Plant Lectin Can Target Receptors Containing Sialic Acid, Exemplified by Podoplanin, to Inhibit Transformed Cell Growth and Migration

Cancer is a leading cause of death of men and women worldwide. Tumor cell motility contributes to metastatic invasion that causes the vast majority of cancer deaths. Extracellular receptors modified by α2,3-sialic acids that promote this motility can serve as ideal chemotherapeutic targets. For example, the extracellular domain of the mucin receptor podoplanin (PDPN) is highly O-glycosylated with α2,3-sialic acid linked to galactose. PDPN is activated by endogenous ligands to induce tumor cell motility and metastasis. Dietary lectins that target proteins containing α2,3-sialic acid inhibit tumor cell growth. However, anti-cancer lectins that have been examined thus far target receptors that have not been identified. We report here that a lectin from the seeds of Maackia amurensis (MASL) with affinity for O-linked carbohydrate chains containing sialic acid targets PDPN to inhibit transformed cell growth and motility at nanomolar concentrations. Interestingly, the biological activity of this lectin survives gastrointestinal proteolysis and enters the cardiovascular system to inhibit melanoma cell growth, migration, and tumorigenesis. These studies demonstrate how lectins may be used to help develop dietary agents that target specific receptors to combat malignant cell growth.


Introduction
Over 1 in 4 people are diagnosed with cancer at some point in their life [1]. These cancers are the leading cause of death of men and women under 85 years old [2]. Tumor cell motility contributes to metastatic invasion that causes the vast majority of cancer deaths [3,4]. Extracellular receptors that promote this motility can serve as ideal chemotherapeutic targets [5][6][7][8][9]. Expression levels of receptors with a2,3-sialic acid residues are closely associated with the invasive and metastatic potential of many cancers [10][11][12] including skin cancer [13,14]. Podoplanin (PD PN) represents one of these receptors.
PDPN is a unique transmembrane receptor that promotes tumor cell motility. PDPN expression can be induced by tumor promoters including TPA, oncogenic Ras, and Src [15][16][17]. For example, we have reported that Src utilizes the focal adhesion adaptor protein Cas to induce PDPN expression in order to promote tumor cell migration [15]. PDPN regulates the activities of effectors including ezrin, Rho, and Cdc42 to mediate filopodia formation and promote tumor cell migration, invasion, and metastasis [5,[18][19][20][21][22].
PDPN is found at the invasive front of many tumors, which is consistent with its role in promoting malignant invasion [5,20,23]. For example, PDPN expression is strongly induced in about 40% of breast cancers [20,24], 50% of oral cancers [23,25,26], and 80% of skin cancers [27,28]. The bulk of the PDPN protein, about 150 amino acids, lies outside of the cell and could serve as an ideal target to combat cancer invasion and metastasis [5,29].
Here, we show that a lectin with affinity for O-linked carbohydrate chains containing sialic acid binds to PDPN to inhibit transformed cell growth and motility at submicromolar concentrations. This approach demonstrates how lectins may be used as dietary agents that target specific receptors to combat malignant cell growth.

MASL targets PDPN on Src transformed cells
We have recently reported that the Src tyrosine kinase utilizes the adaptor protein Cas/BCAR1 to augment PDPN expression in order to promote tumor cell motility [15]. PDPN is also profoundly involved in the ability of transformed cells to escape contact normalization (growth control by surrounding nontransformed cells) and ''break out'' of their microenvironment to become malignant or metastatic [15]. Consistent with these previous reports, Src transformed cells expressed more PDPN (see Figure 1a and 1b) and migrated more than nontransformed controls ( Figure 1c). As described above, the extracellular domain of PDPN is O-glycosylated with sialic acid, a2,3 linked to galactose. Scholl et al have reported that PDPN associates with lectin from seeds of the legume tree Maackia amurensis (MASL) [30], which has an affinity for O-linked carbohydrate chains containing sialic acid [51,52]. These observations led us to investigate the ability of MASL to target PDPN on transformed cells.
As shown in Figure 2a, MASL associated with PDPN on the membrane of Src transformed cells. Note that MASL did not target the membrane of cells that did not express PDPN, which, along with functional studies described below, attests to its preferential targeting. In addition, MASL associated with PDPN in lysates from transformed cells during affinity precipitation experiments as shown in Figure 2b.

