Palladium nanoplates scotch breast cancer lung metastasis by constraining epithelial-mesenchymal transition

Metastasis accounts for majority of cancer deaths in many tumor types including breast cancer. Epithelial-mesenchymal transition (EMT) is the driving force for the occurrence and progression of metastasis, however, no targeted strategies to block the EMT program are currently available to combat metastasis. Diverse engineered nanomaterials (ENMs) have been reported to exert promising anti-cancer effects, however, no ENMs have been designed to target EMT. Palladium (Pd) nanomaterials, a type of ENM have received substantial concern in nanomedicine due to their favorable photothermal performance for cancer therapeutics. Herein, Pd nanoplates (PdPL) were found to be preferentially biodistributed to both primary tumors and metastatic tumors. Importantly, PdPL showed a significant inhibition of lung metastasis with and without near-infrared (NIR) irradiation. Mechanistic investigations revealed that EMT was significantly compromised in breast cancer cells upon the PdPL treatment, which was partially due to the inhibition of the transforming growth factor-beta (TGF-β) signaling. Strikingly, the PdPL was found to directly interact with TGF-β proteins to diminish TGF-β functions in activating its downstream signaling, as evidenced by the reduced phosphorylation of Smad2. Notably, TGF-β-independent pathways were also involved in undermining EMT and other important biological processes that are necessary for metastasis. Additionally, NIR irradiation elicited synergistic effects on PdPL-induced inhibition of primary tumors and metastasis. In summary, these results revealed that the PdPL remarkably curbed metastasis by inhibiting EMT signaling, thereby indicating the promising potential of PdPL as a therapeutic agent for treating metastases. results suggest that the PdPL inhibited cell division at a high dose, which is in agreement with the in vivo findings of the inhibitory effect of accumulated PdPL on tumor growth at the tumor sites. Therefore, the anti-metastatic feature of the PdPL can be attributed to the repression of primary tumor cells and the impairment of the propensity of inborn metastasis induced by the PdPL.

are the major reason for adverse disease outcomes among cancer patients. For example, metastatic tumors account for over 90% mortality among breast cancer patients [1]. Metastasis is a consecutive process, which involves the increased shedding of tumor cells from the primary tumor in a direction parallel to the growth of primary tumors, thereby leading to an ongoing delivery of tumor cells to distant organs [2], which essentially determines the prognosis of the patient. Metastasis is a multistep and dynamic cascade that begins from a primary tumor and requires the circulation of metastasizing cells through the lymph or blood, followed by the lodgment of tumor cells in the target organ(s). Metastasis involves a number of complex biological processes, in which epithelial-mesenchymal transition (EMT) largely dictates the invasion/spread and subsequent re-localization of cancer cells in distant organs [3]. EMT is a fine-tuned program controlling the transformation of epithelial cells into mesenchymal-like cells with more fibroblast properties, thereby increasing the invasion and migration capabilities of tumor cells [4]. Among a plethora of upstream regulators of EMT, the transforming growth factor-β (TGF-β) plays a crucial function in driving EMT during metastasis through the successive regulation of the genes responsible for cellular contact, plasticity, motility, stemness, and so on [5]. TGF-β ligands are known to bind with their transmembrane receptors (TGF-βR) to phosphorylate adapter and regulatory proteins, mainly Smad2 and Smad3 (Smad2/3), to orchestrate gene expression and contribute to EMT [6]. Research is being carried out to discover drugs that can block TGF-β signaling [6].
Nanomedicine continues to bring unprecedented advantages in the development of cancer therapeutics [7]. Engineered nanomaterials have been extensively studied for the inhibition and elimination of primary tumor growth; however, studies on the use of ENM-based strategies to selectively target metastases remain rather limited [8]. ENM-based anti-cancer approaches include nanocarrier-assisted chemo/gene delivery, photothermal therapy, and immunotherapy [9]. Although these treatments effectively kill cancer cells, several factors such as drug leakage, multidrug resistance, limited tissue penetration depth of near-infrared (NIR) light, and immune tolerance limit the effectiveness of these treatment, and ENM-based selective anti-metastatic therapies are unavailable. However, the potential of ENMs per se as therapeutic agents in cancer treatment has attracted increased attention. For instance, ENMs including cuprous oxide, zinc oxide, and silver nanoparticles are themselves cytotoxic, but they do not selectively kill cancer cells [10][11][12]. Additionally, ENMs such as gold nanoparticles (AuNPs) and Gd@C 82 (OH) 22 are also known to restrict tumor cell proliferation [13,14]; however, they do not show selectivity towards metastatic tumors. In a recent study, we reported the potential of palladium (Pd)-based ENMs (such as PdPL) as a therapeutic agent in cancer treatment owing to their outstanding photothermal and photoacoustic properties, and desirable bio-/cyto-compatibility [15,16].
Importantly, in comparison with other ENMs, PdPL displays greater photostability, and is more prone to localize to the cancer tissue [17,18]. However, regardless of these encouraging progresses, research on the effect of ENMs on EMT to diminish metastasis has rarely been conducted. Therefore, it is important to conduct studies on the effect of ENMs on EMT to target metastasis.
Herein, we primarily assessed the efficacy of PdPL to treat metastasis using different cancer models. We found that PdPL showed remarkable tropesis to target metastatic cancer cells, and strikingly suppressed lung metastasis. Mechanistic investigation revealed that PdPL partially inhibited the metastatic propensity of cancer cells by constraining EMT which was dependent on the blockade of TGF-β signaling. Thus, PdPL represents a novel agent against cancer metastases.

