Role of the CXCR4-LASP1 Axis in the Stabilization of Snail1 in Triple-Negative Breast Cancer

The CXCL12-CXCR4 axis plays a vital role in many steps of breast cancer metastasis, but the molecular mechanisms have not been fully elucidated. We previously reported that activation of CXCR4 by CXCL12 promotes the nuclear localization of LASP1 (LIM and SH3 protein 1). The nuclear LASP1 then interacts with Snail1 in triple-negative breast cancer (TNBC) cell lines. In this study, we report that the nuclear accumulation and retention of Snail1 was dependent on an increase in nuclear LASP1 levels driven by active CXCR4. The CXCR4-LASP1 axis may directly regulate the stabilization of nuclear Snail1, by upregulating nuclear levels of pS473-Akt, pS9-GSK-3β, A20, and LSD1. Furthermore, the activation of CXCR4 induced association of LASP1 with Snail1, A20, GSK-3β, and LSD1 endogenously. Thus, nuclear LASP1 may also regulate protein-protein interactions that facilitate the stability of Snail1. Genetic ablation of LASP1 resulted in the mislocalization of nuclear Snail1, loss of the ability of TNBC cells to invade Matrigel and a dysregulated expression of both epithelial and mesenchymal markers, including an increased expression of ALDH1A1, a marker for epithelial breast cancer stem-like cells. Our findings reveal a novel role for the CXCR4-LASP1 axis in facilitating the stability of nuclear localized Snail1.


Introduction
Breast cancer (BC) is a heterogeneous disease, and is the second leading cause of death in women among the cancer mortalities. Based on the new statistics from 2019, 3 out of 10 women will develop BC in their lifetime, and 1 in 7 will succumb to BC [1]. Mortality in BC patients is mainly due to metastasis to the lungs, bone, and brain. Triple-negative breast cancer (TNBC) is a highly aggressive form of BC, with an increased risk of relapse, drug resistance, and metastases. The durable or pathological complete response (pCR) is low in TNBC patients [2][3][4][5]. There is mounting evidence to suggest that the CXCL12-CXCR4 axis plays an important role in triple-negative breast cancer (TNBC) progression and metastasis [6][7][8][9][10][11]. Previously, we reported that the N-terminal LIM domain of LIM and SH3 protein 1 (LASP1) directly binds to CXCR4 through its C-terminal tail [12,13]. LASP1 is an adaptor protein, that is highly expressed in various cancer types, and is involved in cell migration when activated by CXCL12 [14]. Nuclear LASP1 expression is high in advanced and aggressive stages of breast cancer, and is a predictor of poor patient outcome [13][14][15]. Interestingly, CXCR4-mediated nuclear profile of LASP1 and SNAI1 is depicted by the beeswarm plot [15] (Figure 1A bottom panel). Snail1 plays a major role with respect to chemoresistance, tumor initiation, and metastasis, but the protein is highly unstable. In order to determine the potential role of the CXCR4-LASP1 axis and other proteins such as A20 in the regulation of the Snail1 stability, we screened several human luminal and basal breast cancer cell lines, for the protein levels of CXCR4, LASP1, LSD1, GSK-3β, Snail1 and A20 by the immunoblot analysis of total cell lysates. We observed that CXCR4 is expressed in all of the basal-like BC (BLBC) cell lines correlating with their invasive potential. The protein was undetectable in luminal BC (LBC) cell lysates in our system. Ectopically expressed CXCR4 in human embryonic kidney-293 (HEK-293) cells served as positive control ( Figure 1B). Next, we found that LASP1 is expressed uniformly high in both LBC and BLBC lines, with the exception of MCF7. LSD1 expression followed a similar profile to that of LASP1, with a tendency to higher levels in BLBC lines. Snail1 levels approximately mirrored the expression of CXCR4. A20 and GSK-3β are mostly expressed in all the cell lines, but we observed that the expression of GSK-3β was higher in MDA-MB-436 and SUM-159 PT. On the contrary, other BLBC cell lines showed lower expression of GSK-3β, further proving the importance of the MDA-Bone-Un cells ( Figure 1C). GSK-3β is a well-known negative regulator for the Snail1 stability. Overall, the protein levels of CXCR4, Snail1, and A20 correlated the most, and only some cancer cell line-specific aberrations were observed.

