Background
Tartrate-resistant acid phosphatase (TRAP/
ACP5) is a metalloenzyme of the category of acid phosphatases [
1] that is synthesized as a monomeric proenzyme (TRAP 5a, 35 kDa) [
2]. A disulfide linked heterodimer (TRAP 5b) with an N-terminal fragment of 20–23 kDa joined to the 16 to 17 kDa C-terminal part originates from post-translational cleavage of the monomeric form, which significantly increases phosphatase activity [
3]. TRAP derived from different mammalian sources reveals almost identical homology at the amino acid sequence level and identical biochemical properties [
4‐
7]. During bone resorption, TRAP is secreted into the resorption lacuna of active osteoclasts, where it dephosphorylates the bone matrix protein osteopontin (OPN), thereby promoting osteoclast detachment and migration [
8]. Additionally, TRAP has been suggested to regulate OPN bioactivity in autoimmune conditions [
9‐
11]. The isoform TRAP 5b was proposed as a serum marker for bone metastases in various types of primary cancers [
12‐
16]. Interestingly, TRAP has also been detected in several cancer cells and tissues (breast, ovarian, cervical cancer and malignant melanoma) and its expression level correlates with the severity of the tumor [
17‐
19]. Moreover, high TRAP expression correlates with reduced tumor- and metastasis-free survival in malignant melanoma [
20], and with decreased overall survival and increased incidence of metastasis in hepatocellular cancer [
21]. In gastric cancer, elevated TRAP expression is an independent risk factor for peritoneal dissemination and is associated with shorter patient survival [
22]. In lung cancer, patients with high TRAP expression had a significantly lower overall survival than the patients with low TRAP expression [
23].
Altogether, previous studies underscore the potential clinical relevance of TRAP to monitor cancer development and progression; nevertheless, underlying cellular and molecular processes remain unclear.
TRAP was shown to interact intracellularly with the Transforming growth factor β (TGFβ) receptor interacting protein-1 (TRIP-1), thereby activating TGFβ receptor type II (TβR2) and osteoblast differentiation through the Mothers against decapentaplegic homolog 2/3 (SMAD2/3) pathway at sites of prior bone resorption [
24]. Furthermore, TRIP-1 knock-down abrogates osteoblast differentiation and proliferation [
25]. TRAP 5a interaction with TRIP-1 has also been demonstrated in mouse pre-adipocytes [
26].
TGFβ ligands exist in three highly homologous isoforms, TGFβ1, TGFβ2 and TGFβ3 and are part of a large family of structurally related secreted cytokines [
27‐
29]. Upon ligand binding to the constitutively active serine/threonine kinase TβR2 [
30], the latter forms a hetero-oligomeric complex with the type I receptor (TβR1). TβR2 then trans-phosphorylates TβR1, ultimately leading to the transcription of various target genes via both SMAD- and non-SMAD mediated pathways (reviewed in [
29,
31‐
34]). TGFβ acts as tumor repressor in early stages of tumorigenesis and as an oncogene in late stages [
28]. For instance, in breast cancer patients expression of TGFβ was increased in tumor tissue [
35,
36] and was associated with disease progression [
36,
37]. TGFβ2 has been proposed as a predictive marker for breast cancer, as high levels of TGFβ2 correlate with advanced tumor stage and shortened survival [
38]. Additionally, TGFβ2 was reported as a catalyzer of TGFβ signaling through an autocrine loop [
39]. Finally, TβR1 contain a single cytoplasmic binding site for Cluster of differentiation 44 (CD44) [
40], suggesting a potential interaction between the TGFβ pathway and CD44, a cancer-associated glycoprotein previously reported as an OPN receptor [
41]. Activated CD44 stimulates the serine/threonine kinase activity of TβR1, which in turns increases SMAD2/3 phosphorylation [
40].
Aim of this study was to delineate by proof-of-concept, how TRAP promotes cellular properties related to metastasis in breast cancer cells at advanced state. As there is only limited knowledge about the molecular perturbations and possible substrates of TRAP, global phosphoproteomic and proteomic analysis was applied to connect possible signaling mechanisms to the TRAP-dependent phenotypic changes.