Synthesis and characterization of PdPL
Hexagonal PdPL was prepared by reducing Pd(acac) 2 using carbon oxide (Fig. S1), as described in our previous reports [19]. Transmission electron microscopy (TEM) analysis revealed that the obtained PdPL was monodispersed with an average particle diameter of approximately 11 nm ( Fig. 1a). Furthermore, atomic force microscopy (AFM) showed that the thickness of the asprepared PdPL was approximately 1.8 nm (Fig. 1b). Additionally, the as-prepared PdPL showed a strong surface plasmon resonance absorption in the NIR region, i.e., 808 nm ( Fig. 1c) (Fig. 1d). The excellent photothermal performance of the PdPL was further verified using the infrared thermal imaging analysis (Fig. 1e). The PdPL showed high photothermal stability, and no agglomeration or aggregation was observed after NIR irradiation (Fig. S2) [19].
Additionally, no visible change was observed in the photothermal efficacy of the PdPL after irradiation for 4 cycles (5-min irradiation followed by a 5-min cooling interval) (Fig. S2a).
A crucial technical limitation of NIR irradiation is its shallow penetration into tissues, which is typically limited to 3.4 cm [20]. To determine if the NIR laser was able to penetrate the chest shield in mice to target metastatic nodules in the lungs (a major target organ for cancer metastases), we assessed the photothermal efficacy of the PdPL under the chest shields that were dissected from mice (Fig. S3). The photothermal performance remained strong under these conditions, as a large increase in temperature (characterized by ΔT/°C) was observed in the PdPL solutions at all concentrations, with a maximum increase of 16.6 ΔT/°C at 100 μg/mL (Fig.   1f). These data indicated that the NIR irradiation adequately penetrated the murine chest, and the induced temperature elevation was sufficient to trigger the ablation of tumor nodules in the lung.

Significant PdPL accumulation in primary tumors and metastatic tumors
The tumor-targeting propensity of the PdPL was evaluated using our previously established mouse model of breast cancer lung metastasis from orthotopic tumors (Fig. 2a 1 ) [21]. In this mouse model, 4T1-derived subline LG12 (4T1-LG12) breast cancer cells showed highly efficient and extremely selective metastasis to the lungs [21]. Four weeks after orthotopic transplantation of 4T1-LG12 cells, after the lung metastases were well-established, 10 mg/kg body weight of PdPL was injected intravenously (i.v.) to the animals. Mice were sacrificed after 6, 12, 24, and 48 h injection, followed by the collection of primary tumors, metastatic lung nodules, and lung tissues adjacent to metastatic loci, and other organs (heart, liver, spleen, kidneys, and brain). The Pd content, measured by the inductively coupled plasma mass spectrometry (ICP-MS), in each organ confirmed that the PdPL mainly accumulated in the liver and spleen, as shown in Fig. 2b, which was consistent with previous observations [22]. In support of the above-mentioned findings, an agglomerated PdPL was also clearly observed in the liver specimens from treated mice (Fig. S4). Notably, the PdPL was remarkably localized in tumors including the primary and metastatic tumors (Figs 2c-e), thus showing PdPL amounts comparable to that observed in the liver and spleen after normalization (Fig. 2b). Additionally  [28,29]. Under this condition, a mechanism similar to that occurring in the primary tumors was proposed to be responsible for the retention of the nanoparticle in the metastatic tumors.
Moreover, the pronounced tropesis of the PdPL to intrude metastatic tumors in the lung can be ascribed to its two-dimensional (2D) nanostructure, in analogy with the molecular basis, as we recently discovered for other 2D nanomaterials, such as graphene oxide [30]. In support of this finding, we found that molybdenum disulfide/graphene oxide (MoS 2 /GO), a 2D nanocomposite, displayed selectively target the lungs, owing to the easy capture of the lung capillary vessels for the GO-protein complexes, as described in our previous report [30]. Therefore, our findings revealed that the PdPL could effectively target primary tumors, and metastatic nodules in the lung.

