Lectin-Mediated Binding of Engineered Lactococcus lactis to Cancer Cells

Lectins have been increasingly utilized as carriers for targeted drug delivery based on their specific binding to glycans located on mammalian cells. This study employed two lectins, B subunit of bacterial Shiga holotoxin (Stx1B) and fungal Clitocybe nebularis lectin (CNL), for surface display on the lactic acid bacterium Lactococcus lactis. The specific adhesion of these engineered, lectin-displaying L. lactis to cancer cells was evaluated. The expression and surface display of both lectins on L. lactis were demonstrated by western blotting and flow cytometry, respectively. MTS assays revealed that recombinant Stx1B had no effect on Caco-2 cell viability at concentrations of ≤25 µg/mL, whereas CNL was non-toxic even at relatively high concentrations of ≤250 µg/mL. Stx1B bound to Caco-2, HT-29 and HeLa cells after 1 h of incubation. CNL bound to Caco-2 cells and recognized several glycoproteins in HT-29 and Caco-2 cell homogenates of which a 70 kDa protein predominated. Confocal microscopy revealed adhesion of Stx1B-displaying L. lactis to HeLa, Caco-2, and, to a lesser extent, HT-29 cells; CNL-displaying L. lactis showed a relatively similar level of adherence to HT-29 and Caco-2 cells. Thus, lectin-displaying L. lactis might serve as a carrier in targeted drug delivery when coupled to a therapeutic moiety.


Introduction
Altered glycosylation patterns and overexpression of specific carbohydrate epitopes are hallmarks of many cancers [1]. Changes in the oligosaccharide structures of tumorassociated glycoproteins or glycolipids include increased N-glycan branching, a higher Oglycan density, and the generation of truncated versions or modification of terminal glycan molecules through sialylation and fucosylation [2]. These alterations can be exploited for targeted therapy, which is one of the goals of precision medicine. Carbohydrate receptors or patterns on the cell surface mediate intercellular interactions [3] and can be recognized by lectins, a heterogeneous group of proteins and glycoproteins with a selective affinity for carbohydrates [4]. Lectins are found in a diversity of organisms ranging from viruses and plants to humans. Human endogenous lectins are involved, through their specific interactions with complex carbohydrates, in numerous physiological and pathological processes, such as intracellular trafficking, recognition processes, cell homing, endocytosis, phagocytosis, and inflammation [5].
Exogenous lectins have been exploited for their directed binding to cell surfaces for targeted cancer therapy, i.e., targeted delivery of anticancer drugs [6][7][8]. For this purpose, a lectin-targeting moiety is conjugated to an anti-cancer agent, such as a monoclonal antibody, peptide, or small chemotherapeutic molecule [9,10]. Since carbohydrate structures are cell monolayer was gently washed with Dulbecco's PBS (DPBS, Gibco). The cells were detached and collected in DPBS, transferred to tubes, and centrifuged at 13,523× g for 20 min 4 • C. Following supernatant aspiration, RIPA lysis buffer (50 mM Tris/HCl pH 8.0, 150 mM NaCl, 1% Triton-100, 0.5% Na-deoxycholate, 0.1% SDS, 1 mM EDTA) with protease inhibitor was added to the pellet and incubated for 30 min on ice. The pellet was then centrifuged at 16,000× g for 20 min at 4 • C, and the protein in the supernatant was collected and stored at −80 • C. Protein concentration was determined by DC Protein Assay (Bio-Rad), and 30 µg of proteins were loaded onto the gel. SDS-PAGE was performed with a Mini-Protean II apparatus (Bio-Rad). Samples were mixed with 2× Laemmli sample buffer and dithiothreitol, and denatured by heating to 100 • C before loading [36]. The Page Ruler Plus (Thermo Fisher Scientific) pre-stained standards were used for molecular weight comparisons. The proteins were transferred to nitrocellulose membranes (GE Healthcare Life Sciences, Marlborough, MA, USA) using semi-dry transfer with a protocol for 1.5 mm gels (Trans-Blot Turbo Blotting System; Bio-Rad). The membrane was blocked with 5% skim dried milk in Tris-buffered saline (TBS) with 0.05% Tween-20 (TBST; 50 mM Tris-HCl, 150 mM NaCl, 0.05% Tween 20, pH 7.5) and incubated overnight at 4 • C with recombinant E. coli-expressed CNL [29] diluted at 1:1200 (5 µg/mL) in 5% skim dried milk in TBST. Following three washes with TBST, the membrane was incubated for 1 h with rabbit polyclonal anti-CNL (1:2000) in 5% skim dried milk in TBST. The detection was performed using horseradish peroxidase-conjugated goat anti-rabbit IgG (1:5000, Jackson ImmunoResearch) and Lumi-Light chemiluminescent reagent (Roche). Images were acquired using a ChemiDoc MP imaging system (Bio-Rad). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as a loading control and was detected by rabbit monoclonal anti-GAPDH (1:5000; Proteintech).