Activation of CXCR4 Signaling by CXCL12 Contributes to an Increased Stability of SNAIL1 through Modulation of Its Upstream Regulators
We employed a bone-tropic metastatic variant of the TNBC cell line, MDA-MB-231 (which is denoted as 'MDA-Bone-Un' or simply Bone-Un) for all the experiments. Bone-Un cells were derived from the MDA-MB-231 cells obtained from murine bone lesions after intracardiac injection. These cells metastasize to the bones when orthotopically implanted into mammary fat pads [48,49].
In order to elucidate the contributions of Akt, GSK-3β, LSD1, A20, and LASP1 to the stability of Snail1 when CXCR4 is activated, cells were incubated with 20 nM CXCL12, for various time points up to 30 min (Figure 2A). The level of CXCR4 peaked at 20 min, followed by a decline at 30 min, but still above the baseline level. Similar to this, the pS473-Akt level was elevated from 15-30 min above the baseline, with the levels normalized to total Akt. The level of pS9-GSK-3β increased steadily up to 20 min, followed by a minor decrease at 30 min with the levels normalized to total GSK-3β. There was only a marginal increase in the LSD1 after stimulation when compared to basal level until 30 min. The total levels of A20 increased in the first 20 min (20-fold increase), followed by a mild decline (13.3-fold) at 30 min. We then examined for the phosphorylation status of LASP1. Phosphorylation of LASP1 at S146 is known to induce nuclear translocation of the protein [50]. In addition, LASP1 contains a second phosphorylation site at Y171, which is reported to alter its cellular localization [51]. There was a marginal increase in the pS146-LASP1 levels at 20 min post stimulation, but the level of pY171-LASP1 increased up to 3.3-fold by 10 min, followed by a decrease to 2.9-fold at 30 min after stimulation. The phosphorylated forms of LASP1 were normalized to total LASP1 levels. Importantly, we examined the level of Snail1 following the stimulation of CXCR4 with 20 nM CXCL12. The levels of Snail1 in the total lysate dramatically increased by 20 min, which was followed by a decline at 30 min, but remained higher than basal level. This increase of Snail1 level correlated with a decrease in its transcriptional target fructose-1, 6-bisphosphatase (FBP1) and an increase in the mesenchymal cell marker vimentin.
Next, we examined the levels of the aforementioned regulators of Snail1 stability in the nuclear fractions. We observed a CXCR4-stimulated increase in nuclear accumulation of pS473-Akt from 10-30 min ( Figure 2B). The observed pattern of accumulation for pS473-Akt level mirrored that of pS9-GSK-3β. Furthermore, we observed an increase in nuclear accumulation of LSD1 and A20 as well. We observed an increase in nuclear localization of pS146-LASP1 similar to the expected lines of previous observations [2]. Interestingly, pY171-LASP1 exhibited a biphasic accumulation in the nucleus though the levels are not high. As reported by our previous findings [14], total LASP1 accumulated Cancers 2020, 12, 2372 4 of 20 upon the stimulation of CXCR4 by CXCL12 ( Figure 2B). We observed that the nuclear accumulation of LASP1 mirrored that of Snail1.
Cancers 2020, 12, x; doi: www.mdpi.com/journal/cancers embryonic kidney-293 (HEK-293) cells served as positive control ( Figure 1B). Next, we found that LASP1 is expressed uniformly high in both LBC and BLBC lines, with the exception of MCF7. LSD1 expression followed a similar profile to that of LASP1, with a tendency to higher levels in BLBC lines. Snail1 levels approximately mirrored the expression of CXCR4. A20 and GSK-3β are mostly expressed in all the cell lines, but we observed that the expression of GSK-3β was higher in MDA-MB-436 and SUM-159 PT. On the contrary, other BLBC cell lines showed lower expression of GSK-3β, further proving the importance of the MDA-Bone-Un cells ( Figure 1C). GSK-3β is a well-known negative regulator for the Snail1 stability. Overall, the protein levels of CXCR4, Snail1, and A20 correlated the most, and only some cancer cell line-specific aberrations were observed. Bottom panel-Beeswarm plot of differential gene expression profile of LASP1 and SNAI1-red symbols-high expression; black symbols-low expression (B). Proteins in the total cell lysates were separated by 10% SDS-PAGE, and subjected to Western blot analysis for differential expression of CXCR4 by probing with anti-CXCR4 antibody in luminal and basal-like breast cancer cell lines. Moreover, β-tubulin served as the loading control (C) Proteins in the total cell lysates were separated by 10% SDS-PAGE and subjected to Western blot analysis for differential expression of proteins  -free survival rate of basal-like, triple-negative breast cancer  patients with grade 3 tumors, with a differential expression of LASP1 and SNAI1 (n = 112). Bottom panel-Beeswarm plot of differential gene expression profile of LASP1 and SNAI1-red symbols-high expression; black symbols-low expression (B). Proteins in the total cell lysates were separated by 10% SDS-PAGE, and subjected to Western blot analysis for differential expression of CXCR4 by probing with anti-CXCR4 antibody in luminal and basal-like breast cancer cell lines. Moreover, β-tubulin served as the loading control (C) Proteins in the total cell lysates were separated by 10% SDS-PAGE and subjected to Western blot analysis for differential expression of proteins involved in Snail1 stabilization such as LASP1, LSD1, Snail1, and A20 in luminal and basal-like breast cancer cell lines. Furthermore, β-tubulin served as the loading control.

Constitutive Activity of CXCR4 Increases Stability of Snail1
In order to further assess the ability of CXCR4 to influence the stabilization of Snail1 levels, we employed a previously documented MCF7 series of cell lines. This includes vector-MCF7 (retroviral vector control) cells, WT-MCF7 (low ectopic expression of wild-type CXCR4) cells, and ∆CTD-MCF7 cells, which ectopically express a truncated CXCR4 that lack C-terminal with traditional phosphorylation sites [52]. In ∆CTD-MCF7 cells, the truncated CXCR4 is reported to demonstrate a gain-of function in signaling abilities. This was evident from a constitutively high ERK1/2 phosphorylation in the absence of CXCL12, and an observed increase in cell motility and mesenchymal markers. Importantly, ∆CTD-MCF7 cells were mesenchymal in morphology that was supported by the loss of classical epithelial markers, such as E-cadherin and the tight junction protein zona occludens-1 (ZO-1) [52]. Additionally, the ∆CTD-MCF7 cells demonstrated a high propensity for lung metastasis in a murine model [53].
Cancers 2020, 12, x; doi: www.mdpi.com/journal/cancers was only a marginal increase in the LSD1 after stimulation when compared to basal level until 30 min. The total levels of A20 increased in the first 20 min (20-fold increase), followed by a mild decline (13.3-fold) at 30 min. We then examined for the phosphorylation status of LASP1. Phosphorylation of LASP1 at S146 is known to induce nuclear translocation of the protein [50]. In addition, LASP1 contains a second phosphorylation site at Y171, which is reported to alter its cellular localization [51].
There was a marginal increase in the pS146-LASP1 levels at 20 min post stimulation, but the level of pY171-LASP1 increased up to 3.3-fold by 10 min, followed by a decrease to 2.9-fold at 30 min after stimulation. The phosphorylated forms of LASP1 were normalized to total LASP1 levels.  Similar to our studies involving the stimulation of endogenous CXCR4, we observed an increase in the levels of pS473-Akt, A20, LASP1 and Snail1 in the constitutively active CXCR4 model. To support the notion that stabilized Snail1 was indeed active, we probed for EMT markers and observed corresponding reductions in the expression of epithelial markers, such as E-cadherin, epithelial cell adhesion molecule (EpCAM), and the metabolic enzyme FBP1. This was accompanied by an increase in the levels of the mesenchymal marker vimentin ( Figure 2C).
Additionally, one dark band of the upregulated endogenous CXCR4 was observed to run a little higher, presumably due to phosphorylation. This was observed with the ectopic expression of constitutively active CXCR4 ( Figure 2C). It is known that the activation of PI3K signaling pathway leads to translocation of phospho-Akt (pAkt) to the nucleus. Nuclear pAkt would then phosphorylate GSK-3β, which ultimately prevents the degradation of Snail1. We observed a 2.7 and 9.7-fold increase in pS473-AKT, with ectopic expression of WT and constitutively active CXCR4, respectively ( Figure 2C).
Interestingly, despite the upregulation of total GSK-3β in the presence of ectopically expressed WT or constitutively active CXCR4, the level of pS9-GSK-3β remained unchanged in this constitutive model. The levels of LSD1, A20 and total LASP1 were also elevated when WT or constitutively active CXCR4 are expressed in MCF-7 cells ( Figure 2C).
Importantly, the levels of Snail1, as well as mesenchymal marker vimentin, were upregulated concomitant with a reduction in the expression of epithelial markers E-cadherin, epithelial cell adhesion molecule (EpCAM), and the metabolic enzyme FBP1 ( Figure 2C).

