Its pharmacologic inhibition leads to multipolar spindle formation and cell death in cancer cells with CA whereas normal cells or non-CA cells exhibit little to no cell death. Overall, TACC3 plays several key functions at the centrosomes of cancer cells by promoting spindle pole positioning and organization, microtubule nucleation and maturation, and centrosome clustering (in case of cancer cells with CA) (Figure 4).

Nuclear functions of TACC3

Besides playing key roles at the spindles and centrosomes, TACC3 has non-canonical functions also in the nucleus. Interestingly, in one of the initial studies that first identified TACC3, it was found to be interacting with the Aryl hydrocarbon receptor nuclear translocator (ARNT, also known as HIF-1β) transcription factor [75]. Although it was first thought to be a member of the bHLH/PAS protein family, later it was realized that it is the third member of the TACC family, i.e. TACC3. HIF-1 is a dimeric protein complex sensing oxygen deprivation, i.e. hypoxia, and is composed of the HIF-1α subunits, dimerizing with ARNT, which allows them to bind to DNA and stimulate the transcription of the target genes [76]. Glutathione S-transferase (GST) pulldown experiments showed that the mouse Tacc3 specifically interacts with Arnt, Arnt2 and weakly with Sim2, another bHLH/PAS protein. Treatment of liver epithelial cells with iron-chelating hypoxia mimics, activating the HIF-1 pathway mediated by the Arnt/Hif-1α dimer in the presence of Tacc3 overexpression, leads to efficient transcriptional activity on the hypoxia response element (HRE) [75]. Human TACC3 was also shown to interact with the PAS-B domain of ARNT at its C-terminal TACC region [33,77], acting as a co-activator that is necessary for transcriptional responses to hypoxia. Overexpressing the TACC domain of TACC3 was enough to increase the HIF-1-dependent expression of its target genes, BNIP3, DEC2, and PGK. The increase in HIF1A transcriptional activity was also demonstrated by HRE reporter assay [33] (Figure 5). Hypoxia is a key feature of almost all solid tumors exhibiting rapid growth, and it is associated with the gain of aggressive traits, such as increased stiffness, epithelial-mesenchymal transition (EMT), and reprogramming of energy metabolism [78]. Thus, by stimulating the HIF-1-dependent expression of the oncogenic targets, TACC3 is likely acting as a key mediator of hypoxia-driven tumor progression in solid tumors.

Figure 5.

The functions of TACC3 and its interaction partners in nucleus. TACC3 in complex with ARNT triggers hypoxia-mediated gene transcription under hypoxia (i). TACC3 in complex with members of the NuRD complex, i.e. MBD2 and HDAC2 suppresses the transcription of tumor suppressors in cancer cells with CA (ii). TACC3 interacts with FOG-1 in the cytosol, and relieves the inhibitory effect on GATA-1, thereby activating transcription (iii). TACC3 is also in complex with the histone acetyl transferase, pCAF and MBD2 to activate gene transcription (iv). This figure is generated in Biorender.com.

Two independent research groups identified TACC3 as an interactor of FOG-1, the partner of the transcription factor, GATA-1 using yeast two-hybrid. One of the studies validated the interaction in the nucleus by performing pulldown assay in the nuclear extracts from HEK293T cells overexpressing tagged proteins [79]. The second study failed to detect TACC3 and FOG-1 in the nuclear extracts and thus proposed a cytoplasmic interaction [80]. According to the latter, the cytoplasmic interaction leads to sequestration of FOG-1 in the cytoplasm, thus reducing nuclear levels and the binding of FOG-1 to GATA-1. FOG-1 can either inhibit or potentiate GATA-1 activity depending on the context [81,82]. Overexpressing TACC3 relieves the inhibitory effect of FOG-1 on GATA-1-mediated transcription, supporting a model where TACC3 competes with FOG-1 for binding to GATA-1, thus inhibiting the cofactor function of FOG-1 for GATA-1 [80] (Figure 5). GATA family members regulate cancer cell survival and proliferation, thus mutations, loss of expression, or overexpression of GATA family members have been associated with various cancers, such as leukemia, colorectal and breast cancers [83]. By being a regulator of GATA-1-driven gene expression, TACC3 is likely the key for the growth and aggressiveness of these GATA-altered tumors.

