do not necessarily reflect the views of UKDiss.com.
Strategies to implement the cytotoxic activity of Saporin-based immunotoxins and chimeras
Since the concept of “magic bullet” was formulated by Paul Erlich (…) there has been a competitive challenge to find out the best toxic substance to be delivered to cancer cells, possibly the one that may avoid cellular resistance and be able to repeat its action inside cells for many times. In this prespective a protein enzyme may have the characteristics matching these requirements, and among the enzymes studied in the last 40 years saporin, first described by the group of Fiorenzo Stirpe (Stirpe et al., 1983), was one of the most appealing candidates. Along this period Saporin has been largely studied as a component of “magic bullets”, tested as antitumour agent in several clinical trials, but facing with problems (common also to other RIPs) that have substantially hampered a therapeutic use of the toxin. This paper suggests possible strategies to improve the efficiency of chimeras (i.e. conjugates or fusion proteins containing saporin as the toxic moiety) wishing a continuation of studies and opening to future development of new formulations as anti-cancer agents.
Saporin structure and function
Saporin is a toxic protein produced by Saponaria officinalis L., a family of Caryophyllacea, belonging to the type I Ribosome Inactivating Proteins (RIPs) family. There toxins differ from type II RIPs that present a lectin B chain (Stirpe et al 1986; Giansanti et al 2010) linked to the A chain by a disulfide bridge, whose reduction is necessary to the make active A chain (Agapov et al 1999; Bellisola et al 2004) effective in damaging ribosomes inside cells.
RIPs posses N-glycosidase activity (EC 22.214.171.124), that removes a specific adenine (A4324 in the 28S ribosomal rRNA) located in a universally conserved GAGA-tetraloop (called α-sarcin/ricin loop). This depurination leads to a permanent inactivation of a large ribosomal subunit, blocking the recognition and binding of the elongation factor EF-1 and the formation of the EF-2 / GTP-ribosome complex making impossible the translocation of the tRNA from Site A to the P site (Endo at al 1998) in the ribosome, thus finally blocking protein synthesis.
Type I RIPs share high pI (> 9.5), while on the contrary type II catalytic A chains are neutral (i.e. pI 7.09 for RTA) or acidic (i.e. pI 5.25 for Abrin A chain) (Stirpe et al 1986). Despite these differences, the 3D-structures of these proteins are mostly super-imposable and the catalytic residues are conserved among all toxins belonging to RIPs family (Savino et al 2000; Azzi et al 2009).
The mature form of Sap-SO6 consists of 253 amino acids of which almost 10% is represented by lysine residues, a condition that gives the toxin an extremely high pI (about 10) (Maras et al 1990).
The crystal structure of the Sap-SO6 (PDB Code: 1QI7) has shown that this protein contains two principal domains: an N-terminal domain with a predominantly beta-sheet structure and a C-terminal domain with a prevalent alpha-helix structure ( Fig. X).
The N-terminal domain is very similar to that of the other RIPs and in this there are six structures arranged to form a beta-sheet mixed with four antiparallel sheets in the center and the other two at the edges that are, instead, parallel.
The C-terminal domain contains eight α-helix with canonical geometry. The helix A and B are part of the cross-connections between the parallel filaments of the β. The helix E and F are contiguous in sequence and a single residue (Phe180) assumes a non-helical conformation, introducing a loop between the two helix (Savino et al 2000).
The active site (Fig. XX) is somewhat preserved between the different isoforms of Saporin, composed by Tyr72, Tyr120, Glu176, Arg179 and Trp208 (De Virgilio et al., 2010).
The residues Glu176, Arg179 and Trp208 take up the same position also in other RIPs and the Tyr72 assumes a different conformations since this residue has proved to be the responsible for the interaction with the target adenine (Savino et al., 2000).
The analysis of the electrostatic potential of the surface (Fig. XXX) indicates a negative potential (colored in red) in the region of the active site. Two glutamate residues (Glu176 and Glu205) are important for this negative charge. The small positive area (colored in blue) is due to the presence of only one arginine (Arg209) at the entrance and two others (Arg136 and Arg179) within the active site (Fermani et al. 2005).
Sap-SO6 isoform is extremely resistant to high temperatures, to denaturation with urea or guanidine and to the attack of proteolytic enzymes [Santanchè et al., 1997]. Moreover, it is also very stable in response to chemical modifications, such as those used for derivatization or conjugation procedures (Bolognesi et al. 1992). Altogether these characteristics made Sap-SO6 an interesting candidate for the realization of immunoconjugates.
In order for Saporin to be a therapeutic molecule, it must be able to enter its target cell and thus overcome the barrier constituted from the cell membrane. While type II RIPs, due to the presence of a lectin moiety, have easy access to the cell using the mechanism of endocytosis and following a retrograde transport travelling through the Golgi and the ER, type I RIPs have low absorption efficiency (Bolognesi et al. 2012) and generally follow intracellular pathways different from those of the prototype Ricin toxin, the most studied type II RIP (Fabbrini et al., 2003; Vago et al., 2013; Ippoliti & Fabbrini, 2014).
With regard to Saporin, the observations that some cell types show a moderate resistance to its cytotoxicity and other cellular types are more sensitive, led some researchers to identify specific surface receptors that mediate the internalization of this toxin. A mechanism of endocytosis mediated by the α-2-macroglobulin receptor [Cavallaro et al., 1995) was proposed in which, however, a discrepancy between the levels of the receptor protein and the cytotoxicity induced by the Saporin was revealed. Some cell lines, positive or negative for the presence of the aforementioned receptor, show a similar sensitivity to the effects of Sap-SO6 suggesting also a mechanism of endocytosis internalization independent of this receptor (Bagga et al., 2003).