LASP1 Associates with Snail1 and Other Proteins that Promote Snail1 Stability
To further clarify the role of LASP1 in the stabilization of nuclear Snail1, we first investigated its association with known proteins which contribute to the stability of Snail1. This was accomplished through a glutathione S-transferase (GST) pulldown approach. We purified full length LASP1 and various domains of LASP1 fused to GST and examined for proteins that would associate from total lysates of MDA-Bone-Un cells ( Figure 3A). Regions of LASP1 that associate with Snail1, A20, Akt, LSD1, and GSK-3β were identified. Previously, we reported that LASP1 associates with Snail1 endogenously in the nucleus only when TNBC cells were stimulated with CXCL12 [14]. Therefore, we employed HEK-293 cells with the ectopic stable and transient expression of CXCR4 and Flag-Snail1. Nuclear lysates were prepared following stimulation of CXCR4-HEK-293 cells with 20 nM CXCL12 and incubated with GST-full length LASP1 and its various domains. Both full length LASP1 and its SH3 domain robustly associated with nuclear Snail1. There was some association observed between the LIM domain of LASP1 and Snail1. The NRLD of LASP1 failed to associate with Snail1 under our conditions ( Figure 3B). We then examined at the association of LASP1 with Snail1 regulators in MDA-Bone-Un lysates. The association of A20 mirrored that of Snail1, except for the protein more robustly associated with the LIM domain of LASP1. A similar binding pattern was detected for LSD1, i.e., association with full length LASP1 as well as with the LIM and SH3 domain, but not with the NRLD ( Figure 3C). For Akt, association to the full length LASP1 and SH3 domain of LASP1 was observed. GSK-3β was the only protein that associated with the NRLD domain of LASP1 ( Figure 3D).
Cancers 2020, 12, x 6 of 20 various domains of LASP1 fused to GST and examined for proteins that would associate from total lysates of MDA-Bone-Un cells ( Figure 3A). Regions of LASP1 that associate with Snail1, A20, Akt, LSD1, and GSK-3β were identified. Previously, we reported that LASP1 associates with Snail1 endogenously in the nucleus only when TNBC cells were stimulated with CXCL12 [14]. Therefore, we employed HEK-293 cells with the ectopic stable and transient expression of CXCR4 and Flag-Snail1. Nuclear lysates were prepared following stimulation of CXCR4-HEK-293 cells with 20 nM CXCL12 and incubated with GST-full length LASP1 and its various domains. Both full length LASP1 and its SH3 domain robustly associated with nuclear Snail1. There was some association observed between the LIM domain of LASP1 and Snail1. The NRLD of LASP1 failed to associate with Snail1 under our conditions ( Figure 3B). We then examined at the association of LASP1 with Snail1 regulators in MDA-Bone-Un lysates. The association of A20 mirrored that of Snail1, except for the protein more robustly associated with the LIM domain of LASP1. A similar binding pattern was detected for LSD1, i.e., association with full length LASP1 as well as with the LIM and SH3 domain, but not with the NRLD ( Figure 3C). For Akt, association to the full length LASP1 and SH3 domain of LASP1 was observed. GSK-3β was the only protein that associated with the NRLD domain of LASP1 ( Figure 3D). and its domains fused to GST were incubated with 200 µg of nuclear lysate derived from 293-CXCR4 cells that were transiently expressing Flag-Snail1-WT and stimulated with 20 nM CXCL12. The association of Flag-Snail1 with full length LASP1 and its domains were analyzed by 10% SDS-PAGE, followed by immunoblotting; n = 3 (C) and (D) Association of LASP1 and its domains with proteins that contribute to the stabilization of the Snail1. Nuclear lysates were prepared from MDA-Bone-Un cells that were serum-starved and stimulated with 20 nM CXCL12. 1.5 nmol of each of the GST and LASP1 and its domains fused to GST were incubated with 250 µg of nuclear lysate derived from CXCL12-stimulated MDA-Bone-Un cells. The association of endogenous A20, LSD1, AKT and GSK-3β with full length LASP1 and its domains were analyzed by 10% SDS-PAGE followed by immunoblotting; n = 3. of Flag-Snail1 with full length LASP1 and its domains were analyzed by 10% SDS-PAGE, followed by immunoblotting; n = 3 (C) and (D) As sociation of LASP1 and its domains with proteins that contribute to the stabilization of the Snail1. Nuclear lysates were prepared from MDA-Bone-Un cells that were serum-starved and stimulated with 20 nM CXCL12. 1.5 nmol of each of the GST and LASP1 and its domains fused to GST were incubated with 250 µg of nuclear lysate derived from CXCL12-stimulated MDA-Bone-Un cells. The association of endogenous A20, LSD1, AKT and GSK-3β with full length LASP1 and its domains were analyzed by 10% SDS-PAGE followed by immunoblotting; n = 3.
2.5. Nuclear LASP1 Co-Immunoprecipitates with Snail1, A20, GSK-3β, and LSD1 To validate our findings from the GST pull-down approach, we performed a co-immunoprecipitation experiment, to examine the endogenous association of nuclear LASP1 with Snaill, A20, GSK-3β, and LSD1, when CXCR4 is activated. We indeed demonstrate that nuclear localized LASP1 associated with Snail1, A20, GSK-3β, and LSD1 endogenously in a CXCL12-dependent manner. Importantly, this association could be abrogated by treatment with AMD3465, supporting the specificity of CXCR4 ( Figure 4A).