TACC3 was also found to interact with the histone acetyltransferase (HAT), pCAF as well as the methyl-binding domain protein 2 (MBD2) at its TACC domain as shown by GST pull-down and co-immunoprecipitation experiments [10]. Overexpressing TACC3 reactivates the transcription at the methylated promoters. Mechanistically, it was found that pCAF binds to TACC3/MBD2 and the resulting complex increases HAT activity, thus leading to the transcription of target genes [10]. In cancer cells with CA, we recently demonstrated that TACC3 interacts with MBD2 and HDAC2, the two important members of the chromatin remodeling NuRD complex, via its TACC domain [11]. We showed that it further stimulates the transcriptional suppressor function of MBD2 and HDAC2, thus repressing tumor suppressor gene expression (Figure 5). Inhibiting TACC3 with BO-264 leads to the loss of interaction of TACC3 with the NuRD complex members, removing them from the chromatin, and thus activating the transcription of the tumor suppressors, leading to G1 arrest and apoptosis in breast cancer cells with CA [11]. Interestingly, the HDAC inhibitor, trichostatin (TSA) was shown to downregulate TACC3 mRNA in cholangiocarcinoma cells [84], suggesting that there might be a feedforward loop between TACC3 and HDACs. Overall, besides its canonical functions on spindles and centrosomes, TACC3 is also localized to nucleus and regulates gene transcription via interacting with transcription factors or chromatin remodelers (Figure 5).

TACC3 as a potential biomarker in cancer

TACC3 is overexpressed at mRNA or protein levels and associated with worse overall, disease-free or progression-free survival in a wide range of tumors, including breast [11,67], cervical [85], ovarian [86], non-small cell lung cancer (NSCLC) [87,88], gastric [30], renal [89] cancers, osteosarcoma [90], glioma [38], esophageal squamous cell carcinoma [91], colorectal [29], hepatocellular [92], and head and neck cancers [11] and lymphomas [93]. TACC3 expression was also shown to be upregulated specifically in aggressive tumors, such as tumors with CA [11] or in metastatic prostate cancer [28], and its high expression is associated with worse metastasis-free survival in osteosarcoma [90].

One of the most common fusions of FGFR in cancer is the FGFR3-TACC3 fusion, constituting approximately half of the FGFR gene fusions [39]. The FGFR3-TACC3 fusions have been detected in clinical tumors of various cancers including glioblastoma multiforme (GBM) [94], bladder urothelial tumors [95], NSCLC [96], cervical cancer [97] and TNBC [98]. It can be detected by FGFR3 immunopositivity [99] or by RNA-FISH in the FFPE samples [100]. The fusion-bearing tumors exhibit distinct histopathological and molecular features [99] and are enriched during acquired resistance to EGFR or MEK-targeting therapies [101–103]. The fusion is associated with a stronger risk of bladder cancer, especially for women [95]. In glioma, FGFR3-TACC3 fusion was found to be mutually exclusive with the IDH mutations [104] which are associated with better clinical outcome [105]. However, among IDH wildtype patients, presence of FGFR3-TACC3 fusion is associated with better overall survival [104], warranting more research to understand the roles of FGFR3-TACC3 fusions in predicting survival and therapy response in different cancers. Overall, high expression of TACC3 or the FGFR3-TACC3 fusion protein is associated with tumor aggressiveness and worse survival in many cancers, showing their biomarker potential (Table 1).

Table 1.

Clinical association of TACC3 alterations in different cancers.

Type of cancer	Clinical association	Level of association	Reference
Breast cancer	High TACC3 – associated with stage, grade and worse overall survival	Protein & mRNA	[11,67]
Prostate cancer	High TACC3 – associated with metastasis status, tumor stage, total prostate–specific antigen (PSA) level, and Gleason score, and worse disease–free survival	mRNA, protein	[26,67]
NSCLC	High TACC3 – associated with differentiation, stage, grade, histologic type, smoking status and worse recurrence–free and overall survival	mRNA and protein	[87,88,96]
Osteosarcoma	High TACC3 – associated with worse metastasis–free survival	Protein	[90]
Gastric cancer	High TACC3 – associated with extracapsular extension of the tumor, tumor relapse and worse overall survival	mRNA, protein	[30,67]
Lymphoma	High TACC3 – associated with worse overall survival	Protein	[93]
Cervical cancer	High TACC3 – associated with tumor histological type, nerve invasion, differentiation, parametrium invasion, Ki-67, overall, disease–free and recurrence–free survival	Protein	[85]
Renal cancer	High TACC3 – associated with worse overall survival	mRNA	[89]
Esophageal squamous cell carcinoma	High TACC3 – associated with differentiation, lymphoid nodal status and worse overall survival	Protein	[91]
Head and neck cancer	High TACC3 – associated with worse overall survival	mRNA	[11]
Colorectal cancer	High TACC3 – associated with clinical stage, T classification and M classification, worse overall and disease–free survival	Protein	[29]
Glioma	High TACC3 – associated with worse overall survival	mRNA	[38]
	FGFR3-TACC3 – associated with better overall survival	mRNA	[104]
Bladder cancer	FGFR3-TACC3 – associated witha stronger risk of bladder cancer, especially for women	Gene	[95]
Hepatocellular cancer	High TACC3 – associated with worse overall and disease–free survival	Protein	[92]