Discussion
TRAP has been shown to be associated with tumor progression in several types of cancer and suggested to be clinically relevant as a marker for peritoneal dissemination in gastric cancer [
20‐
22]. Recently, TRAP was identified as a pro-invasion oncogene and a prognostic marker in melanoma [
20]. Moreover, TRAP’s functional role in invasion, cell motility and metastasis was suggested to be mediated through phosphorylation of focal adhesion complexes [
20].
In this study, we investigated the effects of TRAP expression on cell properties related to the development of metastasis and on the cellular signaling of the MDA-MB-231 breast cancer cell line at the system level. The small molecule 5-phenylnicotinic acid (5-PNA), recently identified as a specific inhibitor of TRAP activity and TRAP-dependent migration and invasion, was employed as a tool to investigate the molecular events mediated by TRAP [
45,
59]. We demonstrate that TRAP regulates metastasis-related features such as anchorage-independent and –dependent cell growth, proliferation, migration and invasion. Global proteomics and phosphoproteomic analyses showed that regulated events upon TRAP perturbation are mainly proteins and phosphorylation sites involved with cellular adhesion and extracellular matrix (ECM) organization. Based on these analyses and on the literature, we hypothesized that TGFβ pathway-associated proteins and three previously unreported intracellular phosphorylation sites of CD44 mediate the observed TRAP-dependent cellular phenotypic properties.
TRAP has been proposed as a differentiation and growth factor for cells of hematopoietic origin [
60]. Effects of TRAP expression on cell transformation and tumor progression have been clinically validated in melanoma, as well as proven in vitro by an anchorage-independent growth assay [
20]. TRAP overexpression correlates with increased tumor size and poor differentiation in hepatocellular cancer [
21]. We consolidated and expanded these findings by using the breast cancer cell line MDA-MB-231, showing that TRAP overexpression increases cell growth as well as colony formation and cell proliferation above basal levels. Moreover, a higher number of TRAP-overexpressing cells compared to control cells were actively proliferating after 48 h serum starvation, indicating a lower requirement for exogenous growth stimulation.
Morphological changes such as cell rounding or cell spreading upon perturbation of TRAP expression have been reported in melanoma cells [
20]. Modulation of TRAP expression was shown to impact the migration and invasion of melanoma and hepatocellular carcinoma cells both in vitro and in vivo
, when either non-invasive cancer cells, expressing low amounts of TRAP or metastatic high-TRAP expressing cells were subjected to upregulation or knockdown, respectively [
20,
21]. Moreover, also in non-malignant epithelial cells TRAP expression was linked to a regulation of cell migration [
61]. This study demonstrates that TRAP overexpression enhances the elongated phenotype, migration and invasion capabilities of invasive breast cancer cells. Importantly, the elongated morphology and migration were regulated by TRAP in a dose-dependent manner. The presence of ECM proteins and basement membrane proteins Collagen IV and Laminin I increased transwell migration of TRAP3
high cells as compared to control cells, underscoring the role of TRAP during the invasive process. Transwell migration was particularly increased in the presence of osteopontin (OPN), a highly phosphorylated ECM protein previously suggested to be a physiological substrate for TRAP [
10], and involved in the progression of TRAP-related pathologies such as the immuno-osseous disorder Spondyloenchondrodysplasia [
9,
62]. OPN has been reported as a ligand to the CD44 receptor [
41] and was shown to increase osteoclast migration [
8], which is blunted upon antibody-mediated blocking of CD44 [
63]. Inhibition of TRAP by the small molecule inhibitor 5-PNA was previously reported to decrease TRAP3
high cells migration and invasion [
45]; here we showed that also proliferation of TRAP3
high cells is reduced to basal levels upon treatment with 5-PNA, altogether providing evidence that the above mentioned phenotypes of TRAP-overexpressing MDA-MB-231 cells are attributable to the overexpression of TRAP.