Cancers 2020, 12, x 7 of 20 2.5. Nuclear LASP1 Co-Immunoprecipitates with Snail1, A20, GSK-3β, and LSD1 To validate our findings from the GST pull-down approach, we performed a coimmunoprecipitation experiment, to examine the endogenous association of nuclear LASP1 with Snaill, A20, GSK-3β, and LSD1, when CXCR4 is activated. We indeed demonstrate that nuclear localized LASP1 associated with Snail1, A20, GSK-3β, and LSD1 endogenously in a CXCL12dependent manner. Importantly, this association could be abrogated by treatment with AMD3465, supporting the specificity of CXCR4 ( Figure 4A). In some conditions, Bone-Un cells were pre-incubated with the CXCR4 antagonist AMD-3465 for 30 min prior to the addition of the ligand CXCL12. LASP1 was immunoprecipitated with mouse anti-LASP1 (8C6 clone) antibodies from 250 µg of the nuclear extracts from MDA-Bone-Un cells and analyzed by 10% SDS-PAGE followed by immunoblotting for associated Snail1, A20, GSK-3β, and LSD1. 'M' represents the mock co-immunoprecipitation performed with isotype control IgG1 antibodies. Bottom panel: 15 µg of nuclear lysates were resolved by 10% SDS-PAGE, and immunoblotted for Snail1, A20, GSK-3β, LSD1 and LASP1 blot. (B) Differential association of phosphorylated forms of LASP1 to endogenous Snail1 and its regulators-1.5 nmol of each of the GST, LASP1-WT and its phosphomimetic (S146D and Y171D) and phosphonull (S146A and Y171A) mutants fused to GST were incubated with 250 µg of nuclear lysate derived from CXCL12-stimulated MDA-Bone-Un cells. The differential association of endogenous Snail1, GSK-3β, A20 and LSD1 with full length LASP1-WT and its mutants were analyzed by 10% SDS-PAGE, followed by immunoblotting; n = 3. (C) Occupancy of LASP1 at the E-cadherin promoter region-Serum-starved MDA-Bone-Un cells were stimulated with 20 nM CXCL12 for 20 min. In some conditions, Bone-Un cells were pre-incubated with the CXCR4 antagonist AMD-3465 for 30 min prior to the addition of the ligand CXCL12. The chromatin fragments were prepared and subjected to chromatin immunoprecipitation (ChIP) analysis. The data from quantitative, real-time PCR were statistically analyzed and shown as the Mean ± SD, n = 3 independent biological repeats; p < 0.05.
2.6. Differential Association of Phosphorylated Forms of LASP1 with Snail1, A20, GSK-3β, and LSD1 Next, we wanted to clarify the role of the phosphorylation status of LASP1 and its association with Snail1 and its regulators. We purified full length GST-LASP1 carrying various phosphomimetic and phosphonull substitutions at S146 and Y171. The phosphorylation status of LASP1 did not In some conditions, Bone-Un cells were pre-incubated with the CXCR4 antagonist AMD-3465 for 30 min prior to the addition of the ligand CXCL12. LASP1 was immunoprecipitated with mouse anti-LASP1 (8C6 clone) antibodies from 250 µg of the nuclear extracts from MDA-Bone-Un cells and analyzed by 10% SDS-PAGE followed by immunoblotting for associated Snail1, A20, GSK-3β, and LSD1. 'M' represents the mock co-immunoprecipitation performed with isotype control IgG1 antibodies. Bottom panel: 15 µg of nuclear lysates were resolved by 10% SDS-PAGE, and immunoblotted for Snail1, A20, GSK-3β, LSD1 and LASP1 blot. (B) Differential association of phosphorylated forms of LASP1 to endogenous Snail1 and its regulators-1.5 nmol of each of the GST, LASP1-WT and its phosphomimetic (S146D and Y171D) and phosphonull (S146A and Y171A) mutants fused to GST were incubated with 250 µg of nuclear lysate derived from CXCL12-stimulated MDA-Bone-Un cells. The differential association of endogenous Snail1, GSK-3β, A20 and LSD1 with full length LASP1-WT and its mutants were analyzed by 10% SDS-PAGE, followed by immunoblotting; n = 3. (C) Occupancy of LASP1 at the E-cadherin promoter region-Serum-starved MDA-Bone-Un cells were stimulated with 20 nM CXCL12 for 20 min. In some conditions, Bone-Un cells were pre-incubated with the CXCR4 antagonist AMD-3465 for 30 min prior to the addition of the ligand CXCL12. The chromatin fragments were prepared and subjected to chromatin immunoprecipitation (ChIP) analysis. The data from quantitative, real-time PCR were statistically analyzed and shown as the Mean ± SD, n = 3 independent biological repeats; * p < 0.05. 2.6. Differential Association of Phosphorylated Forms of LASP1 with Snail1, A20, GSK-3β, and LSD1 Next, we wanted to clarify the role of the phosphorylation status of LASP1 and its association with Snail1 and its regulators. We purified full length GST-LASP1 carrying various phosphomimetic and phosphonull substitutions at S146 and Y171. The phosphorylation status of LASP1 did not robustly regulate the LASP1-Snail1 association, although pS146D-LASP1 showed a lower affinity. Interestingly, this mutant did not associate with GSK-3β. On the contrary, A20 associated with all forms of LASP1, except for the Y171F mutant. Lastly, there appears to be no preferential association between the phosphomimetic and phosphonull mutations of LASP1 with LSD1, but the WT full length LASP1 did not associate very well ( Figure 4B).