Saporin as a component of immunotoxins and chimeras (general, to be written)
Saporin has been linked to a very large number of antibodies to generate “immunotoxins” directed against several cancer forms, with particular interest in haematological malignancies (Polito et al., 2015). Pioneering work was done on chemically linked immunotoxins by the group of Stirpe in Bologna and Flavell in Southampton, directing those molecules against lymphoma antigens such as CD2, CD7, CD19 and CD22 (Tazzari et al., 1994; Morland et al., 1994; Flavell et al., 1994; Flavell et al., 1995; Flavell et al., 1997; Flavell et al., 1998; Flavell et al., 1998; Flavell et al., 2000). In all these studies the immunotoxins were demonstrated to act at sub-nanomolar concentration in vitro and in vivo they could induce survival in 40-50% animal models. The first reported clinical trial with a saporin based immunotoxin is dated back to 1992 (Falini et al., 1992; Pasqualucci et al., 1995), when an anti CD30-saporin was used to treat refractory non-Hodgkyn lymphoma patients. The study first demonstrated a clinical efficacy in reducing the tumour mass by approximately 60%. The effect was limited to 2-4 months but clearly demonstrated the possibility to use saporin-based immunotoxins to treat haematological malignancies. In this study problems connected to the treatment were evidenced, due to side effects of saporin, mainly reported to be caused by vascular leak syndrome (see below) and the production of antibodies against both the toxin and the monoclonal (Pasqualucci et al., 1995).
Immunological reactions elicited by the use of RIPs
It is reported that ITX based therapy may cause some several side effects like such as formation of antibodies against the antibody and the RIP (causing allergic reactions, and even anaphylactic shock), capillary leak syndrome, hepatotoxicity, renal insufficiency, fatigue, myalgia or fever.
Some of the problems related to immuntoxins safety have been solved engineering humanized toxins, by which for example vascular leak syndrome (VLS) could be reduced by the identification of the residues involved in the interaction of the toxin with epithelial cells. (Baluna et al 2000; Lindstrom et al. 2001; Smallshaw et al. 2003). VLS is indeed not a prerogative of plant toxins, as it was observed also for immunotoxins containing Pseudomonas toxin domains: possible strategies for the reduction of non-specific toxicity suggested in this case the use of non-steroidal anti-inflammatory drugs (Siegall et al 1997).
Strategies to improve the toxicity of Saporin-based chimeras: co-treatments with specific cytosolic entry enhancers or anti-neoplastic drugs
To increase the toxicity of saporin-based immunotoxins on different tumor targets, several co-treatments in the literature have been described that tend to maximize the toxin therapeutic effect.
While the route followed by ricin holotoxin in intoxicated cells has been intensively studied and its Golgi retrograde transport elucidated demonstrating that the toxin exploits ER chaperones (calnexin/calreticulin) cycling between this compartement and the TGN to reach the ER where A and B chain of holotoxin dissociate thanks to the action of Protein Disulphide Isomerase (PDI) (Spooner 2004) and thioredoxin-reductases reducing the disulphide bond (Bellisola et al. 2004). Free RTA then mimicks as a misfolded protein and is therefore targeted for proteasomal degradation (retrograde transport of ricin) Bo Van Deurs et al., 1986; Rapak et al., 1997; Lord and Spooner 2011) by interacting with EDEM (ER degradation-enhancing α-mannosidase I-like protein 1) and EDEM-2 which are key players in redirecting aberrant proteins for ERAD (ER-associated protein degradation). Once in the cytosol the A chain may rapidly refold and depurinates the ribosomes (toxin 2017). Type I RIPs are believed to pass through the endo-lysosomal compartment but as shown for saporin should bypass the Golgi-mediated transport route (Vago, 2005; toxins 2017).
Endosomes constitute a complex tubulo-vacuolar endomembrane system of which early endosomes and late endosomes/lysosomes are major constituents (Ruenberg J et al. 2004).Tubular extensions are formed on vacuolar early endosomes to remove specific cargo proteins, and vesicular invaginations sequester other cargos onto internal vesicles that bud from the limiting membrane into the vacuolar lumen. Early endosomes may also transform into endocytic carrier vesicles which then fuse with late endosomes/lysosomes. Sandvig and colleagues showed in 1992 for the first time that an exogenous bacterial toxin – the RIP Shiga toxin – could reach the ER following a retrograde transport route (Sandvig et al 1992) to finally retrotranslocate to the cytosol, so to inhibit protein synthesis. Retrograde trafficking routes from early, recycling and late endosomes seem to exist in parallel (Johannes and Wunder 2011) This landmark paper established the existence of a transport connection between plasma membrane/endosomes and TGN/Golgi complex/endoplasmic reticulum. Ricin uses the retrograde transport route passing throug the Golgi and ending into the ER For saporin and other type I RIPs this is presumably not the main cytosolic entry pathway (MSF, RV, R Ippoliti 2013, Toxins 2017)
Escape from the endo-lysosomal membrane system represents another bottleneck for type I RIPs and several potential endosomal escape enhancers have been investigated . For instance, endosomal membranes could be permeabilized using lypopolyamines or DMSO, leading to an increased access of saporin to the cytosolic compartment, but not that of RTA (Geden, 2007).
One of these most popular treatments is based on the use of glycosylated triterpenoid saponins (extracted from plant sources) that may induce stimulation of the endocytosis of saporin by the target tumor cells (Weng et al. 2008; 2009).