Immunofluorescence Staining
CNL [30] at concentrations of 0.1, 1, and 10 µg/mL was added to the Caco-2-coated coverslips. Following 1 h of incubation, cells were washed with PBS, fixed and permeabilized in 4% paraformaldehyde in PBS for 20 min, and incubated in 0.1% Triton X-100 in PBS for 5 min. Non-specific staining was blocked with 1% bovine serum albumin in PBS for 1 h. CNL was detected by incubating cells with 20 µg/mL of affinity-purified rabbit anti-CNL primary antibody [30] for 1 h followed by a subsequent incubation with the secondary Alexa Fluor 555-conjugated goat anti-rabbit antibody (1:1000; Life Technologies, Carlsbad, CA, USA) for 1 h in 1% bovine serum albumin in PBS. After each step, the cells were washed three times with PBS. Cells incubated with primary and secondary antibodies in the absence of CNL were used as controls for nonspecific binding. Immunostained cells were mounted onto slides with the DAPI-containing mounting agent and visualized by confocal microscopy.

Confocal Microscopy
The slides were imaged with a confocal microscope (LSM-710, Carl Zeiss, Oberkochen, Germany). Images were collected using a 63× immersion oil objective with settings to detect brightfield, DAPI, Alexa 488, and Alexa 647. Images were prepared using the ImageJ software.

Cell Viability Assay
The effect of recombinant lectins on Caco-2 cell viability was assessed with MTS colorimetric assay as described previously [37]. Caco-2 cells were seeded onto 96-well culture plates (3 × 10 4 cells/well) and treated with increasing concentrations (0.1, 1, 10, 25, 100, and 250 µg/mL) of Stx1B [24] and CNL [30] in complete medium. For controls, medium devoid of lectins was added. The cells were incubated for 24 h and 48 h at 37 • C. Their viability was assessed using the CellTiter 96 Aqueous One Solution Cell Proliferation Assay (Promega, Madison, WI, USA) according to the manufacturer's instructions. Absorbance was measured with a microplate reader (Tecan) at a wavelength of 492 nm. The experiment was performed in triplicate, and the results were normalized to the controls (mean values of treated versus untreated cells).

Bacterial Strains and Growth Conditions
The bacterial strains used in this study are listed in Table 1.   The lectin genes stx1B (GenBank accession number: CP050498.1) and cnl (UniProt accession number: B2ZRS9, L. lactis codon-optimized, encoding non-dimerizing mutant) were amplified from pET28-Stx1B [24] and gBlock (IDT), respectively, by PCR using the primers specified in Table 1. Amplicons were cloned into the pGEM-T Easy plasmid and then transferred to the plasmid pSDBA3b via the BamHI/EcoRI restriction sites to yield pSD-Stx1B and pSD-CNL. FLAG-tag (24 bp) was inserted via NcoI/BamHI as described previously [42]. The whole cassette encoding the spUsp45-Stx1B-acmA3b fusion protein was transferred to the first multiple cloning site (MCS 1) in the dual promoter plasmid pNZDual via the NcoI/XbaI restriction sites. Finally, IRFP was cloned to the second multiple cloning site (MCS 2) of the plasmid pNZDual via the NdeI/XhoI restriction sites, yielding pNZD-Stx1B-IRFP. The cassette encoding the spUsp45-CNL-acmA3b fusion protein was similarly transferred to the MCS 1 in the dual promoter plasmid with IRFP in the MCS 2, yielding pNZD-CNL-IRFP.