Suppression of lung metastasis by PdPL in various murine models
The anti-metastasis efficiency of the PdPL was assessed using various mouse models (Fig. 2a).
First, the effects of the PdPL administration on the metastatic propensity of the i.v. injected 4T1- LG12 cells (210 4 cells/mouse) were assessed with and without NIR irradiation (2 W/cm 2 , 5 min/each side) (Fig. 2a 2 ). The bioluminescence data showed that lung metastases were observed in all groups 2 weeks after the i.v. injection of the 4T1-LG12 cells (Fig. 3a), which was in agreement with our previous report [21]. Mice were randomly divided into four groups including the control group: (equal volume of phosphate-buffered saline [PBS]); NIR group: NIR irradiation (2 W/cm 2 , 5 min/each side); PdPL group: PdPL injection (10 mg/kg body weight); and NIR+PdPL group: PdPL injection (10 mg/kg body weight) and NIR irradiation (2 W/cm 2 , 5 min/each side). The term 5-min/each side refers to irradiation of the back and chest of the mouse for 5 min. During the 15-day treatment period, we found that in comparison with the untreated groups, the metastatic tumor growth of the PdPL-and NIR+PdPL-treated groups was restrained immediately after PdPL administration, particularly in combination with NIR irradiation (Fig. 3 metastasis formation or growth (Fig. 3).
To corroborate these observations, histological examination with hematoxylin-eosin (H&E) staining was performed (Fig. 3c). Multiple lung nodules were observed in the untreated mice and the NIR-treated mice; however, a few nodules were found in the lungs of the PdPL-treated mice, and almost no modules were found in the lungs of NIR+PdPL-treated mice (Fig. 3c). Nodule counts confirmed these differences, as the PdPL treatment led to a 51% reduction in nodule number, while the NIR+PdPL treatment caused 82% reduction, compared to the untreated controls ( Fig. 3d, P<0.05). Additionally, the H&E staining also revealed that the PdPL and NIR+PdPL treatment caused no overt damage to the normal lung tissues and other organs (heart, liver, spleen, and kidneys) (Figs 3c and S5).
To substantiate the above-mentioned findings, we investigated the suppressive effect of the PdPL on the lung metastasis from orthotopic primary breast tumors in a murine model (Fig. 2a 1 ).
Briefly, mice were injected with 4T1-LG12 cells ( Similar to the above-mentioned results as described in the metastasis model (Fig. S6), Figure S10 shows a marked decline of the primary tumors with approximately 40% drop in the tumor weight upon PdPL treatment (10 mg/kg body weight) (P<0.05), compared to that in the control group.
Furthermore, as shown in Fig. S11, histological examination with H&E staining indicated that the PdPL did not cause significant injuries to the heart, liver, spleen, lung, and kidney.
Additionally, the levels of inflammatory cytokines, including interferon (IFN)-γ, interleukin (IL) 6, and tumor necrosis factor (TNF)-α, were not significantly induced in the PdPL-treated mice, compared to those in the control mice (Fig. S7). Furthermore, no abnormal changes were observed in the PdPL-treated mice for the blood tests including white blood cell (WBC), We studied the effect of the PdPL on the proliferation in various cell lines. As shown in Figs S12 and S13, the PdPL did not inhibit cell division at lower concentrations, however, a significant repression of proliferation was observed at the highest concentration, 100 μg/mL, at

Alteration of the intrinsic metastatic propensity of cancer cells by PdPL
We determined the underlying molecular mechanisms of the PdPL-induced suppression of the metastatic propensity of breast cancer cells (Fig. 4a). The cellular morphology of the 4T1-LG12 cells after PdPL treatment substantially changed from a spindle-like shape to a pebble-like shape, namely from a "mesenchymal" to an "epithelial" shape, and their length greatly decreased post treatment, compared to the untreated cells (Fig. 4b) LG12 cells, a significant decrease of tumor nodules (65% decrease) was found in the lungs of the mice that acquired the PdPL-pretreated B16F10 cells, compared to the mice that received control cells (Fig. S15, P<0.05). These results revealed that the PdPL considerably undermined the inherent metastatic propensity of cancer cells. In fact, some monoclonal antibodies (e.g., 2G7 and 1D11), acting as pan-TGF-β-neutralizing antibodies, have been used to block lung metastases [39,40]. Based on the above-mentioned results, it could be inferred that the PdPL could have functions similar to those of the TGF-β-neutralizing antibody. Therefore, the PdPL harbors a great potential to block TGF-β actions, suggesting that the PdPL is a promising therapeutic agent for treating TGF-β-dependent disorders.

