Several carcinoma, melanoma, lymphoma and leukemia cell lines are more sensitive to CAPs compared to normal cells [46–49]. It is believed that this selectivity is due to a more negatively charged cell surface of the cancer cells. However, we have previously shown that negatively charged GAGs on the cell surface inhibit the cytotoxic activity of CAPs, probably by sequestering the peptides away from the phospholipid bilayer. In the present study three peptides consisting of only 9 amino acids, and with a net positive charge of +6, were tested for their antitumor activity and selectivity. Compared to LfcinB, the 9-mers were more active and killed cancer cells more effectively, showing that the 9-mers are more optimized for antitumor activity than LfcinB. By examining the role of cell surface GAGs on the cytotoxic effect of the 9-mers, we found that cell surface GAGs had a different effect on the cytotoxic activity of this new generation of shorter peptides compared to what we previously reported for the longer naturally occurring LfcinB (25-mer) peptide and the KW5 (21-mer) peptide.
All the three 9-mer peptides displayed a higher activity towards the lymphoma, carcinoma and neuroblastoma cell lines compared to normal endothelial cells, fibroblasts and red blood cells. One exception was the lower activity of LTX-315 against the carcinoma cell lines compared to the endothelial cells.
LTX-315 killed the various tumor cells more efficiently than LTX-302 and LTX-318. However, the LTX-318 and LTX-302 peptides displayed a higher specificity for the tumor cells versus the non-tumor endothelial and fibroblast cells than did LTX-315. These findings are in agreement with our earlier findings that enhanced antitumor activity may result in reduced tumor cell specificity [50, 51].
The relatively higher cytotoxic activity against the lymphoma and neuroblastoma cells compared to the endothelial cells, the fibroblast cells and the red blood cells suggest that differences at the cellular membrane level decide their vulnerability to the peptides. Differences in cell membrane composition, fluidity  and surface area [53, 54] between cancer cells and normal cells may be factors that make the former cells more susceptible to the peptides.
The lack of correlation between the cytotoxic activity of the peptides and the expression of HS on the cell surface of the lymphoma cells indicates that membrane components other than HS affect the susceptibility of the lymphoma cells against the 9-mers. The cell lines that displayed the highest sensitivity against the peptides also had the highest amount of cell-associated CS. It can therefore be speculated if CS is involved in the cytotoxic effect of the peptides. However, the correlation between cell-associated CS and cytotoxicity was not significant.
Both the expression of sialic acids, which is another component of the anionic glycoconjugate cell coat that surrounds cells, and the expression of PS in the outer membrane leaflet have been shown to affect the CAPs interactions with the lipid bilayer [11, 55]. Moreover, the membrane fluidity has been demonstrated to be an important determinant for the selective permeabilization of membranes [56–58].
In order to study the possible contribution of HS to the cytotoxic activity of the 9-mers more directly, the peptides cytotoxic activity was tested against CHO wild-type cells expressing HS on the cell surface and its mutant lacking HS on the cell surface. CHO cells have been widely used to study the role of cell surface GAGs in various processes such as viral infection, growth factor signaling and cell adhesion . The pgsA-745 cells have defective xylosyltransferase, an enzyme necessary for biosynthesis of HS and CS . Although CHO cells are derived from normal tissue, both CHO-K1 and pgsA-745 induce solid tumors when injected into immunodeficient mice [60, 61]. By examining the expression pattern of GAGs on the cell surface of the CHO-K1 cells, we found that the cell surface PGs primarily contained HS chains. This expression profile, in which HS is the dominant type of cell surface GAGs, is common among most cell types . Our experiments with CHO cells clearly indicate that cell surface GAGs increase the cytotoxic effect of LTX-302 and LTX-318. However, the cytotoxic effect of LTX-315, which lysed the cells more efficiently, was not influenced by cell surface GAGs.