The Expression of Stx1B and CNL Fusion Proteins in L. lactis
Overnight cultures of L. lactis harboring the appropriate plasmids were diluted (1:100) in 10 mL of fresh GM-17 medium containing chloramphenicol (10 µg/mL) and biliverdin (15.5 µg/mL), and grown to an optical density of A600 = 0.8-1.0. Fusion protein expression was induced with nisin (25 ng/mL; Fluka Chemie AG, Buchs, Switzerland) [38]. After 3 h of incubation at 30 • C, 1 mL of the cultures was stored at 4 • C for flow cytometry analysis. For SDS-PAGE analysis, the remaining cell culture was centrifuged at 5000× g for 10 min, and the cell pellets were resuspended in 400 µL of phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 12.5 mM Na 2 HPO 4 , 1.98 mM KH 2 PO 4 , pH 7.4) and stored at −20 • C. To evaluate binding to cancer cells, L. lactis cultures were centrifuged at 5000× g for 10 min, washed twice with PBS, resuspended in PBS to an optical density of A 600 = 0.8, and stored at 4 • C. Before the adhesion assay, L. lactis were resuspended in RPMI 1640 medium with L-glutamine and HEPES (Lonza, Basel, Switzerland).

Fluorescence Measurements of IRFPs
Aliquots of cell cultures (200 µL) with an optical density of A600 = 0.8 were transferred to black, flat-bottom 96-well plates (Greiner, Kremsmünster, Austria). Fluorescence was measured on an Infinite M1000 microplate reader (Tecan, Männedorf, Switzerland), with excitation/emission at 690 nm/713 nm. Two technical replicates of the measurements were performed.

Flow Cytometry
L. lactis cultures (20 µL) in the stationary growth phase were added to 500 µL of TBS (50 mM Tris-HCl, 150 mM NaCl, pH 7.5) and centrifuged at 5000× g for 5 min at 4 • C. The pellets were resuspended in 500 µL of TBS containing rabbit polyclonal anti-FLAG-tag (1:500; Proteintech) or anti-CNL (1:1000) for the detection of Stx1B and CNL, respectively [43]. After 2 h of incubation at room temperature (RT) with constant shaking at 100 rpm, the cells were washed three times with 200 µL of TBS with 0.1% Tween-20 (0.1% TBST) and resuspended in 500 µL TBS containing an Alexa Fluor 488-labelled anti-rabbit antibody (1:2000; Cell Signaling Technology). After 2 h of incubation at RT with constant shaking at 100 rpm, the cells were washed three times with 200 µL of 0.1% TBST and finally resuspended in 500 µL TBS. The samples were analyzed using a flow cytometer (FACS Calibur; Becton Dickinson, Franklin Lakes, NJ, USA) with excitation/emission at 488/530 nm in the FL1 channel. The geometric mean fluorescence intensity of at least 20,000 L. lactis cells in the appropriate gate was measured.

L. lactis Cell Adhesion Assay
HeLa, HT-29, and Caco-2 cells were seeded onto sterilized coverslips (8 mm diameter #1.5) in 24-well plates. The seeding concentrations were determined to reach the desired confluence (5 × 10 4 /well for HeLa, 1 × 10 5 /well for HT-29, and 1.5 × 10 5 /well for Caco-2). After 48 h, the medium was aspirated, and 500 µL of the L. lactis culture (at A 600 = 0.8) in RPMI was added to each well. The cells were incubated with L. lactis cultures for 2 h at 37 • C with 5% CO 2 . Following incubation, cells were gently washed twice with PBS to remove unattached L. lactis, fixed in 4% paraformaldehyde in PBS for 20 min at RT, washed twice with PBS, mounted with the DAPI-containing mounting agent, and visualized by confocal microscopy.

Statistical Analyses
Statistical analyses were performed using the GraphPad Prism 6 software. The data are presented as mean ± standard deviation. Student's t-tests were used to assess significant differences between the lectin-displaying L. lactis and respective controls.