Nuclear LASP1 Associates with the Promoter of E-cadherin in A CXCL12-Dependent Manner
Our previous [14] and current evidence had demonstrated a nuclear interaction between LASP1, Snail1, and its regulators. We next assessed the functional outcome of the nuclear localized LASP1 and its interaction with Snail1 and LSD1, by examining the occupancy of nuclear LASP1 at the E-cadherin promoter (a well-documented Snail1 and LSD1 target), using a chromatin immunoprecipitation (ChIP) assay. We observed a 2-fold enrichment of LASP1 at the E-cadherin promoter following stimulation with CXCL12. This enrichment was significantly abrogated by pre-treatment of MDA-Bone-Un cells with the CXCR4 antagonist AMD-3465 ( Figure 4C).

Genetic Ablation of LASP1 Led to Loss of the Ability to Invade Matrigel
To demonstrate the importance of LASP1 in the Snail1-EMT pathway, we knocked out LASP1 by employing a CRISPR-Cas9 approach in MDA-Bone-Un cells. Single cell colonies were isolated and showed no change in morphology ( Figure 4A, left panel). Successful knockout was verified by immunoblotting ( Figure 4A, right panel). Next, we investigated for the ability of the LASP1-KO MDA-Bone-Un cells to invade Matrigel. In CRISPR-control cells, incubation with 25 nM CXCL12 induced a 4-fold increase in the number of cells that invaded compared to unstimulated cells. On the contrary, we observed impairment in the invasion of LASP1-KO MDA-Bone-Un cells following CXCL12 stimulation ( Figure 4B).
Cancers 2020, 12, x 8 of 20 Cancers 2020, 12, x; doi: www.mdpi.com/journal/cancers Interestingly, this mutant did not associate with GSK-3β. On the contrary, A20 associated with all forms of LASP1, except for the Y171F mutant. Lastly, there appears to be no preferential association between the phosphomimetic and phosphonull mutations of LASP1 with LSD1, but the WT full length LASP1 did not associate very well ( Figure 4B).

Nuclear LASP1 Associates with the Promoter of E-cadherin in A CXCL12-Dependent Manner
Our previous [14] and current evidence had demonstrated a nuclear interaction between LASP1, Snail1, and its regulators. We next assessed the functional outcome of the nuclear localized LASP1 and its interaction with Snail1 and LSD1, by examining the occupancy of nuclear LASP1 at the Ecadherin promoter (a well-documented Snail1 and LSD1 target), using a chromatin immunoprecipitation (ChIP) assay. We observed a 2-fold enrichment of LASP1 at the E-cadherin promoter following stimulation with CXCL12. This enrichment was significantly abrogated by pretreatment of MDA-Bone-Un cells with the CXCR4 antagonist AMD-3465 ( Figure 4C).

Genetic Ablation of LASP1 Led to Loss of the Ability to Invade Matrigel
To demonstrate the importance of LASP1 in the Snail1-EMT pathway, we knocked out LASP1 by employing a CRISPR-Cas9 approach in MDA-Bone-Un cells. Single cell colonies were isolated and showed no change in morphology ( Figure 5A, left panel). Successful knockout was verified by immunoblotting ( Figure 5A, right panel). Next, we investigated for the ability of the LASP1-KO MDA-Bone-Un cells to invade Matrigel. In CRISPR-control cells, incubation with 25 nM CXCL12 induced a 4-fold increase in the number of cells that invaded compared to unstimulated cells. On the contrary, we observed impairment in the invasion of LASP1-KO MDA-Bone-Un cells following CXCL12 stimulation ( Figure 5B).

Cytoplasmic Localization of Snail1 Following the Genetic Loss of LASP1
Finally, we examined for the level of Snail1 and the factors that affect its stability of Snail1 in LASP1-KO MDA-Bone-Un cells by immunoblotting total cell lysates. Surprisingly, the level of Snail1 protein was elevated by 4.8-fold when LASP1 was genetically ablated ( Figure 4C). Importantly, we did not observe a similar elevation of vimentin, a protein normally reflecting the activity of Snail1. The epithelial markers EpCAM and FBP1 were elevated. In addition, we also observed an increase in the aldehyde dehydrogenase (ALDH) expression ( Figure 4C) and activity ( Figure 4E) following LASP1 knockout, which is indicative of a shift towards the epithelial phenotype [54], along with the retention of some mesenchymal markers. This further confirms that the cells exhibit a hybrid phenotype with both epithelial and mesenchymal characteristics. Moreover, the level of E-cadherin remained repressed. In contrast to LASP1-KO MDA-Bone-Un cells, the genetic loss of LASP1 in a model HEK-293 cell line showed the expected Snail1 downregulation by 0.6-fold. The levels of the epithelial markers, such as E-cadherin (1.9-fold), EpCAM (2.1-fold) and FBP1 (2.5-fold) were upregulated as expected. Concurrently, we also observed an increased ALDH1A1 expression ( Figure 4D) and ALDH1A1 activity ( Figure 4E). Under two-dimensional culture conditions, we analyzed for the levels of cytokeratin14 (CK14; a mesenchymal marker) and cytokeratin18 (CK18; an epithelial marker) upon genetic loss of LASP1. There was a marginal increase (1.5-fold) in CK18, but the level of CK14 was not altered ( Figure S1).
To further analyze the unexpected upregulation of Snail1 in MDA-Bone-Un cells after LASP1 knockout, we compared the levels of proteins that confer or affect the stability of Snail1 in nuclear lysates ( Figure 4F). In the nuclear lysate, the levels of Akt, GSK-3β, LSD1 and A20 were downregulated upon genetic loss of LASP1. The level of pS473-Akt was reduced by 0.5-fold, and a marginal decrease in pS9-GSK-3β was noticed ( Figure 4F).
To corroborate this upregulation in the cytoplasmic and downregulation in the nuclear levels of Snail1, we examined the subcellular distribution of Snail1 upon the genetic loss of LASP1 by confocal microscopy ( Figure 4G). Immunofluorescent staining of Snail1 in CRISPR-control and LASP1-KO MDA-Bone-Un cells revealed a pattern which matched the distribution observed in the biochemical fractionation experiments involving nuclear and cytosolic fractions. Figure 5 depicts a pictorial representation of the CXCR4-LASP1 axis contributing to Snail1 stability.
Cancers 2020, 12, x 11 of 20 Cancers 2020, 12, x; doi: www.mdpi.com/journal/cancers Figure 6. A working model illustrating the role of CXCR4-LASP1 axis in promoting the stabilization of nuclear Snail1. Left panel: A proposed model illustrating the cellular situation when CXCL12 is absent-In the absence of any activation of CXCR4 by CXCL12, the constitutively phosphorylated LASP1 at S146 (pS146-LASP1) is bound to inactive, unliganded CXCR4 at the plasma membrane. Phosphorylation of Snail1 by GSK-3β on specific residues results in cytosolic localization of Snail1. Cytoplasmic Snail1 undergoes additional phosphorylation by GSK-3β, which marks Snail1 for ubiquitination and subsequent proteasomal degradation. Gene expression of E-cadherin is not repressed by Snail1, though repression by other transcription factors such as ZEB1 is not ruled out. This happens in epithelial cancer cells. Right panel: Activation of CXCR4 by CXCL12 leads to the phosphorylation of LASP1 at Y171 and translocation of LASP1. Additionally, active CXCR4 induces phosphorylation/activation of Akt (pS473-AKT) through the PI3K pathway, which traverses to the nucleus and initiates the phosphorylation/inhibition of nuclear GSK-3β (pS9-GSK-3β). Concurrently, the CXCR4 signaling increases the nuclear levels of A20 and LSD1. Each of these proteins induces the stabilization of Snail1 in a variety of ways. A20 mediates monoubiquitination of Snail1 thus reducing its affinity for GSK-3β. Nuclear shuttled LASP1 and elevated LSD1 level may physically shield Snail1 and prevent access of GSK-3β to Snail1. The association of pLASP1 with Snail1, pAKT, A20 and LSD1 in the form of several complexes of differing compositions will protect and stabilize Snail1. The stabilized Snail1 represses the promoter region of E-cadherin. This happens in mesenchymal cancer cells. Snail1: Snail family transcriptional repressor; LASP1: LIM and SH3 protein 1; Akt: Protein kinase B; GSK-3β: Glycogen synthase kinase-3β; LSD1: lysine-specific histone demethylase 1; A20: TNF-α induced protein 3; E-cad: E-cadherin; Ub: Ubiquitination.