The use of saponin SO-1861 (extracted from roots of S. officinalis) in co-treatment with EGF-Sap (Sap3-EGF, i.e. saporin fused to epidermal growth factor) was used to reduce the Immunotoxin dosage that resulted more effective than conventional ITx alone against TSA-EGFR target cells with a high level of synergy and much reduced side-effects (Thakur M. et al. 2013). In details 0.1 g/treatment Sap3-EGF were required to achieve more than 90% tumor reduction in the presence of saponin whereas 5 g/treatment were necessary to obtain a regression of only 33%, without saponins (Fuchs et al., 2007).
Moreover Saponin SA1641 (extracted from Gypsophila paniculata L.) was used in co-treatment with a chimera containing a PDZ domain as the targeting moiety instead of a classical monoclonal antibody. The PDZ domain from hCASK (Human calcium/calmodulin-dependent serine protein kinase) binds extracellular CD98 in epithelial cells and this antigen is recognized as a marker for several human tumors and particularly considered as a negative prognostic marker for human glioblastoma. The recombinant fusions ITx with one or two hCASK-PDZ domains with the ribosome inactivating protein Saporin (Sap6) was tested on several glioblastoma cell lines. Particularly Saponin SA1641 was able to increase the toxicity of Saporin and h(CASK)2-SAP in U87 Glioblastoma cell line (IC50 <10-11M and <<10-12M respectively), and allowed hCASK-SAP to reach an IC50 of 10-10 M (Giansanti et al. 2015). When we compared the performances of anti CD22 single chain variable fusion constructs/scFV to identify novel recombinant single chain antibodies showing nanomolar affinity for the CD22 antigen, a known marker for B-cell lymphomas and some leukaemias, fused either to saporin or to Pseudomonas aeginosa PE40 bacterial-derived toxin we found comparable cytotoxicities between the plant or bactarial derived ITxs (please see the section below) (Erri Provenzano et al. 2016). However, Saponinum album saponin enhancer augmented the PE40 based ITx by 170-fold whilst the saporin-based ITx was augmented more than 6 orders of magnitude by this co-treatment (David Flavell et al. unpublished results). This great difference in the saponin enhancer effects on PE40-ITx versus saporin-based ITxs reflects the preferential interaction of some type I RIPs with these triterpenoid glycosylated saponins (von Mallinckrodt, B et al 2017) .
In general, endosomal escape enhancers i) should not exert toxicity per se for the target cells ; ii) their co-adjuvant role requires their concomitant presence at the site of dislocation during the passage of the toxin “en route” – to its final destiny iii)-. Nor should these agents facilitate toxin cytosolic entry into off-target cells; While these optimal conditions should be ideally tested in in vivo systems, unfortunately, most of the data published using endosomal escape enhancers to date have been mostly examined in vitro, which remains a serious limitation for their therapeutic use. However some have demonstrated efficacy in animal models (Bachran, et al. 2009) . For a comprehensive description of endosomal enhancers used to potentiate plant RIPs-based ITxs please refer to (Fuchs et al. 2017)
Another innovative methods to enhance the delivery and uptake of therapeutical agents, as saporin, is the encapsulation of drugs using specific exosomal membranes modified with different peptides including a cationic pH-sensitive fusogenic peptide known as GALA peptide (Toxin 2017). These viral-derived peptides are able to enhance the disruption of the endosomal membrane, leading to a more efficient cytosolic release of proteins. Nakase and coworkers used these GALA peptides, one having sequence WEAALAEALAEALAEHLAEALAEALEALAA (amino acid one letter) designed to mimic a viral specific fusion peptide sequence that mediates the escape of viral genes from acidic endosomes to the cytosol. The GALA peptide converts its molecular structure from random coil to a helical when the pH drops from 7.5 to 5.0, passing from endocytic vesicles to acidifed endosomes, leading to membrane fusion and release of the content of the endosome. This technique was used to deliver saporin, leading to the efficient induction of cytotoxicity in HeLa targeted cells (Nakase I et al. 2015: Toxins 2017)
MACROPINOCYTOTIC PATHWAYS AS POSSIBLE EFFICIENT ENTRY PATHS?
Macropinocytosis is upregulated in several cancer cells and is a more efficient internalization pathway capable of bulk receptor-mediated endocytosis with macropinosome vesicles having an average of 200-300 nm diameter, well-suited for nanodelivery approaches, as well. Using an automated microscopy-based imaging platform to identify phage antibodies co-localizing with macropinocytotic markers, Ha and co-authors screened their single chain antibody libraries enriched in internalizing antibodies by a laser capture microdissection selection, which bound preferentially to tumor cells (Ruan, et al. 2006). They now have selected phage displayed antibodies that favour this peculiar entry route which were retransformed into full lenght humanized IgG molecules and showed specific targeting to the ephrin type A-receptor 2, a receptor known to be involved in tumor invasion and metastases dissemination (Ha, Bidlingmaier, et al. 2017) that has been a target for cancer therapies (Ha et al, 2017). Cancer cell lines overexpressing ephrin type A-receptor 2 were used to show efficient killig by a derivatized biotinylated IgG termed HCA-F1 followed by reaction with a streptavidin-conjugated saporin to produce a specific IT that killed DU145 prostate cancer cells with an Ec50 of around 19 pM after 96h exposure. The same IT was found uneffective against receptor negative cells. The authors observed that either fusions made with sc-Fv or with the full IgG can be used for validation studies, since no major differences were observed in the internalization patterns between these two antibody formats. In addition, since the epitope bound by this novel HCA-F1 anti-EphA2 antibody shows cross-species specificity, therefore toxicity profiles could be also first tested in small rodents. This is a first example of a phage antibody library allowing for screening macropinocytotic antibodies that could become an useful tool to select new agents targeting tumor cells which exploit efficient macropinocytotic entry route.