CNL-Binding Glycoproteins in HT-29 and Caco-2 Whole-Cell Lysates
The presence of the CNL-binding glycoproteins in HT-29 and Caco-2 cells was determined by exposing the blot of whole-cell lysates to CNL and performing detection with anti-CNL antibodies. In both HT-29 and Caco-2 cell lysates, CNL bound to several glycosylated proteins, of which a protein with a molecular weight of 70 kDa exhibited the most CNL binding (Figure 1). A protein of similar size was previously detected in HT-29 cells with the LacdiNAc-specific lectin Wisteria floribunda agglutinin [44].
The presence of the CNL-binding glycoproteins in HT-29 and Caco-2 cells was determined by exposing the blot of whole-cell lysates to CNL and performing detection with anti-CNL antibodies. In both HT-29 and Caco-2 cell lysates, CNL bound to several glycosylated proteins, of which a protein with a molecular weight of 70 kDa exhibited the most CNL binding (Figure 1). A protein of similar size was previously detected in HT-29 cells with the LacdiNAc-specific lectin Wisteria floribunda agglutinin [44].

Binding of Recombinant Lectins to HeLa, HT-29, and Caco-2 Cancer Cells
The FITC-labelled recombinant protein Stx1B (Stx1B-FITC) was tested for binding to HeLa, HT-29, and Caco-2 cells. Compared to the control, significant binding of Stx1B-FITC was observed, confirming the presence of its corresponding carbohydrate ligand Gb3 in all the cell lines included in the study. The amount of Stx1B-FITC binding to HeLa cells was larger than that to HT-29 and Caco-2 cells (Figure 2A). The binding of recombinant CNL to Caco-2 cells was analyzed with immunocytochemical staining using CNL-specific antibodies. Treatment of Caco-2 cells with CNL at concentrations of 0.1, 1, and 10 µg/mL resulted in dose-dependent CNL binding. After 1 h of incubation, a significant number of Caco-2 cells bound CNL ( Figure 2B), and this increased substantially after 48 h of incubation ( Figure 2C). No signal was detected in untreated cells labelled with antibody alone.

Binding of Recombinant Lectins to HeLa, HT-29, and Caco-2 Cancer Cells
The FITC-labelled recombinant protein Stx1B (Stx1B-FITC) was tested for binding to HeLa, HT-29, and Caco-2 cells. Compared to the control, significant binding of Stx1B-FITC was observed, confirming the presence of its corresponding carbohydrate ligand Gb3 in all the cell lines included in the study. The amount of Stx1B-FITC binding to HeLa cells was larger than that to HT-29 and Caco-2 cells (Figure 2A). The binding of recombinant CNL to Caco-2 cells was analyzed with immunocytochemical staining using CNL-specific antibodies. Treatment of Caco-2 cells with CNL at concentrations of 0.1, 1, and 10 µg/mL resulted in dose-dependent CNL binding. After 1 h of incubation, a significant number of Caco-2 cells bound CNL ( Figure 2B), and this increased substantially after 48 h of incubation ( Figure 2C). No signal was detected in untreated cells labelled with antibody alone.

The Effect of Stx1B and CNL Lectins on Caco-2 Cell Viability
The effect of lectins on Caco-2 cell viability was evaluated after 24 and 48 h using the MTS assay. Recombinant protein Stx1B [24] at concentrations of up to 25 µg/mL did not affect cell viability, while higher concentrations decreased the number of viable cells ( Figure 3A). Conversely, CNL at concentrations of up to 250 µg/mL exerted no effect on cell viability after 24 and 48 h of exposure ( Figure 3B). Of note, the cells were slightly more viable in the presence of CNL; this is in line with experiments analyzing monomeric CNL cytotoxicity in Jurkat human leukemic T cells [30].

Genetic Constructs for Stx1B and CNL Lectin Display on L. lactis
Two fusion genes were constructed to display lectins on the surface of L. lactis. The genes for the lectins Stx1B and CNL were fused with the gene for the Usp45 secretion signal at the 5 -end and with the gene for the non-covalent cAcmA surface anchor at the 3 -end, as described previously [45]. The gene for the FLAG-tag was added for Stx1B detection. The genes were cloned under the control of the NisA promoter into our previously reported dual promoter plasmid pNZDual [41], with two multiple cloning sites (MCS 1 and MCS 2) to enable the simultaneous expression of two fusion proteins. The gene for IRFP was  Table 1 and shown schematically in Figure 4.

The Effect of Stx1B and CNL Lectins on Caco-2 Cell Viability
The effect of lectins on Caco-2 cell viability was evaluated after 24 and 48 h using the MTS assay. Recombinant protein Stx1B [24] at concentrations of up to 25 µg/mL did not affect cell viability, while higher concentrations decreased the number of viable cells ( Figure 3A). Conversely, CNL at concentrations of up to 250 µg/mL exerted no effect on cell viability after 24 and 48 h of exposure ( Figure 3B). Of note, the cells were slightly more viable in the presence of CNL; this is in line with experiments analyzing monomeric CNL cytotoxicity in Jurkat human leukemic T cells [30].