Discussion
In this study, we dissected the contributions of the CXCR4-LASP1 axis on the stability of Snail1 in TNBC cells. Snail1 is a key EMT-transcription factor that regulates cell migration, invasion, breast cancer stemness, and drug resistance in clinical TNBC cases of tumor relapse and therapy failure [30,55,56]. Therefore, the regulation of Snail1 is also critical for the control of primary tumor progression and for the regulation of many steps in metastasis [56]. In the present study, we observe upregulated expression levels of CXCR4, A20, and Snail1 in basal-like TNBC cells, suggesting that these proteins might be involved in the stabilization of Snail1. A20 is known to promote the stability of Snail1 through its monoubiquitination of Snail1 [45,57]. Nuclear pS473-Akt is an active serine situation when CXCL12 is absent-In the absence of any activation of CXCR4 by CXCL12, the constitutively phosphorylated LASP1 at S146 (pS146-LASP1) is bound to inactive, unliganded CXCR4 at the plasma membrane. Phosphorylation of Snail1 by GSK-3β on specific residues results in cytosolic localization of Snail1. Cytoplasmic Snail1 undergoes additional phosphorylation by GSK-3β, which marks Snail1 for ubiquitination and subsequent proteasomal degradation. Gene expression of E-cadherin is not repressed by Snail1, though repression by other transcription factors such as ZEB1 is not ruled out. This happens in epithelial cancer cells. Right panel: Activation of CXCR4 by CXCL12 leads to the phosphorylation of LASP1 at Y171 and translocation of LASP1. Additionally, active CXCR4 induces phosphorylation/activation of Akt (pS473-AKT) through the PI3K pathway, which traverses to the nucleus and initiates the phosphorylation/inhibition of nuclear GSK-3β (pS9-GSK-3β). Concurrently, the CXCR4 signaling increases the nuclear levels of A20 and LSD1. Each of these proteins induces the stabilization of Snail1 in a variety of ways. A20 mediates monoubiquitination of Snail1 thus reducing its affinity for GSK-3β. Nuclear shuttled LASP1 and elevated LSD1 level may physically shield Snail1 and prevent access of GSK-3β to Snail1. The association of pLASP1 with Snail1, pAKT, A20 and LSD1 in the form of several complexes of differing compositions will protect and stabilize Snail1. The stabilized Snail1 represses the promoter region of E-cadherin. This happens in mesenchymal cancer cells. Snail1: Snail family transcriptional repressor; LASP1: LIM and SH3 protein 1; Akt: Protein kinase B; GSK-3β: Glycogen synthase kinase-3β; LSD1: lysine-specific histone demethylase 1; A20: TNF-α induced protein 3; E-cad: E-cadherin; Ub: Ubiquitination.