MAJOR BOTTLENECKS & POSSIBLE SOLUTIONS
The successful clinical development of Immunoconjugates has been hampered by several heavy side effects including nonspecific toxicity and vascular leak syndrome. Already in 1988 a Cancer Treatment research’s review stated: “A major problem with this ITx approach is that it is likely that both the antibody and the toxin components of these conjugates will be immunogenic (in humans). Both antitoxin and antixenogenic immunoglobulin responses have been shown to occur in animals after infusion of IT… (Lambert et al 1998)
Major drawbacks when using murine targeting domains (monoclonal antibodies), as well as plant domains as therapeutic agents, are represented by their potential immunogenicity, especially after repeated administrations (refs). ITxs which contain fully humanized antibodies or antibody fragment lacking Fc portions partially could resolve these issues by avoiding the induction of human anti-mouse antibodies (HAMA) directed to the Fc murine fragments in treated patients (Holmes et al 2006) In addition, bacterial or plant toxins can also trigger the formation of neutralizing antibodies, hindering their efficacy. Many efforts are being made to decrease immunogenicity of the plant toxin moiety; one first choice which is possibly more easy to achieve is the masking of the therapeutic molecules by use of polyethylenglycol (PEG) derivatization techniques (Molineux 2003)
Clinical studies administrating in patients ITxs made with the plant Rip saporin did not unfortunatelly progress beyond early phase I studies, due to side-effects and/or immunogenicity issues. Moreover, for tumor penetration and distribution inside solid tumor masses the classic ITx format should be better engineered by reducing the size and choosing humanized scFv formats to try minimising immune responses through PEGylation masquing strategies or through specific de-immunization strategies, as for the de-immunized-bouganvillea RIP termed De-bouganin .(improved biopharmaceuticals for oncology. Pharmacotherapy 2003, 23, S3–S8 .] Type I RIP, bouganin from the leaf of Bougainvillea spectabilis Willd. was mutated to create a de-immunized variant termed “DeBouganin” in which most of the recognized T-cell epitopes could be identified and removed, in this case without affecting RIP’s activity. Bouganin has LD50 exceeding 30mg/kg (versus 5-10mg/kg of saporin) and turns out ot to have a much lower systemic toxicity as compared to saporin. Glen C. MacDonald’s group identified and removed T-cell epitopes using Biovation (Merck subsidiary) technology synthetizing 15 aminoacid long peptides covering the mature toxin protein with a 12-residue overlap among these different peptides and incubating them in the presence of peripheral blood mononuclear cells isolated from 20 naive donors using 96 well-microplates assays aimed at identify stimulating reactive peptides. Using a peptide threading software they could detect key reactive aminoacids for MHC class II binding. They expressed the de-bouganin mutants with an hexahistidine tag demonstrating that the T-depleted De-bouganin variants show comparable activity in reticulocyte lysates but minimal T-cell activation. They used the De-bouganin for targeting Epithelial cell adhesion molecule EpCAM a surface antigen confined to the baso-lateral surface on normal epithelia but highly overexpressed in several tumors. This patented bouganin variant was genetically fused to an anti ephitelial cell adesion molecule (EpCAM) Fab fragment via a peptide linker containing a furin proteolytic site, creating VB6-845 IT for treatment of solid tumors and successfully tested on EpCAM-positive human tumor xenograft model in SCID mice. Interestingly,the moderate affinity anti-EpCAM antibodies show lower toxicity profile than higher affinity counterparts and thus, a low affinity antibody was used to produce a fusion to the anti EpCAM termed VB6-845 which was characterized in vitro for killing capability and pre-clinically in xenograft mouse model with NIH:OVACAR_3 tumors. A comprehensive series of pharmacology and toxicology studies were conducted using VB6-845, including a phase I trial in patients bearing advanced squamous cell carcinomas of the head and neck. The data showed that VB6-845 could reduce or stabilize tumors about 70% of patients with a maximum tolerated dose of 280 mg of IT, administered daily for five days. These ITx was well tolerated with the only adverse effects being pain due to the intratumoral injections and transiently elevated liver enzymes.
More recently when De-Bouganin was conjugated to Trastuzumab (anti Her2/Neo) T-deB was also shown to be more potent than a Trastuzumab-emtansine T-DM1 and was able to overcome drug resistance due to anti-microtubule agents such as Emtansine or methy(MMAE). Tumor cells surviving anti-microtubule conjugated drugs (T-DM1 or T-MMAE) importantly remained still sensitive to another anti-HER2 C6.5 Diabody fusion made again using De-variant . Resistance phenotypes comprised expression of Multiple Drug Resistance (MDR) efflux pumps, Bcl-2 members and other pathways, to be as yet identified.