Expression of Lectin Fusion Proteins in L. lactis
Both lectins Stx1B (Uniprot ID: Q7DH26) and CNL (UniProt ID: B2ZRS9) have low molecular weights (Stx1B, 7.7 kDa; CNL, 15 kDa) and are therefore suitable for L. lactis expression and surface display. The expression of Stx1B and CNL in fusion with Usp45 secretion signal and cAcmA anchoring domain was confirmed in L. lactis cell lysates with SDS-PAGE followed by western blot analysis ( Figure 5A) using the anti-FLAG and anti-CNL antibodies, respectively. The size of Stx1B fused to the cAcmA anchoring domain was calculated to be around 34 kDa (FLAG-tag ∼1 kDa, Stx1B ∼7.7 kDa, and cAcmA ∼25 kDa), which corresponds to the size detected in cell lysates of L. lactis harboring pNZD-Stx1B-IRFP ( Figure 5). The size of CNL fused to the cAcmA anchoring domain was calculated to be around 40 kDa (CNL monomer ∼15.9 kDa and cAcmA ∼25 kDa), which is somewhat smaller than the size (∼45 kDa) detected in cell lysates of L. lactis harboring pNZD-CNL-IRFP ( Figure 5). No bands were detected in control L. lactis harboring IRFPcontaining plasmid or empty plasmid (pNZ8148). The concomitant expression of IRFP was confirmed by fluorescence intensity measurements in both lectin-displaying L. lactis strains ( Figure 5B). The level of IRFP co-expressed in the Stx1B-displaying strain was somewhat lower than that expressed in the control strain. Conversely, the CNL-displaying strain expressed a similar amount of IRFP as the control strain.

Genetic Constructs for Stx1B and CNL Lectin Display on L. lactis
Two fusion genes were constructed to display lectins on the surface of L. lactis. The genes for the lectins Stx1B and CNL were fused with the gene for the Usp45 secretion signal at the 5′-end and with the gene for the non-covalent cAcmA surface anchor at the 3′-end, as described previously [45]. The gene for the FLAG-tag was added for Stx1B detection. The genes were cloned under the control of the NisA promoter into our previously reported dual promoter plasmid pNZDual [41], with two multiple cloning sites (MCS 1 and MCS 2) to enable the simultaneous expression of two fusion proteins. The gene for IRFP was included in MCS 2 to enable the visualization of the L. lactis. The fusion genes in pNZDual are listed in Table 1 and shown schematically in Figure 4.

Expression of Lectin Fusion Proteins in L. lactis
Both lectins Stx1B (Uniprot ID: Q7DH26) and CNL (UniProt ID: B2ZRS9) have low molecular weights (Stx1B, 7.7 kDa; CNL, 15 kDa) and are therefore suitable for L. lactis expression and surface display. The expression of Stx1B and CNL in fusion with Usp45 secretion signal and cAcmA anchoring domain was confirmed in L. lactis cell lysates with SDS-PAGE followed by western blot analysis ( Figure 5A) using the anti-FLAG and anti-CNL antibodies, respectively. The size of Stx1B fused to the cAcmA anchoring domain was calculated to be around 34 kDa (FLAG-tag ∼1 kDa, Stx1B ∼7.7 kDa, and cAcmA ∼25 kDa), which corresponds to the size detected in cell lysates of L. lactis harboring pNZD-Stx1B-IRFP ( Figure 5). The size of CNL fused to the cAcmA anchoring domain was calculated to be around 40 kDa (CNL monomer ∼15.9 kDa and cAcmA ∼25 kDa), which is somewhat smaller than the size (∼45 kDa) detected in cell lysates of L. lactis harboring pNZD-

Genetic Constructs for Stx1B and CNL Lectin Display on L. lactis
Two fusion genes were constructed to display lectins on the surface of L. lactis. The genes for the lectins Stx1B and CNL were fused with the gene for the Usp45 secretion signal at the 5′-end and with the gene for the non-covalent cAcmA surface anchor at the 3′-end, as described previously [45]. The gene for the FLAG-tag was added for Stx1B detection. The genes were cloned under the control of the NisA promoter into our previously reported dual promoter plasmid pNZDual [41], with two multiple cloning sites (MCS 1 and MCS 2) to enable the simultaneous expression of two fusion proteins. The gene for IRFP was included in MCS 2 to enable the visualization of the L. lactis. The fusion genes in pNZDual are listed in Table 1 and shown schematically in Figure 4.