Discussion
In this study, we dissected the contributions of the CXCR4-LASP1 axis on the stability of Snail1 in TNBC cells. Snail1 is a key EMT-transcription factor that regulates cell migration, invasion, breast cancer stemness, and drug resistance in clinical TNBC cases of tumor relapse and therapy failure [30,55,56]. Therefore, the regulation of Snail1 is also critical for the control of primary tumor progression and for the regulation of many steps in metastasis [56]. In the present study, we observe upregulated expression levels of CXCR4, A20, and Snail1 in basal-like TNBC cells, suggesting that these proteins might be involved in the stabilization of Snail1. A20 is known to promote the stability of Snail1 through its monoubiquitination of Snail1 [45,57]. Nuclear pS473-Akt is an active serine kinase that stabilizes Snail1 through the inactivating phosphorylation of nuclear GSK-3β on S9. We observed that, in the presence of active CXCR4, the levels of both pS473-Akt and pS9-GSK-3β increased, indicating that signaling from CXCR4 might stabilize Snail1. The previous reports indicate that GSK-3β actively controls the localization of Snail1 through the phosphorylation of Snail1 protein which targets it for cytoplasmic degradation [41]. Thus, inactivation of GSK-3β enzymatic activity by CXCR4 signaling would have significant impact on Snail1 stability. We indeed observed an increase in the level of Snail1 around 20-30 min post stimulation with CXCL12. A similar observation was made in the constitutively active CXCR4 model in MCF7 cells. Recently, it has been shown that pS146-LASP1 can bind to CXCR4 at the plasma membrane and upon activation of CXCR4, LASP1 is released and traverses to the nucleus [13,51]. Interestingly, we made a novel observation that the phosphorylation of LASP1 on Y171 increases upon activation of CXCR4 in TNBC cells. In normal HEK cells, this phosphorylation is less dramatic [58], maybe due to the lower expression of tyrosine kinases in non-cancer cells.
In all, the phosphorylation status of LASP1 may be context dependent. To investigate the importance of p-LASP1 in our system, we employed GST-LASP1 pulldown studies, with phosphomimetic substitutions demonstrating that LASP1, phosphorylated on either S146 [50] or Y171 [59], can indeed associate with Snail1 (though Snail1 associated with pS146-LASP1 with a reduced affinity). It is possible that the nuclear association of LASP1 with Snail1 can physically shield it from GSK-3β-mediated phosphorylation. Additionally, through its association with LSD1, LASP1 may form a LASP1-Snail1-LSD1 tripartite complex, which could additionally protect Snail1 from GSK-3β [46,[60][61][62][63]. Eventually, nuclear LASP1 may also stabilize Snail1 by orchestrating its monoubiquitination via A20. Interestingly, the levels of A20 and LSD1 plummeted upon knockout of LASP1 in MDA-Bone-Un cells. We observed the association of AKT to the SH3 domain of LASP1, while another study found that the NRLD domain of LASP1 interacted with AKT [51]. This may be due to the stimulation of our TNBC cells, which led this differential association of the domains of LASP1 to AKT. In fact, LASP1 is nuclear in high grades of breast cancer. Nuclear localization of LASP1 and the 10-year survival rate are inversely correlated; i.e., poor patient outcome with the increasingly nuclear localized LASP1 [14,15]. In support of this, we observed that the high gene expression of LASP1 and SNAI1 in grade 3 tumors correlated with low probability of RFS in a cohort of 112 TNBC patients that we described in Figure 1A. The nuclear co-abundance of LASP1 and Snail1 in high grade tumors will most likely increase the propensity for local invasion and metastasis.
To support the notion that the activity of Snail1 was indeed impaired in LASP1 KO cells, we examined the downstream EMT characteristics in these cells. Bone-Un LASP1 KO cells had an impaired ability to invade Matrigel. Importantly, the levels of epithelial markers and the transcriptional repressor targets for Snail1 such as EpCAM and the metabolic enzyme FBP1 increased [36]. Moreover, we observed an increase in ALDH-positive epithelial cancer stem-like cells. At the same time, the vimentin level remained the same, implicating a possible hybrid phenotype, having both epithelial and mesenchymal markers in TNBC cells [64,65]. Importantly, Snail1 was predominantly localized to the cytosol and depleted in the nucleus.
We are also the first, to the best of our knowledge to provide the functionality of the LASP1-Snail1 interaction by demonstrating that LASP1 occupied the E-cadherin promoter in a CXCL12-dependent manner, probably as a co-factor together with Snail1 and LSD1, shown here for the first-time. The E-cadherin level was not rescued upon genetic loss of LASP1, indicating several layers of repression for E-cadherin in TNBC cells. However, when LASP1 was knocked out in HEK-293 cells, the level of E-cadherin was indeed restored. Similarly, when LASP1 was stably silenced, E-cadherin level was rescued back in MCF7 and MDA-MB-361 luminal breast cancer cell lines in our previous studies [14]. Earlier work also discussed nuclear LASP1 as part of the transcription factor AP-1 [66].
Overall, our study indicates that the CXCR4-LASP1 axis can contribute to the nuclear stability of Snail1 through several mechanisms. This includes the activation of Akt, induction of nuclear localization and/or retention of A20 and LSD1, and a potential for nuclear LASP1 to physically shield and protect Snail1 from GSK-3β induced phosphorylation and subsequent degradation in the cytosol. The resulting nuclear retention of Snail1 may contribute to a migratory and invasive phenotype, which then aids in the many steps of TNBC metastasis in vivo.

Cell Culture
The human breast cancer cell lines BT20 These plasmids encode the Cas9 nuclease and non-specific 20 nucleotide guide RNAs (denoted CRISPR control). The GFP-sorted cells were cultured, and the level of LASP1 expression in each single cell colony was verified by immunoblotting. HEK-293-HA-CXCR4 LASP1 KO and CRISPR control cells were generated previously [67].

Preparation and Immunoblotting of Nuclear Extracts
Intact nuclei and cytosolic fractions were separated by employing the Nuclei EZ Prep kit (Sigma-Aldrich, St. Louis, MO, USA), according to the manufacturer's instructions. Pure nuclei were extracted from cellular lysates with nuclear extraction buffer (50 mM Tris, pH 8.0, 350 mM NaCl, 1 mM DTT, protease and phosphatase inhibitors (5 mM MgCl 2 and 50 units of DNAse I), and incubating for 2 h at 4 • C [14]. Total protein in the clarified nuclear lysate was quantified using a Bradford protein assay (Bio-Rad, Hercules, CA, USA; Cat. No. 5000006). Moreover, 25-50 µg of total protein from the nuclear extract was separated by 10% SDS-PAGE and transferred onto nitrocellulose membrane (0.45 µM) overnight. The membranes were then blocked with 5% nonfat dry milk (NFDM) in TBST (20 mM Tris, pH 7.6, 100 mM NaCl, 0.1% Tween-20) for 1 h. Subsequently, the membranes were rinsed with TBST and incubated with primary antibodies in 5% BSA or NFDM for 2 h at 4 • C. Subsequently, the membranes were washed thrice with TBST, and incubated with secondary antibodies (1:5000), conjugated with horseradish peroxidase (HRP) in 1% BSA or NFDM in TBST for 1 h at R.T. The membranes were washed thrice with TBST, and the proteins were detected by enhanced chemiluminescence (ECL) Prime kit (GE Healthcare Life Sciences; Cat. No. RPN2232,).

Immunoblotting
The 25-50 µg of total or nuclear lysates were separated by 10% SDS-PAGE and transferred onto nitrocellulose membrane (0.45 µm), overnight at 25 • C. Then, the membranes were blocked with 5% nonfat dry milk (NFDM) in TBST (20 mM Tris, pH 7.6, 100 mM NaCl, 0.1% Tween-20). After this, the membranes were washed with TBST and incubated with the following primary antibodies in 5% BSA or NFDM for 2 h at 4

Nuclear co-Immunoprecipitation
Serum-starved Bone-Un Cells were stimulated with 20 nM CXCL12 (ligand for CXCR4) for 20 min at 37 • C. Some batches of cells were pre-incubated with AMD3465 prior to the addition of CXCL12. The clarified nuclear lysate was prepared as described in aforementioned 'Immunoblotting of nuclear extract'. LASP1 was immunoprecipitated with a mouse monoclonal anti-LASP1 clone-8C6, which was generated against amino acid sequence 151-162 (SYRRPLEQQQPH) within the LASP1 linker domain (not present in LASP2), and associated proteins were analyzed as described [12,14,69].