Antineoplastic treatments with ITX can sometimes be ineffective when only one construct is administered. To enhance and optimize anti-cancer approaches more ITX could be administered simultaneously. As a‘proof of principle’ of the possibility of targeting two overlapping but distinct subpopulations of cells by using a secondary immunotoxin treatment is the dual treatment used against Glioblastoma Multiforme (GMB) by combining the use of the “Mab-Zap” saporin immunotoxin system developed by (Targeting systems?). This innovative system consist of NG2 antibody-SAP and GD3A antibody-SAP that can be used in sequence. Neuron-glia 2 (NG2), is a transmembrane chondroitin sulphate proteoglycan, present on developing glial cells, while GD3A, is a ganglioside expressed on cell surface of developing migratory glia, both are found re-expressed in GBM. This sequential combinatorial ablation of both NG2 and GD3A-expressing cells resulted in significant reduction in GBM cell viability as compared to the single epitope targeting as controls (Higgins SC et al. 2015).
Another example of a co-treatment with two different saporin-based immunotoxins was demonstrated by Flavell and colleagues by using chimaeric anti-CD20 antibody rituximab (Rituxan) and anti-CD19 immunotoxin BU12-SAPORIN on Ramos cell line (human lymphoma) both in vitro (inducing apoptosis) and in vivo. This combined treatment made use of two immunotherapeutic reagents directed against different B-lineage molecules on the target cell surface and resulted in an increase in both the potency and fidelity of the immunospecific attack (Flavell at al. 2006)
Rituximab was used also conjugated to seed extracted saporin in combination with another ITx containing saporin-S6. These two ITxs were obtained by chemical conjugation of the IgG with saporinS6. Rituximab/saporin-S6 and OM124/saporin-S6 are directed respectively against CD20 and CD22, both antigens highly expressed on Non-Hodgkin’s lymphomas (NHLs) and found particularly expressed at high levels on normal mature B-cells and on a large population of B-lymphoma cells but absent in the normal tissues and hematopoietic stem cells.
Another way to increase the toxicity of these two saporin-based ITxs was obtained by the simultaneous administration of classical chemiotherapeutic agents such as PS-341, MG-132, and fludarabine (Polito L. et al 2017).
Another interesting improvement of the Saporin-S6-Rituximab was obtained by chemical conjugation through an artificial disulfide bond. In this way, they could obtain two species having differing MW and inhibitory activity. The HMW-ITx (Dimeric) and LMW-ITx (Monomeric) had sigmilar activity in inhibitin cell-free protein synthesis but in two CD20+ lymphoma cell lines, Raji and D430B, HMW-IT was found more cytotoxic than LMW-IT.
This double ITx system (HMW and LMW-ITx) can be seen as a potentially versatile system, because can be used as endocytosis enhancer using a dimeric HMW (for antigens with low internalization rate) or as improved penetrating systems for solid tumors due to small dimensions (Bortolotti M et al 2016).
ENHANCED DELIVERY THROUGH Phototherapy approaches
The classical approach to site-specific drug delivery of a target-specific immunotoxin against tumors cells requires an improved endocytic uptake and efficient release of the therapeutic molecule from intracellular compartments, mainly endocytic vesicles as stated above, into the cytosol. Nevertheless, this approach can show several limitations such as the low rate of penetration through the membranes of endocytic vesicles and/or compartments and degradation of the therapeutic macromolecules by lysosomal enzymes (Berg et al 2010).
To partially circumvent this problem, another new technology has been recently developed: Photochemical internalization (PCI). The PCI is not specific for few plant type I Rips as the saponins as coadjuvants but may help to release mostly endocytosed macromolecules into the cytosol being based on the use of specific photosensitizers located within endocytic vesicles that, upon activation by light, can induce the drug release due to photochemical rupture of endocytic membranes (Berg K et al 2010).
Weyergang and colleagues demonstrated that PCI significantly increases about 1000-fold the toxicity of EGF–saporin on NuTu-19 cells (rat epithelial ovarian cancer cell line, higly EGFR-positive cells) respect to PCI untreated cells (Weyerganag A et al. 2006).
The evolution of PCI technology is the in vivo treatments using laser-controlled endosomal escape in localized tumors. An example of this application is the treatment with a CD133-targeting immunotoxin AC133–saporin (PCIAC133–saporin) that colocalize with the PCI-photosensitizer TPCS2a in a CD133 positive cells (CD133 is a cancer stem cell (CSC) marker). After light exposure, a cytosolic release of AC133–saporin in target cells was induced . The advantages of this thecnology are the ability of PCI-photosensitizer to preferentially accumulate in tumor tissue, the possibility to trigger in vivo locally the PCI photosensitizer in tumor tissue and finally the tumor-selectivity of the immunotoxin (Bostad M et al. 2015).
PCI is very helpful also for treatment of CSCs expressing CD44, that characterize tumor cells highly resistant to ROS attack, enabling these cells to be more resistant to chemo- and radiotherapy. Moreover CD44 has also been reported to be involved in multidrug resistance (MDR). Bostad and coworkers made an ITX using a biotnylated-CD44 mAb and a avidin-Sap. After administration of the PCI-photosensitizer TPCS2a and ITX in 7 cancer cell lines of carcinoma and sarcoma origin. It was observed an efficient and specific cytotoxicity in CD44-expressing but not in CD44-negative cancer cells (Bostad M et al. 2014).
Another antibody that was improved thanks to PCI is the commercially humanized HER2 mAb, trastuzumab. This Antibody was used in chemotherapeutic treatment of HER-2 positive breast cancer. Unfortunately the acquired resistance to trastuzumab treatment is not rare and difficult to solve. A novel strategy to overcome this problem is PCI. Using biotinylated-Trastuzumab and avidin-Saporin was realized a novel ITX for PCI. A new ITx was realized couplig trastuzumab to saporin via biotin-avidin protocol, as above described, and administered with PCI-photosensitizer TPCS2a on trastuzumab-resistant HER2+ Zr-75-1 cells prior to light exposure (i.e. “light after” procedure), increasing citotoxic effect because “light first” procedure can reduce the trastuzumab-induced HER2 endocytosis of ITx (Berstad MB et al 2012).