MASL inhibits Src transformed cell growth and migration
We hypothesized that binding of MASL to PDPN would inhibit transformed cell migration. As shown in Figure 2c, MASL significantly inhibited transformed cell migration, with 385 nM and 770 nM inhibiting cell migration by over 25% and 50% respectively. In contrast, MASL did not inhibit the migration of nontransformed cells in a dose dependent fashion at these concentrations.
To verify the functional relevance of MASL targeting PDPN on cell migration, we investigated its effects on nontransformed cells transfected with PDPN or empty parental vector. Since PDPN expression is sufficient to promote cell migration [5,15,30], nontransformed cells transfected with PDPN migrated several fold more than control transfectants (Figure 3a and 3b). In addition, MASL reduced the migration of these nontransformed PDPN transfected cells in a dose dependent fashion ( Figure 3b). For example, 385 nM MASL decreased the migration of PDPN transfectants by over 40%.
As shown in Figure 3c, in addition to inhibiting cell migration, MASL was also toxic to PDPN expressing cells in a dose dependent fashion. In contrast, MASL did not inhibit the viability of empty vector transfectants in an equally dose dependent fashion. For example, 1540 nM MASL decreased Trypan blue exclusion of PDPN transfectants by over 70%, but control transfectants by only about 30%.

MASL targets PDPN to inhibit melanoma cell growth and motility
Studies indicate that PDPN expression is strongly induced in about 80% of skin cancers [27,28]. Consistent with its role in tumor cell invasion and metastasis, malignant B16 melanoma cells expressed higher levels of PDPN ( Figure 4a) and migrated significantly better ( Figure 4b) than syngeneic nontransformed Melan-a melanocytes. As shown in Figure 4c, MASL effectively suppressed melanoma cell migration at concentrations of 308 nM or less. In addition to inhibiting melanoma cell migration, MASL also inhibited melanoma cell growth in a dose responsive manner ( Figure 4d). Moreover, MASL was significantly more toxic to B16 melanoma cells than Melan-a cells (p,0.001 by ANOVA).
Transwell chambers were used to further investigate the effects of MASL on melanoma cell growth and migration. As shown in Figure 4e, while 385 nM MASL was not significantly toxic to B16 melanoma cells (p.0.2 compared to controls), migration through 8 micron pores was decreased by over 40 fold. These data indicate that MASL suppressed melanoma cell migration prior to inhibiting cell viability.
We employed siRNA to verify the effects of PDPN and MASL on melanoma cell growth and migration. PDPN siRNA effectively decreased B16 Pdpn expression levels ( Figure 5a) and cell migration ( Figure 5b). As shown in Figure 5c, this decreased PDPN expression resulted in a 25% decrease in MASL toxicity. These data indicate that while PDPN may not be the only receptor targeted by MASL on these melanoma cells, it is a functionally relevant receptor that can be targeted to prevent melanoma cell growth and migration.

MASL may induce caspase independent necrosis of melanoma cells
Some lectins used as potential anticancer agents, including Mistletoe lectins, contain a distinct N-glycosidase motif that acts as a ribosome inactivating peptide (RIP) [45]. Since MASL does not contain this peptide, inhibition of tumor cell viability by MASL did not result from RIP activity. In general, cell death can occur by apoptosis, necrosis, or autophagy [53,54]. MASL did not induce characteristics of autophagy such vacuolarization or the formation of autophagosomes, or characteristics of apoptosis including nuclear chromatin condensation, membrane blebbing, or chromosomal DNA fragmentation. Instead, MASL induced morphological changes typical of necrosis including cell swelling, membrane rupture, and nuclei fragmentation.
Apoptosis usually results from the induction of caspase activity which cleaves poly(ADP-ribose) polymerase (PARP) into an 89 kD fragment [55,56]. As shown in Figure 6, while PARP cleavage was clearly detected in cells treated with puromycin which induces caspase mediated apoptosis [57], only slight PARP cleavage was seen in melanoma cells treated with MASL (p.0.2 by ANOVA).  These observations suggest that MASL can reduce tumor cell growth by caspase independent necrotic mechanisms. This property may be particularly advantageous when applied to malignancies that resist apoptosis induced by classical cancer agents.