Expression of Lectin Fusion Proteins in L. lactis
Both lectins Stx1B (Uniprot ID: Q7DH26) and CNL (UniProt ID: B2ZRS9) have low molecular weights (Stx1B, 7.7 kDa; CNL, 15 kDa) and are therefore suitable for L. lactis expression and surface display. The expression of Stx1B and CNL in fusion with Usp45 secretion signal and cAcmA anchoring domain was confirmed in L. lactis cell lysates with SDS-PAGE followed by western blot analysis ( Figure 5A) using the anti-FLAG and anti-CNL antibodies, respectively. The size of Stx1B fused to the cAcmA anchoring domain was calculated to be around 34 kDa (FLAG-tag ∼1 kDa, Stx1B ∼7.7 kDa, and cAcmA ∼25 kDa), which corresponds to the size detected in cell lysates of L. lactis harboring pNZD-Stx1B-IRFP ( Figure 5). The size of CNL fused to the cAcmA anchoring domain was calculated to be around 40 kDa (CNL monomer ∼15.9 kDa and cAcmA ∼25 kDa), which is somewhat smaller than the size (∼45 kDa) detected in cell lysates of L. lactis harboring pNZD-

Surface Display of Lectins on L. lactis
The statistically significant surface display of lectins on L. lactis was confirmed with flow cytometry. A characteristic shift in mean fluorescence intensity was observed for Stx1B-displaying and CNL-displaying strain in comparison to control L. lactis harboring empty plasmid or pNZD-IRFP ( Figure 6). Flow cytometry also verified that lectins were accessible on the L. lactis cell surface.

The Adhesion of Lectin-Displaying L. lactis to Cancer Cells
The binding of Stx1B-displaying L. lactis to cancer cells was analyzed. A large number of Stx1B-displaying L. lactis adhered to HeLa cells. A significant number of Stx1B-displaying L. lactis clusters also adhered to Caco-2 cells, whereas only a few L. lactis clusters interacted with HT-29 cells ( Figure 7A). The ability of CNL-displaying L. lactis to adhere to HT-29 and Caco-2 cells was also examined. A modest number of CNL-displaying L. lactis directly interacted with HT-29 and Caco-2 cells grown in the form of islets ( Figure 7B). Most of the L. lactis adhered to the outer edges of the larger islets, while few L. lactis adhered to the inner areas of the islets. Some individual cancer cells and small islets appeared surrounded by the adhered CNL-displaying L. lactis. No L. lactis were visible in control cell cultures incubated with IRFP-expressing L. lactis, which demonstrates that L. lactis without lectin displayed on their surface do not bind to cells.
CNL-IRFP ( Figure 5). No bands were detected in control L. lactis harboring IRFP-containing plasmid or empty plasmid (pNZ8148). The concomitant expression of IRFP was confirmed by fluorescence intensity measurements in both lectin-displaying L. lactis strains ( Figure 5B). The level of IRFP co-expressed in the Stx1B-displaying strain was somewhat lower than that expressed in the control strain. Conversely, the CNL-displaying strain expressed a similar amount of IRFP as the control strain.

Surface Display of Lectins on L. lactis
The statistically significant surface display of lectins on L. lactis was confirmed with flow cytometry. A characteristic shift in mean fluorescence intensity was observed for Stx1B-displaying and CNL-displaying strain in comparison to control L. lactis harboring empty plasmid or pNZD-IRFP ( Figure 6). Flow cytometry also verified that lectins were accessible on the L. lactis cell surface.