Therapeutic potential of inhibiting TACC3 in cancer

The potential of TACC3 as a therapeutic target in cancer

Given its overexpression in cancers, its key roles on spindles, at centrosomes, and in transcriptional regulation as discussed above, TACC3 depletion causes cell death irrespective of the p53 status and interferes with key oncogenic processes in cancer cells, leading to blockage of tumor growth and aggressiveness (Figure 6). In hepatocellular carcinoma, silencing TACC3 inhibits proliferation, clonogenicity, and cancer stem cell-like phenotype [92]. It also induces p21 expression, G1 arrest and triggers cell death in HCT-116 hepatocellular carcinoma cells in a p53- and p38-dependent manner [106]. In cholangiocarcinoma, TACC3 knockdown induces G2/M cycle arrest and suppresses the invasion, metastasis, and proliferation both in vitro and in vivo [84]. In cervical cancer cells, TACC3 was shown to promote proliferation, transforming capability, migration as well as the expression of EMT-related markers via activation of PI3K/AKT and ERK signaling pathways [107]. TACC3 inhibition is also effective in inducing multipolar spindle formation, mitotic arrest, and apoptosis in Burkitt lymphoma and T-cell acute lymphoblastic leukemia (T-ALL) [108], suggesting that TACC3 may be an attractive therapeutic target also in hematologic cancers. Furthermore, despite the lack of preclinical studies, TACC3 inhibition presumably has therapeutic potential also in hypoxic tumors or tumors with GATA alterations, given its pivotal roles in modulating HIF-1- and GATA-1-dependent gene transcription [75,80]. Overall, TACC3 was shown to regulate many key oncogenic processes in cancer cells (Figure 6) and was proven as a highly attractive therapeutic target.

Figure 6.

TACC3 coordinates key oncogenic processes in cancer cells. left. TACC3 causes G1/S progression via p38 and p53-dependent inhibition of p21. It induces cell proliferation, migration, and invasion via activating PI3K/AKT and ERK. Furthermore, TACC3 also causes mitotic progression and inhibition of apoptosis. Right. In cancer cells with amplified centrosomes, TACC3 has spatiotemporal functions in cell cycle. TACC3 interacts with the NuRD complex in nucleus, thereby inhibiting the transcription of tumor suppressors (e.g. p21 and APAF1), leading to G1/S progression and inhibition of apoptosis. It also leads to centrosome clustering in mitotic cells upon interaction with KIFC1 on centrosomes, thereby providing bipolar cell division, mitotic progression, and cell survival. This figure is generated in Biorender.com.

There are a few TACC3 inhibitors that have been developed so far. KHS101 was the first TACC3 inhibitor that was initially shown to selectively induce neuronal differentiation and exert similar effects to TACC3 knockdown in neural progenitor cells [109]. It suppresses cell growth, motility, EMT, and breast cancer cell stemness while inducing apoptosis in breast cancer cells [110]. KHS101 is also effective in hepatocellular carcinoma cells and suppresses cell growth and sphere formation as well as the expression of stem cell transcription factors, including Bmi1, c-Myc and Nanog. These effects are also recapitulated upon silencing TACC3, suggesting that TACC3 could be a novel regulator of cancer stem cell-like characteristics in hepatocellular carcinoma (HCC) cells [92]. KHS101 was also shown to indirectly destabilize TACC3 and HIF-1α, although it does not directly bind to the interacting region [77]. Despite the evidence on direct binding using recombinant TACC3 protein, the domain of TACC3 binding to KHS101 has not been identified. Furthermore, KHS101 was demonstrated to specifically bind and inhibit the mitochondrial chaperone, HSPD1 in glioblastoma cells, leading to cytotoxicity [111], which overall suggest that KHS101 is likely an unspecific inhibitor of TACC3 or may have other off-target effects. Another small-molecule inhibitor that was shown to target TACC3 was SPL-B which was identified using a chemical array screening followed by visual screening to assess mitotic arrest [112]. SPL-B was shown to bind to TACC3 by pull-down experiments although the exact binding region has not been identified. SPL-B induces aberrant spindles and mitotic arrest in ovarian cancer cells in a dose-dependent manner [112]. TACC3 has also been degraded using the IAP-based PROTACs, called SNIPER (Specific and Non-genetic IAP-dependent Protein ERaser) technology. SNIPER(TACC3) was one of the first SNIPERs to be developed, which induces poly-ubiquitylation and proteasomal degradation of TACC3 by the ubiquitin ligase APC/C(CDH1), thus reducing TACC3 protein levels [113]. However, SNIPER(TACC3) uses KHS101 to direct the eraser complex to TACC3, suggesting the presence of potential off-target activity.