Moreover Vikdal and colleagues demonstreted that PCI can increase also the accumulation of free saporin and subsequent enhanced PDT induced cytotoxicity in the HUVEC cell line (vascular endothelial) spect to HT1080 cel line (fibrosarcoma). This strateg can be used as an antivascular strategy in the anticancer therapy (Vikdal et al 2013).
An alternative approach to overcome the potential barriers that can reduce the cellular uptake and intracellular release of saporin is the use as carrierr of generation 4 polyamidoamine (PAMAM) dendrimers. This construct is able to improve its endocytic uptake, passive tumor targeting, and implemented the photochemical internalization (PCI) technology and facilitate Saporin cellular uptake and cytosolic release (Lai PS et al. 2008).
Recently a PCI-based drug delivery was applied to EpCAM positive cancer cells. A novel ITx formed by a EpCAM-targeting mAb 3–17I linked to saporin was realized and in vitro test shows strongly and selectively reduction of cellular viability, proliferative capacity, and colony forming ability in breast carcinoma, pancreatic adenocarcinoma, and colon adenocarcinoma cell lines (Lund E et al. 2014).
As a further and innovative procedure for the implementation of RIPs toxicity, plasmid co-transfection with two toxins was used in a in vitro model. This tecnique, also called Toxic gene therapy (or suicidal gene therapy) was tested on various cancer cell lines (HeLa, U87, 9L, and MDA-MB-435) using two different gene toxins inserted in a two distinct mammalian gWIZ plasmids: pGEL (gWIZgelonin) and pSAP (gWIZsaporin). The cotrasfection treatment in mammal tumor cell lines using cationic polyethyleneimine (PEI) is able to induce a high level of cytotoxicity (Min KA et al, 2016).
Improvement of saporin production by optimization of recombinant expression
To express plant protein toxins and especially recombinant fusion chimaeras, based on saporin or Type I RIPs it would be advisable to be able to preferentially use eukaryotic expression systems. First of all because type I RIPs, as saporin, are indeed, secretory plant proteins and also because most of the suitable targeting domains would greatly benefit from being expressed in the microenvironment of an endoplasmic reticulum (ER) in order to undergo proper folding and being subjected to quality control mechanisms before extracellular polypeptide secretion occurs. However, this exogenous toxin expression could lead easily to auto-intoxication related issues. The principal way Ricinus communis protects itself from self-intoxication is to produce a long inactive precursor where the nascent polypeptide is efficiently co-translationally inserted inside the lumen of ER to avoid ricin mislocalization in the cytosol (where target ribomes are) (23). In fact, orphan ricin A polypeptide chains when expressed in Tobacco leaf protoplasts are retained in the ER and retrotranslocated to the cytosol for proteasomal degradation (Di Cola et al. 2001). The saporin precursor, instead follows a quite different fate in this plant model system: in contrast to what observed with ricin A chain, saporin polypeptides were found to be efficiently secreted into the incubation media with protoplast’s intoxication demonstrated to be due to a stress-dependent pathway. Saporin cytotoxicity apparently involved the action of a few toxin molecules that were at least first partially inserted into ER membranes (Marshall et al 2011). By investigating the behaviour of the plant saporin signal peptide we showed that it may act as an ER-stress responsive element (Marshall et al. 2011), recently demonstrating that other type I RIP signal peptide(s) may act as ER stress-sensors, as well (Toxins 2017). To date bacterial, model plant cells and more recently, yeast strains have been used to produce RIPs or RIP-based chimeric fusions. However, common problems faced during recombinant production of type I RIPs or their derived chimaeras resided in their intrinsic toxicity towards the host ribosomes. Indeed, even the initial attempts to express recombinant Type I RIPs in E. coli were also found problematic, because, upon induction of RIP’s expression the bacterial growth rate was significantly impaired, as earlier reported in the case of Mirabilis antiviral protein , PAP , dianthin (26) well as saporin [27;28] and more recently a saporin L1/L3 vacuolar leaf variant could not even be bacterially expressed (REF Yuan te al.). Although toxin expression could be tightly regulated by employing the E. coli strain BL21(λDE3)pLysS to get satisfactory yields in several instances, endotoxin A contamination and processing of the insoluble bacterial pellets remained sometimes problematic (de Virgilio et al. Toxins 2010). Strikingly, while Saporin L3 at was found too toxic even during the simple propagation of plasmids containing the RIP’s DNA in host bacteria, the L3 variant was instead reported to be expressed only in P. pastoris under an alcohol-oxygenase (AOX1) tight regulated promoter and by inserting the prepro-alpha factor signal peptide at the NH2-terminus of saporin L3, as we also previously reported for high level expression of saporin-6 seed isoform in P. pastoris (Lombardi et al 2010). The choice of a tight inducible regulated system such as AOX-1 together with an efficient signal peptide such as the one of the prepoalpha-factor or alternatively by using those present in ER chaperones such as Immunoglobulin binding protein (BIP) or protein disulphide isomerase (PDI) (Toxin 2017) is also of outmost importance to avoid toxin mislocalization into the eukaryotic cytosol. This microbial protein expression system has shown several advantages over other eukaryotic or prokaryotic expression systems: ability of P. pastoris to grow at very high cell-densities (with a rapid growth rate) and almost protein-free extracellular medium; high levels of foreign proteins can be expressed under the methanol-induced Promoter AOX1; correct posttranslational modifications like N-glycosylation (although with different oligosaccharide side chain lenght),
methylation, acylation, proteolytic maturation and proper subcellular targeting of the heterologous polypeptide (Cereghino, J et al 2013)
Furthermore, this yeast expression system is under intensive investigation to eliminate the disadvantages of methanol-oxygen consumption in fed-batch large scale fermentors, Wang and collaborators engineer a new methanol-free P. pastoris by modifying transcription factors regulating AOX1 promoter and developing an efficient glucose-glycerol-shift induction bioprocess control for foreign protein expression. The newly constructed strain could efficiently replace the traditional glycerol-methanol induction in the wild-type for insulin model protein expression, exhibiting a more economic, safe and environment-friendly impact having great potential for biopharma industry (Wang et al 2017)
In addition, another crucial parameter for high levels of protein expression that should be further optimized is codon-usage.