The Adhesion of Lectin-Displaying L. lactis to Cancer Cells
The binding of Stx1B-displaying L. lactis to cancer cells was analyzed. A large number of Stx1B-displaying L. lactis adhered to HeLa cells. A significant number of Stx1B-displaying L. lactis clusters also adhered to Caco-2 cells, whereas only a few L. lactis clusters interacted with HT-29 cells ( Figure 7A). The ability of CNL-displaying L. lactis to adhere to HT-29 and Caco-2 cells was also examined. A modest number of CNL-displaying L. lactis directly interacted with HT-29 and Caco-2 cells grown in the form of islets ( Figure 7B). Most of the L. lactis adhered to the outer edges of the larger islets, while few L. lactis adhered to the inner areas of the islets. Some individual cancer cells and small islets appeared surrounded by the adhered CNL-displaying L. lactis. No L. lactis were visible in HT-29 and Caco-2 cells was also examined. A modest number of CNL-displaying L. lactis directly interacted with HT-29 and Caco-2 cells grown in the form of islets ( Figure 7B). Most of the L. lactis adhered to the outer edges of the larger islets, while few L. lactis adhered to the inner areas of the islets. Some individual cancer cells and small islets appeared surrounded by the adhered CNL-displaying L. lactis. No L. lactis were visible in control cell cultures incubated with IRFP-expressing L. lactis, which demonstrates that L. lactis without lectin displayed on their surface do not bind to cells.