We found that soluble CS and HS inhibited the cytotoxic activity of LTX-302 and LTX-315 against the CHO cells. The stronger inhibition of the cytotoxic activity obtained by HS compared to CS indicates that the peptides bound more strongly to HS than to CS. This was confirmed by affinity chromatography, which exhibited a higher affinity of the peptides to HS compared to CS. The difference in the affinity could be explained by the higher conformational flexibility in HS compared to the more rigid CS , as the peptides may require a high flexibility in the molecules they bind to. The cytotoxic activity of LTX-318 was not affected by the presence of soluble CS or HS. Considering the low activity that LTX-318 displayed against the CHO-K1 cells, the cytotoxic concentration of the peptide might be too high in order for the amount of exogenous CS and HS to affect the activity. The finding that the peptides interact more strongly with HS, together with the higher amount of HS chains attached to syndecans and glypicans compared to CS [27, 28], strongly indicates that HS and not CS is the major interaction site for the 9-mers.
Despite having the same net positive charge, LTX-315 and LTX-318 showed a higher affinity for HS in comparison to LTX-302. The difference in affinity to HS may be due to the position of the basic residues in the peptides. In addition to cationic residues, the CAPs include lipophilic residues, which are important for interactions with the lipid layer of the cell membrane leading to an irreversible membrane destabilizing effect. The relative positions of the lipophilic and cationic residues affect the flexibility of CAPs, which permit the transition from its solution conformation to its membrane-interacting conformation [63, 64]. Both the position of the cationic residues and the relative flexibility of the three 9-mers can therefore affect their interaction with cell surface GAGs.
The ability of the 9-mer peptides and LfcinB to interact with GAG chains will increase the cell surface concentration of the peptides. However, the finding that cell surface HS can act as a facilitator for small CAPs is in contrast to our recent report which shows that the longer lytic peptides LfcinB and KW5 displayed a higher cytotoxic activity against the GAG-deficient cell line . The inhibitory effect of GAGs on the cytotoxic activity of LfcinB could be due to the higher affinity for HS compared to the 9-mer peptides. The LfcinB peptide has a higher net positive charge (+8) than the 9-mers, which may explain its higher affinity for HS. However, it has been documented that the affinity for HS is only partly correlated with the net charge of the peptides [45, 65]. Several studies have demonstrated that peptide analogues with arginine residues bind more tightly to heparin-like molecules than comparable analogues substituted with lysine [65–68]. The 9-mers have no arginine residues in their sequences, while the LfcinB peptide contains five arginine residues. It is believed that the tighter interaction observed for arginine is due to a strong hydrogen bond formation between the guanidine group of arginine and sulfate. The presence of arginine residues in LfcinB might therefore also contribute to the higher affinity for HS compared to the 9-mers. The difference in the size of the 9-mers and LfcinB peptides could also affect the affinity for HS due to differences in the flexibility of the secondary structure. Whereas the LfcinB peptide forms a stabilized amphiphatic β-sheet, the smaller peptides might have a higher plasticity of their secondary structure, thus leading to a less defined binding domain for GAGs.
Hence, the difference in HS affinity between LfcinB and the 9-mers seems to affect the mechanism of action of LfcinB and the 9-mers differently. We therefore propose a mode of action model in which both the LfcinB peptide and the 9-mers are attracted to the anionic glycoconjugate cell coat that surrounds cells. This anionic cell coat consists of both GAGs and sialic acids. The repeating disaccharide structures of HS containing multiple sulfate groups are larger and more negatively charged than sialic acids, which is a monosaccharide with a carboxylic acid group. A stronger electrostatic interaction is therefore expected to occur between CAPs and HS in comparison to sialic acids. In order for the CAPs to exert their permeabilization effect leading to cell death, they have to navigate through this anionic cell coat to reach the phospholipid bilayer. The inhibitory effect of HS on the cytotoxic activity of LfcinB shows that the anionic cell coat may play a limiting role in the cytotoxic activity of LfcinB, in which HS at the cell surface of target cells hinders LfcinB from reaching the phospholipid bilayer. Furthermore, LfcinB that complex with cell surface HS may not be in close enough proximity to the cell surface to destabilize the membrane. The cytotoxic activity of the 9-mers is not inhibited by cell surface HS, thus suggesting that the 9-mers are attracted to HS without being captured. A higher amount of the 9-mer peptides will therefore reach the phospholipid bilayer compared to LfcinB.