Impairment of the EMT signaling pathways by PdPL
To understand the molecular basis for the reduction of the metastatic potential of tumor cells by the PdPL, we conducted close mechanistic investigations. First, we established that higher concentrations of PdPL (up to 100 μg/mL) were not overtly toxic to the 4T1 cells (Fig. S16). Our previous research also found little toxicity of PdPL in non-tumor normal cells, including NIH-3T3 (normal mouse fibroblasts) and QSG-7701 (normal human liver cells) [18]. These combined data demonstrated that PdPL elicited little toxicity to normal cells. Moreover, our data showed that PdPL could be readily taken up by the 4T1-LG12 cells (Fig. S17). Additionally, the ICP-MS analysis revealed an increase in cellular uptake of PdPL over the time course from 6 to 12 and 24 h in the 4T1-LG12 cells responding to PdPL at 50 μg/mL (Fig. 5a, P<0.05 Since PdPL has a high photothermal efficacy when internalized by the 4T1-LG12 cells, NIR irradiation should be able to kill the cells. As shown in Fig. 5b, a dose-dependent phototoxicity was observed, which was consistent with the in vivo results, as described above ( (Fig. 5f). Additionally, a consistent increase in the occludin and decrease in the βcatenin levels were demonstrated in the 4T1-LG12 cells in vitro upon PdPL treatment (Fig. 5g).
It is generally believed that EMT is the driving force for cancer metastases, as the EMT program essentially remodels the cytoskeleton from the polar epithelial cell phenotype to an actin stress fiber dominated phenotype indicative of mesenchymal cells [5,46]. Specifically, the enhanced βcatenin signaling promotes metastasis in various cancers, due to the transcriptional activation of the genes responsible for EMT induction [47]. In contrast, the loss of occludin, which is necessary for the integrity of tight junctions between epithelial and endothelial cells, can result in the loss of cell-to-cell adhesion, thereby leading to reinforced EMT [48]. Therefore, the differential changes of these EMT-associated genes undermine the inherent EMT of 4T1 cells, namely the reprogramming to MET, in response to the PdPL treatment. Additionally, the transcriptome analysis also revealed that a number of proliferation-related genes were disturbed upon PdPL treatment, such as the cleavage stimulation factor subunit 3 (Cstf3), lymphocyte antigen 6 family member e (Ly6e), and glutaminase 2 (Gls2) (Figs 5e and S21), further supporting the PdPL-induced suppression of the tumor cell growth.
Blockade of TGF-β signaling by PdPL and damage to the metastatic potential of cancer cells Further, we attempted to probe the upstream regulators that may be responsible for the compromised EMT induced by the PdPL. Since TGF-β plays a central role in governing EMT by regulating some target genes [6,49], we compared our RNA-Seq results with the previously reported TGF-β target genes [50,51], and a wealth of common genes were found in between. To substantiate these changes, the target genes were further assessed by RT-qPCR in the cells upon the PdPL treatment. As shown in Fig. S23, blunted TGF-β signaling upon PdPL treatment was confirmed by the differentially expressed genes including Id1, Id3, Lcn2, and Ndrg1 in the 4T1 and 4T1-LG12 cells (P<0.05 and P<0.0001). As previously established, Smad2/3 phosphorylation is the most important surrogate to recognize TGF-β signaling activation [52,53].
Therefore, phosphorylated Smad2 (P-Smad2) was examined using western blotting in tumor specimens. As shown in Fig. 6a, a decrease in P-Smad2 concentration was found in the tumors from mice upon PdPL treatment, especially PdPL+NIR treatment, thereby indicating the significant inhibition of TGF-β signaling via Smad2/3 in the tumors upon PdPL treatment.
However, it should be noted that TGF-β conducts both canonical Smad-dependent signaling and Smad-independent signaling to promote EMT. In fact, as described above, our data also recognized Smad-independent targeting of downstream of TGF-β, such as Ocln [54], and the PdPL-induced changes in these genes were assumed to contribute to the inhibition of EMT. To EMT is a complex biological process that drives tumor metastasis; however, this process has not been fully understood, and TGF-β is not the only modulator in this process. Thus, we LncRNAs and miRNAs showed that some LncRNAs and miRNAs were predicted to target genes involved in EMT-related signaling pathways, such as miRNAs (e.g., novel_mir111 and novel_mir52) and LncRNAs (e.g., NONMMUT075484.1 and NONMMUT036870.2) (Figs S25 and S26). Although these findings are very intriguing, the PdPL may also target other proteins and RNAs to compromise EMT. Therefore, extensive study on the detailed regulation network response to PdPL is required.