More recently, we have developed a novel TACC3 inhibitor, BO-264 which was shown to bind to the TACC domain of TACC3 using various biochemical assays, including Drug Affinity Responsive Target Stability (DARTS) in cancer cells overexpressing truncated vs. full versions of TACC3, and Isothermal titration calorimetry (ITC) [11,67]. The specificity of BO-264 to TACC3 was demonstrated using shTACC3 cells which exhibit reduced response to BO-264 in terms of mitotic cell death. Notably, BO-264 demonstrates superior antiproliferative activity to KHS101 and SPL-B, and it strongly induces cell death especially in aggressive breast cancer subtypes, basal and HER2-positive, via spindle assembly checkpoint – dependent mitotic arrest, DNA damage, and apoptosis, while the cytotoxicity against normal breast cells is negligible. Oral administration of BO-264 significantly impairs tumor growth in immunocompromised and immunocompetent breast and colon cancer mouse models and increases survival without any major toxicity [67].

The role of TACC3 in aggressive cancers with centrosome amplification or drug resistance

In cancer cells with CA which is the hallmark of aggressive cancers, we identified TACC3 as a novel CA-directed dependency, driving cell growth by forming distinct functional interactomes during cell cycle progression [11]. We showed that inhibiting TACC3 in mitotic cells blocks the formation of TACC3/KIFC1 complex, leading to centrosome de-clustering, formation of multipolar spindles and activation of spindle assembly checkpoint (SAC) that ultimately results in mitotic cell death. On the other hand, TACC3 inhibition in interphase cancer cells with CA blocks TACC3/HDAC2/MBD2 complex, leading to enhanced transcription of cyclin-dependent kinase inhibitors (e.g. p21 and p16) and apoptosis regulators (e.g. APAF1), ultimately causing p53-independent G1 arrest and strong apoptosis [11]. Overall, TACC3 regulates survival of cancer cells with CA by forming cell cycle stage-specific interactions with KIFC1 and NuRD complex (Figure 6).

Besides being a key vulnerability in cancers with CA, TACC3 inhibition has also been tested in combination with chemotherapy or targeted therapy agents in a few studies. In cervical cancer cells, nontoxic amounts of paclitaxel exacerbate TACC3 inhibition-induced senescence onset [114]. Furthermore, overexpressing mouse Tacc3 in HeLa cells diminishes paclitaxel-induced G2/M arrest, apoptosis and growth inhibition [115]. In another study, combination of SNIPER(TACC3) with the proteasome inhibitor, bortezomib, was shown to increase stress-induced vacuole formation and lead to a synergistic increase in the anti-cancer activity in several cancer cell lines [116]. We recently showed that TACC3 inhibition revives T-DM1-induced immunogenic cell death by inhibiting microtubule polymerization and causing mitotic cell death that ultimately increases cytotoxic T-cell infiltration, thereby overcoming T-DM1 resistance in HER2+ breast cancer [117]. Despite these few studies, the in-depth characterization of the molecular mechanisms of TACC3-mediated drug resistance has largely been missing and necessitates further research.

The FGFR3-TACC3 fusion, its oncogenic functions, and its potential as a therapeutic target

FGFR3-TACC3 fusions, which are one of the most common fusions of FGFR in cancer [39], are formed by rare intrachromosomal rearrangements located within 150 kb of the FGFR3 gene and the TACC domain of the TACC3 gene, both located on chromosome 4p16 [94]. The fusion protein is mostly located at the membrane and also in vesicle-like structures [118] and in nucleus [119]. It has constitutive kinase activity and induces downstream PI3K/AKT and MAPK pathways. It also induces mitotic, chromosomal segregation defects, and triggers aneuploidy, in part due to the sequestration of the endogenous TACC3 protein away from the spindles [94].