For many years it has been empirically assumed that a relationship between ribosome speed in translation and the resulting levels of an exogenous protein synthesized may exist. Codon-usage choice evolved as means to optimize single mRNA translation essentially by changing the rare codons present in the heterologous gene into those preferred by the expression system (codons found in most highly expressed endogenous proteins). This is a common approach that we also employed for optimal expression of seed saporin and its derived fusion chimeras in P. pastoris (refs). Protein expression in eukaryotic cells is not solely controlled by translation initiation factors. In exponentially growing cells, polypeptide elongation factors (eEF1A, eEF2, and eEF3) may exert the strongest translational control (Firczuk et al.2013). Chu and colleagues recently reported that when rare codons are positioned nearby the start of the coding region, pausing interferes with making the initiation codon available for loading of the next 40S subunit which may be rate‐limiting for initiation and therefore for the overall protein synthesis. Translation efficiency is therefore the result of the interdependence of ribosome association either by de novo initiation or by recycling where both elongation and initiation factors are contributing to final protein expression levels. Concerning translation of protein toxins it is outmostly important that the polypeptide is efficiently segregated in the ER lumen as the signal peptide emerges from the polysomes. As we demonstrated for saporin in Tobacco protoplasts, BIP (or PDI) signal peptides may be among the most efficient ones (Marshall 2011)
Working on codon-usage optimization we could obtain just 10 times lower levels of secreted active saporin as compared to SAP-KQ an inactive saporin catalytic mutant.
Intrinsic host cell toxicity was consistently decreased using this strategy also for expression of chimaera such as ATF-saporin (better expressed in P. pastoris with a codon-usage optimised ATF domain (human) and single chain variable domain scFV fusions . An anti CD22 scFv termed 4KB was obtained from the subcloniong of VH and VL regions of a parental monoclonal antibody 4kb128 used to construct a saporin-based immunoconjugate (REFS). The scFv showed could be internalized by target CD22+ human Daudi cells being also competed for by the parental monoclonal CD22 antibody REF.
Interestingly, we have also compared expression of ITs made by using two toxic enzymes (both able to block protein translation, one of bacterial origin (PE40) endowed with EF-tu2 ADP-ribosylation activity, the other being saporin) wiht the aim of evaluating optimal microbial expression of various scFv fusion formats evaluating the performance of the different fusion constructs, with respect to yields from E. coli or P. pastoris cultures and their specific killing efficacy. A number of fusion constructs were designed and expressed either in E. coli or in Pichia pastoris and the resulting fusion proteins affinity-purified. Protein synthesis inhibition assays showed that the selected recombinant ITs were active, having comparable IC50 (inhibitory concentration by 50%) in the nanomolar range. Overall our results confirm that E. coli is the system of choice for the expression of recombinant fusion toxins of bacterial origin that otherwise underwent proteolytic degradation in yeast cells whereas we further demonstrate that saporin-based ITs are best expressed and recovered from P. pastoris cultures after yeast codon-usage optimization. Codon-optimization of the scFv domain would appear important to increase the potential number of secreting clones able to produce at least 1-2 mg/L of fusion protein. In addition, among the alternate design options, the best performing ones were those having a yeast codon-optimised anti CD22 sequence assembled with a 18-amino hydrophilic 218-flexible linker GSTSGSGKPGSGEGSTKG, amino acid one letter code that showed enhanced resistance to proteolysis with reduced aggregation of scFvs when expressed in bacterial systems (Withlow et al., 1993; Rosemblum et al., 2003)joining the VH and VL codon-optimized variable chains and fused to the N-terminus of mature saporin through a trialanine linker ,
The different scFv versions having either a (G4S)3 linker or other formats were either not producing viable clones or gave rise to mutated saporin fusions (inactive).