Dietary MASL inhibits melanoma cell migration and tumorigenesis
As described above, some dietary lectins can survive digestive processing to combat cancer in vivo [43][44][45][46]. Serum was taken from mice fed MASL and then examined for its effects on melanoma cell migration in vitro to determine if this dietary lectin can survive gastrointestinal proteolysis to remain biologically active in the circulatory system. As shown in Figure 7, serum from mice fed 100 or 200 mg/kg MASL inhibited melanoma cell migration by about 30% or over 80% compared to serum from mice fed no MASL, respectively. These data indicate that dietary administration of MASL can result in biologically relevant levels of circulating product that are sufficient to inhibit melanoma cell migration. Since MASL, or at least its biological activity, was resistant to gastrointestinal proteolysis we examined the effects of dietary MASL on tumor cell growth in vivo. As shown in Figure 8a, oral administration of 25 mg/kg of MASL once a week inhibited the subcutaneous growth of melanoma cells in mice by approximately 50% (p,0.05 by ANOVA). Moreover, we found no adverse effects on mouse health or physiology over the course of these experiments based on animal behavior, weight, and organ analysis after dissection.
As expected, PDPN expression was evident in melanoma cells of tumors in vivo. As shown in Figure 8b, PDPN expression was more intense in tumor cells from control animals than animals treated with MASL. In addition, tumor cells from MASL treated animals appeared more epithelial in nature, with a relatively flattened morphology and more restricted PDPN staining than tumors from untreated animals.

Dietary MASL inhibits the formation of blood channels lined by melanoma cells
Although its molecular target has not been defined, Mistletoe lectin is used as an anticancer agent that can reduce tumor vascularization in vivo [46,58,59]. Aggressive melanoma cells are known for their ability to form blood vessels lined with tumor cells by a process referred to as ''vasculogenic mimicry'' [60]. Interestingly, as shown in Figure 8b and 8c, these blood filled vascular spaces lined by tumor cells were much smaller and less

Human melanoma cells express PDPN and respond to MASL
We examined cell lines and clinical specimens to further investigate PDPN expression and MASL sensitivity in human melanoma cells. As shown in Figure 9a, PDPN expression was drastically increased in primary and metastatic melanoma specimens from every patient that we examined. As with mouse melanoma cells, PDPN expression was also much higher in human melanoma cell lines than normal skin. In addition, as with mouse cells, MASL significantly suppressed human melanoma cell growth and migration at concentrations of less than 300 nM (Figure 9b  and 9c). Moreover, as with mouse cells, MASL inhibited human melanoma cell motility prior to effecting cell growth. For example, 308 nM MASL inhibited HT-144 cell migration by over 10 fold, but cell growth by less than 40%.

MASL consists of 2 subunits of identical primary amino acid sequence
Lectins are a complex family of proteins. MASL is a tetramer formed by 32 kD and 37 kD subunits. The 32 kD subunit is called Maackia amurensis hemagglutinin (MAH) and the 37 kD subunit is called Maackia amurensis leukoagglutinin (MAL) or Maackia amurensis mitogen (MAM) [52,61,62]. LC-MS-MS analysis shown in Figure 10 indicates that both of these subunits consist of the same primary amino acid sequence that contains a single cysteine toward the carboxyl terminal at position 243. Unique molecular weights and other properties of these subunits are likely due to specific modifications including asparagine glycosylation and deamidation events. The presence of a cysteine residue in MASL has been reported for the 37 kD subunit (MAL) [62]. In contrast, sequences reported for the 32 kD subunit (MAH) do not contain a cysteine, but contain a serine in its place [52,61,62]. Our finding that both 32 kD and 37 kD subunits contain this cysteine explains their dimerization into respective 64 kD and 74 kD subunits.