Discussion
In this study, we utilized lectins to target L. lactis to cancer cells via their interaction with glycans on the cancer cells' surface. Various cancer tissues produce oligosaccharides that differ from the glycosylation patterns in non-malignant tissue. Lectins possess a high level of specificity for the tumor-associated carbohydrates and are therefore considered for selective delivery of anticancer agents to tumors. For this purpose, we constructed two recombinant L. lactis strains that displayed the Gb3-recognizing lectin Stx1B and LacDiNAc-recognizing lectin CNL on their surface. We demonstrated adhesion of these lectin-displaying L. lactis onto HT-29, Caco-2, and HeLa cancer cells.
First, the colorectal cancer cell lines HT-29 and Caco-2 were tested for the expression of the Gb3 receptor (for Stx1B), while HeLa cells were included as a control, as they are known to express Gb3 and bind Stx1B. We confirmed successful Stx1B-FITC binding to all three cell lines tested. HeLa cells exhibited the largest amount of binding, which is in agreement with the previously observed high expression of the Gb3 receptor in this cell line [46,47]. Less binding was observed to the tested colorectal cell lines, which is also in accordance with previously reported data [48,49].
Unlike for the Gb3 receptor, data on the expression of the CNL target LacdiNAc in HT-29 and Caco-2 cells is scarce. In HT-29 cells, LacdiNAc has been demonstrated to be the ligand for macrophage galactose-type lectin [50] and Wisteria floribunda agglutinin lectin [51], whereas the presence and identity of CNL-binding glycoproteins remain largely unknown. We, therefore, performed western blotting of HT-29 and Caco-2 whole-cell lysates, probed them with CNL, and analyzed the interaction of CNL with glycoproteins from the lysates. Multiple bands were detected, suggesting that CNL recognizes more than one glycoprotein in the lysate of tested cells. Although their identity is not yet known, we found that a 70 kDa glycoprotein, which exhibited the highest CNL binding, was also previously detected in HT-29 cells with the LacdiNAc-specific lectin Wisteria floribunda agglutinin [45]. We further examined whether CNL binds to intact Caco-2 cells via its target glycoproteins. Immunocytochemical analysis with specific anti-CNL antibodies showed successful dose-and time-dependent accumulation of CNL on the surface of Caco-2 cells. The substantial binding was achieved at 1 µg/mL of CNL. However, the selectivity of CNL for Caco-2 and HT-29 cells could not be assessed due to the lack of appropriate control (cell line with confirmed absence of LacdiNAc).
To test the effect of the recombinant lectins Stx1B and CNL on Caco-2 cell viability, we performed a colorimetric MTS assay. Cell viability was assessed after 24 and 48 h incubations with increasing concentrations of the recombinant lectins. CNL has been previously shown to have an antiproliferative effect on human leukemic T lymphocytes (Jurkat cells), decreasing cell viability by >50% at the highest concentration (100 µg/mL) [30]. The effect was dependent on bivalent binding of homodimeric CNL to cell-surface carbohydrates. In this study, we used a non-dimerizing CNL, which caused no effect on Caco-2 cell viability, even at the relatively high concentration of 250 µg/mL. Although limited to 48 h incubation, our results substantiate previous findings regarding a lack of cytotoxicity of non-dimerizing CNL mutant against Jurkat cells [30] and reinforce the conclusion that monomeric CNL is generally non-toxic. Conversely, recombinant Stx1B decreased cell viability at concentrations above 25 µg/mL, consistent with a previous observation that Stx1B is capable of inducing apoptosis even in the absence of the toxic subunit StxA [11]. This is also in agreement with the measured apoptotic effect in HeLa and Vero cells in which recombinant Stx1B concentrations of >20 µg/mL significantly decreased cell viability [10].
After verifying the presence of lectins' target sites on HeLa, HT-29 and Caco-2 cells, we displayed Stx1B and CNL, individually, on the surface of L. lactis and assessed the adhesion of engineered bacteria on target cancer cells. To achieve surface display on L. lactis, Stx1B and CNL were fused with an Usp45 secretion signal and cAcmA anchoring domain. IRFP was co-expressed for the purpose of detection and visualization. A significant level of Stx1B and CNL expression and surface display on L. lactis was verified by western blotting and flow cytometry, respectively. Multiple bands, observed by western blotting, were probably the consequence of cAcmA degradation and could be prevented by using L. lactis NZ9000∆htrA strain as previously reported [22]. The co-expression of IRFP was confirmed by fluorescence measurements.
Finally, we evaluated the binding of engineered Stx1B and CNL lectin-displaying L. lactis to cancer cells. For Stx1B-displaying L. lactis, adhesion assays were performed on HeLa, HT-29, and Caco-2 cells. While only a few Stx1B-displaying L. lactis cells adhered to HT-29 cells, many L. lactis clusters attached to the surface of Caco-2 and HeLa cells. The extent of L. lactis binding was greater in Caco-2 cells than that in HT-29 cells, which is surprising as HT-29 cells were reported to express more Gb3 than Caco-2 cells [49]. Nevertheless, the binding of Stx1B to Caco-2 cells has already been demonstrated [27], which substantiates our observations. The binding of CNL-displaying L. lactis to HT-29 and Caco-2 cells was also assessed. CNL-displaying L. lactis adhered to HT-29 and Caco-2 cells. As a control, L. lactis that did not display lectins was used and did not adhere to cancer cells, confirming that the binding to the cancer cell surface is lectin-mediated. More Stx1B-displaying L. lactis adhered to Caco-2 cells than to HT-29 cells; conversely, more CNLdisplaying L. lactis adhered to HT-29 cells than to Caco-2 cells. These observed differences in cell specificity might be attributed to the different expression levels of target receptors on the surface of cancer cells. Apart from that, efficient ligand-receptor interaction for active targeting depends on variety of factors, including the availability of the receptor on the cell surface, the rate of internalization over shedding of the surface receptor following ligand binding and the affinity of the ligand for the receptor [52]. Overall, both Stx1Band CNL-displaying L. lactis showed a weak to moderate level of adhesion to HT-29 and Caco-2 cancer cells in comparison to the high level of adhesion of Stx1B-displaying L. lactis to HeLa cells. Nevertheless, this study has demonstrated that lectin-mediated targeting of bacteria to the cancer cells is feasible. Moreover, a further increase in cell adhesion could be accomplished by testing additional lectins or by displaying multiple lectins, including those with cytoadhesive characteristics, such as wheat germ agglutinin that was shown to improve association of drug-loaded polymeric nanospheres to Caco-2 cells [53]. The engineered, lectin-functionalized, bacteria would be particularly suitable for delivery of therapeutic proteins to the gastrointestinal tract via oral route for the treatment of colorectal cancer, thanks to the ability of L. lactis to survive harsh conditions in gastrointestinal tract, as well as the stability of lectins in gastro-intestinal fluids and their resistance to degradation by digestive processes [54].

Conclusions
This is the first report on lectin-displaying bacteria engineered for glycan-targeting cancer therapy. Two lectins that target tumor glycoproteins were displayed on the surface of L. lactis. The ability of lectin-displaying L. lactis to specifically adhere to the surface of cancer cells and the ability of recombinant lectins to bind to these cells was demonstrated. Taken together, our results suggest that lectins, displayed on L. lactis, can achieve active targeting of cancer cells via their interactions with glycans on the surface of cancer cells. The use of lectins as potential drug carriers has been previously shown, and this study demonstrates that their applicability could be broadened by their combined use with live food-grade bacteria.