Glutathione S-Transferase (GST) Pulldown Assay
For the GST pull-down assay, glutathione S-transferase (GST), GST-LASP1, and its domain fusion proteins constructs (GST-LIM, GST-NRLD (Nebulin repeats and linker domain), GST-SH3 (C-terminal domain)) were expressed in BL21 bacteria. The GST fusion proteins were isolated as described [12,14,67]. The beads were aliquoted and stored at −80 • C. Furthermore, 1.5 nmol of the GST-LASP1 (Full length), GST-LIM (N-terminal domain), GST-NRLD, GST-SH3 (C-terminal domain) and GST bound to glutathione beads were washed thrice with binding buffer (50 mM Tris-HCl pH 7.5, 0.05% Triton X-100, 100 mM NaCl -Co-IP buffer) and incubated with nuclear lysate overnight at 4 • C. Subsequently, the beads were washed thrice with Co-IP buffer and the bound proteins were eluted with 2×Laemmli buffer and resolved by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). This was followed by immunoblotting for proteins associated to LASP1 and its various domains.

Subcellular Localization of Snail1
The CRISPR control and LASP1-KO cells were seeded onto glass cover slips coated with Collagen IV, and allowed to attach and spread overnight. The cells were fixed on the following day with 4% paraformaldehyde and permeabilized by Triton X-100. After overnight incubation with Snail1 antibody (Cell Signaling Technology, Cat. No. 3879), the coverslips were washed thrice with phosphate-buffered saline, pH 7.5 (PBS). Snail1 was visualized with the Alexa Fluor 488 conjugated goat anti-rabbit secondary antibody at RT for 2 h. Cellular outline was visualized by decorating the filamentous actin (F-actin) with FITC-Phalloidin (Life Technologies, Carlsbad, CA, USA, Cat. No. A12379) for 30 min, followed by DRAQ5 (Thermo Scientific, Cat No. 62251) for 10 min, to visualize the nuclei at RT. The coverslips were air-dried and mounted on glass slides. The subcellular distribution of Snail1 was analyzed by Leica TCS SP5 multiphoton laser scanning confocal microscopy. Analysis of images was performed using the Leica Application Suite X software.

Chromatin Immunoprecipitation (ChIP) Assay
The chromatin was prepared as described [70]. Briefly, serum starved 'Bone-Un' cells were stimulated with and without 20 nM CXCL12, and the protein-DNA complexes were cross-linked with 37% formaldehyde, diluted to a final concentration of 1% for 10 min. The formaldehyde was quenched with 2.5 M glycine diluted to a final concentration of 125 mM for 10 min at RT. Cells were washed and scraped into PBS, pH 7.5, pelleted, washed again and treated with 'Lysis buffer I' (50 mM HEPES-KOH, pH 7.5, 140 mM NaCl, 1 mM EDTA, 10% glycerol, 0.5% IGEPAL CA-630, 0.25% Triton X-100, 1× protease inhibitors), followed by nutation for 10 min at 4 • C to release the nuclei. The nuclei were recovered by centrifugation and resuspended in 1 mL 'Lysis buffer II' (10 mM Tris-HCl, pH 8.0, 200 mM NaCl, 1 mM EDTA, and 0.5 mM EGTA, 1× protease inhibitors), and nutated for 10 min at 4 • C. After centrifugation, the pellet was resuspended in 'Lysis buffer III' (10 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 0.1% sodium deoxycholate, 0.5% N-lauroylsarcosine and 1× protease inhibitors). Samples were then placed on ice and sonicated in 30 s pulses set at 50% amplitude, for a total of 8 min. After the sonication, the samples were centrifuged at 13,000× g for 10 min at 4 • C, and the purified chromatin extract (supernatant) was collected. The chromatin concentration was quantified by Nanodrop. Next, 100 µg of chromatin extract was pre-cleared and incubated with isotype control IgG or anti-LASP1 antibody overnight at 4 • C. On the next day, 80 µL of 'Protein A/G plus Fast Flow'-agarose beads (GE Healthcare Life Sciences) were added, and the incubation continued for an additional 4 h. The beads were washed thrice with wash buffer (50 mM HEPES-KOH, pH 7.5, 500 mM LiCl, 1 mM EDTA, 1.0% IGEPAL CA-630, 0.7% sodium deoxycholate), and finally eluted in elution buffer (50 mM Tris-HCl, pH 8.0, 10 mM EDTA, 1.0% SDS). The cross-links were reversed by heating at 65 • C overnight (~15 h), and the chromatin was extracted with phenol-chloroform, and the DNA was precipitated with 80% ethanol from the aqueous phase. The precipitated DNA was washed twice with 70% ethanol, and the air-dried pellet was resuspended in diluted TE (1 mM Tris, pH 8.0, 0.5 mM EDTA). Furthermore, 3 µL of the final resuspended DNA was subjected to qPCR by employing the human E-cadherin promoter [(+) 5 -GGAGGGGTCCGCGCTGCTGA and (−)5 -AGCTCACAGGTGCTTTGCAG)] and β-tubulin [(+) 5 CTGGACCGCATCTCTGTGTACTAC3 ) and (−) 5 GACCTGAGCGAACAGAGTCCAT)] primers.

Matrigel Invasion Assay
Invasion of CRISPR-control and LASP1-KO cells was assessed by employing a transwell Matrigel assay kit (8-µM pore size) (Corning). Then, 1 × 10 5 cells/well resuspended in 200 µL serum-free medium were added to the upper Matrigel pre-coated chamber (BD Bioscience). The lower chamber of the transwell plate was filled with 600 µL of serum free DMEM, with 20 nM CXCL12 as the chemoattractant. After incubation overnight at 37 • C, the cells attached on the inner surface of the lower chamber were fixed with 100% methanol. The fixed cells were visualized by staining with 0.1% crystal violet (Sigma-Aldrich). The cells that invaded through Matrigel were counted. All the experiments were performed in triplicate.

Kaplan-Meier Survival Analysis
The probability of relapse-free survival (RFS) of basal-like breast cancer patients with a high vs. low expression of LASP1 and SNAI1 was analyzed using the Kaplan-Meier survival analysis (https://kmplot.com/). The 10-year RFS was assessed in 112 TNBC patients (Grade 3 tumors) with the hazard ratio (HR) and log rank p value.

Statistical Analysis
Analysis of the data was performed by employing the software GraphPad Prism ver. 7.0. Statistical significance between groups was determined by Student's t-test. All the experiments had three biological replicates.

Conclusions
We determined that the CXCR4-LASP1 axis contributes to the stability of the EMT-TF Snail1, which is a key mediator of TNBC metastasis.