Discussion
Lectins are promising anticancer agents that can be administered orally [47,50]. For example, dietary Mistletoe lectin has been shown to inhibit lymphoma growth in mice and is used as an adjuvant therapy to treat various forms of cancer in people [43,59]. Plant lectins are resistant to gastrointestinal proteolysis and breakdown by gut bacteria. Wang et al have shown that intact and bioactive peanut agglutinin (PNA) reaches concentrations of approximately 50nM in the serum of people within hours after ingesting 200g of peanuts [63]. Our data indicate that similar levels of lectin, or its bioactive digestion products, may be achieved from ingestion of MASL. These results also indicate that properties of MASL may be used to help develop reagents to prevent or combat melanoma.
Melanoma cell motility contributes to metastatic invasion that causes the vast majority of skin cancer deaths [4]. Robust PDPN expression was found 100% of the 8 melanoma samples examined from cancer patients in this study (see Figure 9a). These data are in contrast to previous studies reporting PDPN expression in less than 20% of spindle cell and other types of melanomas [64,65]. This may be the result of different experimental methods. For example, the NZ1 antibody used here targets an approximately 12 amino acid epitope in the PDPN PLAG domain with a dissociation constant (K D ) below 1 nM, while antibody used for these previous studies targets a larger epitope with lower affinity [66,67]. However, PDPN staining seen in our patient samples may have been due to cancer associated fibroblasts (CAFs), which would not be distinguished from actual melanoma cells by Western blotting. Indeed, PDPN expression in CAFs has been correlated with tumor aggression in cancers including mammary carcinoma [68] and lung adenocarcinoma [69][70][71]. In any case, robust PDPN expression was seen in all of the melanoma cell lines that we examined, indicating that melanoma cells themselves express PDPN (see Figures 4 and 9).
Dietary lectins can block the action of endogenous prometastatic lectins (such as galectins or selectins) to inhibit tumor cell growth, and can be used as medicinal cancer treatments [37,[45][46][47]. While MASL has been shown to decrease viability of different types of transformed cells including melanoma [72,73], its targets and effects on cell migration have not been reported. Our data indicate that MASL be used to target PDPN in order to combat tumor cell growth and migration. Consistent with PDPN acting as a functionally relevant target, these results indicate that MASL inhibits cell motility prior to inhibiting cell viability. For example, 308 nM MASL suppressed melanoma cell migration by over 99%, while inhibiting cell viability by about 20% within the same time period (see Figure 4).
Although some lectins may nonspecifically bind to many glycoproteins, experiments with cloned Maackia amurensis lectin indicate that it can accurately target specific glycoproteins expressed by human cells [74,75]. Our data indicate that MASL, which has a high affinity for O-linked carbohydrate chains containing sialic acid [51,52], targets PDPN to inhibit transformed cell growth and motility at nanomolar concentrations. Interestingly, Maackia amurensis has been used as a medicinal plant for several centuries to treat ailments including cancer in parts of Asia [76,77]. However, as with many traditional medicines, clinical efficacy has been hindered by a lack of mechanistic understanding. Here, we describe how MASL can serve as a potent bioactive plant medicine that targets PDPN to combat cancer.
PDPN expression is relatively low in most normal cells, and found predominantly in tissues including retina, kidney podocytes, lymphatic endothelium, and lung alveolar epithelium [5,78]. Indeed, as shown in Figure 11, PDPN expression in normal tissues is somewhat similar to that of VEGFR2 which serves as a useful chemotherapeutic target [79,80]. Interestingly, PDPN and VEGFR2 are both receptors that promote cell migration and are suppressed during contact normalization by surrounding nontransformed cells in the microenvironment [15].
Our data demonstrate that MASL targets PDPN to inhibit tumor cell growth and motility. However, MASL is bound to target other receptors containing similar sialic acid motifs. In addition, like other reagents, MASL may produce ''off target'' effects in a clinical setting. However, lectin pharmacology is complex and MASL may undergo enzymatic modification to form bioactive compounds in vivo. For example, although legumes including lentils and peanuts contain significant amounts of agglutinating lectins, their ingestion does not result in toxic platelet aggregation [50]. Moreover, MASL caused no notable toxic effects on organ morphology or animal behavior in our studies which followed mice treated for about 3 weeks. As stated above, Maackia amurensis has been used as a medicinal plant for several centuries to treat ailments including cancer in parts of Asia [76,77]. These treatments would result in oral administration of hundreds of mg of MASL per dosage, and have not resulted in reports of toxicity [81]. However, future studies will clearly be needed to further investigate the potential toxic effects of MASL or other lectins that may be considered for clinical use.
Taken together, results from this study suggest that lectins exemplified by MASL could significantly expand our limited arsenal of targeted cancer treatments, particularly anticancer agents that can be administered orally. This paradigm may lead to novel skin cancer treatments, and pave the way to treat other cancers with elevated PDPN expression levels including breast [20,24], glioma [21], and oral cancer [25,26]. However, as stated above, while PDPN served as the model target of this investigation, other receptors with a2,3-sialic acid residues are increased in melanoma and a variety of other cancers, and may also be targeted to some extent by MASL [10,[12][13][14]42].

Ethics Statement
Use of mice to obtain serum was approved by Committee of Animal Experiment in Nagoya University Graduate School of Medicine (Approved ID: 23132). Mouse experimental protocols were approved by the University of Medicine and Dentistry School of Osteopathic Medicine Institutional Animal Care and Use Committee (Approved ID: 10350).