Cancer cells bearing FGFR3-TACC3 fusions are sensitive to FGFR inhibition partly because of dependency on FGFR signaling [118]. However, the response is short-lived due to development of acquired resistance [120]. Unfortunately, cross-resistance is also common among different types of FGFR inhibitors [120], necessitating the identification of novel therapeutic strategies. Targeting ERBB3 or MEK kinases have shown to be effective against some of the fusion-bearing FGFR inhibitor resistant cells [120] although not all fusion-bearing tumors are sensitive to EGFR inhibition [96]. Interestingly, fusion-bearing tumors were also found to be resistant to the chemotherapy [121], temozolomide (TMZ), and targeting heat shock protein 90 (HSP90) potentiates TMZ-induced DNA damage and growth inhibition [122].

The presence of TACC domain is crucial for the activity of the FGFR3-TACC3 receptor, leading to more promiscuous constitutive phosphorylation of tyrosine kinase residues within the FGFR3 protein and activation of the downstream MAPK [118,119]. Interestingly, the oncogenic activity of the fusion protein comprising exons 8–16 of the TACC3 protein (containing the CID domain) is significantly stronger without increased MAPK activity than the fusion bearing exons 11–16 of TACC3, lacking the CID domain. This suggests that there might also be FGFR-independent oncogenic effects of the fusion protein that is likely related to the functions of TACC3. One potential TACC3-specific mechanism could be related to the nuclear functions of the fusion protein as introducing an FGFR3 kinase-dead mutation abrogates the FGFR pathway activation; however, it has no effect on the nuclear localization of the fusion protein [119]. The substantial loss of clonogenicity of the fusion-bearing cells upon deletion of the TACC domain suggests that targeting the TACC domain could be a highly attractive therapeutic strategy to inhibit the growth of FGFR3-TACC3 fusion-bearing tumors, especially those that are resistant to FGFR inhibitors. Indeed, targeting TACC3 using KHS101 reduces the expression of the FGFR3-TACC3 fusion protein and shows anti-tumor effect against FGFR3-TACC3 fusion-transfected cells [97]. We have recently showed that targeting TACC3 using BO-264 induces mitotic cell death in fusion-bearing bladder cancer cells [67]. However, more functional studies are warranted to demonstrate the molecular bases of the TACC3 dependency of FGFR3-TACC3 fusion-bearing cells.

Concluding remarks and future perspectives

TACC3 is a multi-functional protein overexpressed in cancer and plays key roles to drive the survival of the highly aggressive cancer cells, such as cancer cells with CA. TACC3 can be found in different cellular locations, e.g. spindles, centrosomes, and nucleus, and interact with many protein complexes in different cell cycle stages. It has a plethora of key spatiotemporal functions including the canonical spindle and centrosome-related roles, as well as other non-canonical functions involving chromatin remodeling and transcription regulation. The molecular functions of TACC3 in drug resistance are still not fully uncovered and need to be systematically studied in models resistant to different classes of anti-cancer therapies in vitro and in vivo. The roles of TACC3 in suppressing anti-tumor immunity are also just emerging. Future studies may investigate the potential functions of TACC3 in regulating tumor microenvironment and anti-tumor immunity and may test the therapeutic potential of combining TACC3 inhibitors with the established or newly developed immuno-oncology drugs. Furthermore, another future focus could be the investigation of a possible involvement of TACC3 in extracellular matrix modulation, given the inhibitory effects of targeting TACC3 on cell migration and invasion. Above all, the high-throughput characterization of the TACC3 interactomes that are spatiotemporally regulated may uncover the comprehensive map of diverse cellular processes known to be regulated by TACC3 and may further identify novel functions that have so far been unknown. Moreover, novel upstream regulators of TACC3 that facilitate its spatiotemporal functions throughout cell cycle, in addition to the known mitotic upstream kinase, Aurora A is fundamental to elucidate the dynamics of TACC3-driven processes and to uncover novel treatment strategies. Multimodal targeting of different TACC3 interactomes throughout the cancer cells will ultimately lead to highly effective inhibition of tumor growth and aggressiveness, thus encouraging the future clinical studies testing TACC3 targeting to improve the clinical outcome in many different cancers.

Funding Statement

This work is supported in part by National Institutes of Health [NIH, R01CA251374 to O. Sahin].

Disclosure statement

O. Sahin is the co-founder and manager of OncoCube Therapeutics LLC, and the founder and president of LoxiGen, Inc. O. Sahin is also member of the scientific advisory board (SAB) of A2A Pharmaceuticals, Inc. The other authors declare no potential conflicts of interest.

Data availability statement

Data sharing is not applicable to this article as no new data were created or analyzed in this review article.