In parallel, global proteomics analysis of TRAP3high cells revealed regulation of various proteins belonging to the GO terms “biological adhesion” and “ECM organization”. Coherently, an increase in migration and invasion on various ECM and basement membrane proteins was observed in the TRAP3high cells. Enrichment in closely related GO terms, such as “cell adhesion molecule binding” and “cell junction”, was noted when analyzing phosphosites regulated in TRAP-overexpressing cells compared to control cells, further substantiating the involvement of TRAP in these functions. The list of 119 phosphorylation sites downregulated upon TRAP overexpression represent an inventory of putative targets of TRAP phosphatase activity or possible signaling intermediates; among those, eight sites with known regulatory function are involved in DNA damage response, another hallmark of cancer.
Most importantly, we identified a regulation of the TGFβ pathway-associated proteins TGFβ2, TβR1 and SMAD2, as well as a highly significant upregulation of previously unreported phosphorylation sites of CD44 upon TRAP perturbation in the MDA-MB-231 breast cancer cell line. Quantification of expression levels by several methodological approaches confirmed the upregulation of the ligand TGFβ2, which could be reverted by treatment with the TRAP inhibitor 5-PNA. Functional blocking of TGFβ2 or inhibition of TβR1/2 kinase activity restrained the increase in migration and proliferation promoted by TRAP. Antibody-mediated inhibition of CD44 reduced proliferation beyond the basal level of control cells.
Several reports support the concept that TGFβ promotes or restrains cell proliferation depending on the context [
28]. TGFβ inhibited the proliferation of most epithelial cells and its growth inhibitory effect could be partially reverted by treatment with a specific inhibitor [
64‐
66]. Poorly differentiated prostate cancer cells were resistant to TGFβ growth inhibitory effect in vivo [
67]
. Upon functional blocking of TGFβ2 or inhibition of TβR1/2 kinase activity, we observed decreased proliferation of TRAP-overexpressing cells but not of control MDA-MB-231 cells. Such lack of response in the control cells suggests that the malignant MDA-MB-231 cells are resistant to the growth-inhibitory effect of TGFβ, and that TRAP-dependent increase in proliferation of TRAP-overexpressing cells is TGFβ2-mediated. Furthermore, the invasive capacity of malignant breast cancer cells is enhanced by TGFβ1 [
68] and inactivation of TGFβ signaling inhibited invasiveness in vitro and in vivo for colon carcinoma cell lines [
69]. Additionally, here we detected a TRAP-dependent promotion of migration mediated by TGFβ2.
We could exclude a possible contribution of TGFβ1 because treatment with this TGFβ ligand modulated neither migration nor proliferation, as reported previously, despite constitutive expression of its receptors [
68]. Thereby we substantiated that the TGFβ2 isoform is crucial for the TRAP-mediated effects. In support to this notion, TGFβ2 was previously attributed a dominant role in predicting the outcome of breast cancers [
38] and aberrant expression of the TGFβ2 isoform exclusively was induced through an autocrine loop in glioma [
39].
We were also interested in the regulation of CD44, as it has previously been reported to be connected to TGFβ pathway [
70‐
72] and to be phosphorylated by TRβ1 [
40]. Our analysis identified three phosphorylation sites of CD44 as the top upregulated phosphorylation events in TRAP3
high cells. Additionally, CD44 is a receptor for OPN [
41], which is the protein that prominently increased transwell migration in TRAP-overexpressing cells in this study. Functional blocking of CD44 resulted in a decrease of proliferation of TRAP3
high cells beyond the basal level of control cells, but had no effect on cell migration. This suggests that cell migration is regulated by TRAP via TGFβ independently of CD44, but that the basal and TRAP-dependent proliferative activity of the TRAP3
high cells is regulated through both TGFβ signaling and CD44.
As this study is performed in vitro and uses a metastatic cell line, similar experiments in other cancer cell lines at different stages of metastatic progression would be crucial to fully delineate the effects of TRAP, possibly revealing a role in promoting metastasis of non-invasive cell lines as well [
20,
21]. Additionally, knockdown of TRAP in invasive cancer cells expressing high TRAP levels might be employed to test whether their invasive phenotype is TRAP-dependent. Limitations include also a full dissection of the respective TGFβ pathway and a reconfirmation, to allow for a generalization to other TRAP expressing cell lines.