Western blotting
Human tissue collected by the Honest Broker Program was excised and flash frozen in liquid nitrogen at the time of excision during surgery. Western blotting was performed as previously described [15,[84][85][86]. Protein was resolved by SDS-PAGE, transferred to Immobilon-P membranes (Millipore IH1079562), and incubated with antisera specific for PDPN (Sigma NZ1 P0085, Santa Cruz Biotechnology SC23564, or University of Iowa Developmental Studies Hybridoma Bank 8.1.1), active Src kinase (Cell Signaling Technology 2101), PARP (Cell Signaling Technology 9542), MAPK (Cell Signaling Technology 9102), or ß-actin (Sigma A1978). Primary antiserum was recognized by appropriate secondary antiserum conjugated to horseradish peroxidase and detected using Enhanced Chemiluminescence (Millipore WBKLS0500 or Thermoscientific 32106). After blotting, mem- branes were stained with India ink to verify equal loading and transfer.

Affinity precipitation
Cells were lysed in lysis buffer (20 mM Tris.Cl pH 7.4, 150 mM NaCl, 0.5% Triton X-100, 1mM PMSF) on ice for 30 min, clarified by centrifugation, diluted to 1 mg/ml in PBS supplemented with 1mM PMSF and 10 mM MgCl 2 , and incubated with MASL conjugated to agarose beads, or empty beads as controls (Thermo Scientific 20501), on ice for 3 hours. Beads were then washed 4 times with PBS, and eluted in SDS-PAGE sample buffer at 95uC for 5 minutes. Eluted protein was examined along with total cell lysates by Western blotting.

Cell migration and toxicity assays
Cell monolayers were scratched and incubated with MASL for 24 hours to assess cell migration by wound healing assays, which were quantitated as the number of cells that entered an area of the wound indicated in Figure Legends as described [15,84,85]. Alamar blue (Invitrogen DAL1025) was added to cells 24 hours after MASL treatment, and incubated for an additional 8 hours to assay viability, which was detected by fluorescence measurements (ex/em 570/600 nm) as directed by the manufacturer (Invitrogen) in a Tecan GENios microplate fluorescence spectrophotometer. Cells were also stained with 0.2% trypan blue and counted with a hemocytometer to evaluate cytotoxicity. For Transwell migration assays, 600,000 cells were plated in 6-well cluster plates on cell culture inserts with an 8 micron pore size (Transwell-Clear, Costar) and grown for 24 hours. Cells were then released separately from the top of the membrane and the bottom of the membrane. Transwell migration was then quantitated as the percent of cells found on the underside of the membrane as described [15]. All experiments were performed on parallel cultures to control for variations in cell culture conditions.

Ex vivo effects of dietary MASL on cell migration
Mice (C57BL/6) were fed with a 200 mg food pellet containing MASL to achieve doses of 0, 100, or 200 mg/kg 1 hour before blood was taken from subclavian vein. Blood was allowed to coagulate 1 hour at room temperature and clarified by centrifugation for 10 minutes at 20,000 g to obtain serum. This serum was added to cell culture medium (DMEM+10%FBS) to a final concentration of 20%. B16 cell monolayers were scratched, washed with DMEM, and then incubated with these media, or control medium not fortified with mouse serum, for 24 hours to assess their effects on cell migration by wound healing assays, which were quantitated as the number of cells that entered an area of the wound indicated in Figure Legends as described [15,84,85]. Use of mice to obtain serum was approved by Committee of Animal Experiment in Nagoya University Graduate School of Medicine (Approved ID: 23132).

In vivo tumorigenesis
Mice (C57BL/6) were fed with a 200 mg food pellet containing MASL to achieve doses of 0 or 25 mg/kg once weekly, starting the day before subcutaneous inoculation with 100,000 B16 cells in 0.1 ml PBS. Tumor volumes were measured blindly by a trained physician with a single caliper daily. Mice were sacrificed 18 days after inoculation and dissected. Tumors were fixed in formalin, paraffin embedded, sectioned (5 microns), and processed for hematoxylin/eosin staining or immunohistochemistry to detect PDPN expression with 8.1.1 monoclonal antibody as described [15,85,87]. Samples were analyzed on a Zeiss Axiovert microscope with Axiovision software as described [15,85]. Mouse experimental protocols were approved by the UMDNJ-SOM Institutional Animal Care and Use Committee (Approved ID: 10350).

MASL mass spectroscopy
MASL subunits were resolved by reducing and nonreducing SDS-PAGE and stained by Coomassie dye. Bands were then extracted and digested with trypsin. Extracted protein was reduced and alkylated with iodoacetamide, and sequenced using data dependent acquisition by LC-MS/MS on a Thermo Scientific LTQ Orbitrap XL.