The immunogenicity of RIPs: analysis of experimental data and comparative prediction for Saporin
Since RIPs have been suggested as components of immunotoxins, one of the problems suddenly arisen upon patients’ treatment was the immunological reaction. It has been reported that hematological tumour patients respond with the production of antibodies up to about 40% and can undergo eventually multiple treatments with immunotoxins (ref 18#01), but solid tumour patients are instead more prone to immunological response (50-100%) and cannot receive more than a single dose (ref 19#1), thus greatly decreasing the curative potential of these drugs. This discrepancy may be related to the higher accessibility of hematological tumours and to the immuno-compromised state of these patients with respect to solid tumour patients. Notwithstanding the possibility to treat some forms of solid tumours by local administration (i.e. bladder, ref 25-29#1) thus reducing the exposure to the immune system, there is urgent need to identify critical epitopes on RIPs to improve their therapeutic applicability. For these reasons some Ribosome inactivating proteins have been investigated to identify the antigenic determinants that induce immunological response, and recently Cizeau et al. (#1) described a mutagenesis approach to reduce immunogenicity of the RIP bouganin. (PDB:3CTK)
They identified some peptides inducing T cell proliferation and suggested by in silico analysis some residues as possible antigenic determinants for MHCII recognition. They changed single aminoacids or combinantion of them in those critical regions and finally expressed a mutant bouganin (deBouganin) carrying four mutations (V123A, D127A, Y133N, I152A) that didn’t influence the catalytic activity of the enzyme but significantly reduced the immunological response of the toxin. They used this mutant toxin to construct a Fab-bouganin fusion recognizing EpCAM (VB6-485) on the surface of OVCAR3 cells and tested it in a subcutaneous model SCID mouse, where this immunotoxins produced a marked decrease of tumour growth and also 20% of free disease survival. At the moment VB6-485 is under clinical trial phase I by VIVENTIA Biotech Inc.(Winnipeg MB, Canada), as well as other deBouganin-based immunotoxins.
Pioneering work was previously done by Mulot et al. (#4) on Trichosanthin (PDB:1TCS) a well studied RIP since its anti-viral potential (particularly against HIV). This group roughly identified very large peptides (positions 1-72, 101-152 and 153-246) as potential antigenic sites. Due to its large diffusion in the traditional chinese medicine, also other groups investigated the problem of immunogenicity of this toxin: indeed An et al. (#2) have been describing mutants of the RIP Trichosanthin in which two potential antigenic sites (peptides YFF 81-83 and KR 173-174) were selected on the basis of a structural analysis. YFF sequence was selected due to the large side chains of these three aromatic consecutive residues, while KR was identified as a high hydrophilic and exposed loop on the surface of the protein. The mutants were designed as to have the following changes (YFF81-83ACS and KR173-174CG), with the addition of two cysteine residues potentially useful for further conjugation reactions (i.e. Pegylation, #2). The mutants of Trichosanthin showed unaltered enzymatic activity and a reduced immunogenicity in terms of IgE production in immunized animals, but didn’t suppress IgG response. Thus this approach although promising cannot be fully removing immunogenic potentials of this RIP. Gu et al. (ref18#3) identified at least four different epitopes by using antibody competitive assay, while the group of Chan (ref19#3) obtained a reduction of immunogenicity but also of activity deleting the last seven residues at C-terminus of Trichsanthin. Zhu et al. (ref20#3) predicted the presence of two regions of immunogenicity in the peptides 21-27 and 41-48. Cai et al. (ref21#3) identified the region 169-174 partially overlapping the results obtained by An et al (#2). Chan and He (ref22-24#3) successfully mutated residues S7C, K173C and Q219C to add cysteine residues suitable for PEG or dextran modification, able to substantially reduce the immunogenic response to Trichosanthin. Zhang et al. (#3) finally identified two specific residues (Y55 and D78) exposed on the surface of the toxin that upon mutation (Y55G and D78S) strongly reduced the immunogenic response in animals. A representative figure of the antigenic sites in Trichosanthin is reported below (da fare ex novo, questa è presa da #3). It is evident from these data that for Trichosanthin there is a muliplicity of potential immunogenic sites that has not been completely unravelled.
Leung et al. (#5) mapped three different epitopes on the surface of alpha-momorcharin (PDB:1FQ8), another anti-viral and anti-cancer protein widely diffused in chinese medicine. The three regions (1-14, 71-136 and 195-222) were identified using an anti-momorcharin antiserum and peptides corresponding to different regions of the protein. It is worth noting that the peptide 71-136 corresponds to the catalytic region of this enzyme, thus suggesting a neutralization effect of antibody binding in this area.
Different situation may be possible for type II RIPs, since their dimeric composition may allow the catalytic A chain (corresponding to the single typeI RIP chain) to be partially protected to the immune system by the B lectin chain. Anyhow it has been reported that the typeII RIP viscumin A chain exposes new epitopes following internalization of the toxin inside a cell, as a consequence of structural rearrangements (#6). Tonevitsky et al. identified the region 96-105 of the A chain of viscumin by the recognition of monoclonal antibodies as a structural relevant transition region connected to the passage of the toxin from the endovescicular system to the cell cytosol. How much this can be related to the exposure of immunogenic sites in immune cells is not established.
Tommasi et al. (#7) studied the immunological epitopes for ricin A chain focussing the attention to a mini epitope corresponding to the regions I175-Y183/M174-I184, and finding that the residue I175 may be crucial for HLA classII recognition of ricin A chain.
Analysis of the primary structure (Fig.1) of saporin in comparison with Bouganin, Thricosanthin and Alpha-momorcharin shows that the YYF or YFF sequence is present in all toxins but Bouganin, while the KR motif is not significantly represented in conserved positions. Bouganin residues inducing immunological response are not conserved among these group of RIPs. If we look at the aligments of the typeI RIPs for which a crystal structure is known (Fig. 2) we can notice that the YFF motif is largely conserved among typeI RIPs or at least the Y residue is in all the RIPs even though those from Phytolacca americana and Phytolacca dioica share a peculiar YHIF motif in the same region.
This motif is also present in Ricin A chain and Viscumin A chain (Type II RIPs) thus suggesting its possible common role in all RIPs. It is worth noting that these residues are not involved in the catalytic activity of RIPs. If we analyze the structural localization of the three residues (Fig. 4) we can notice that (comparing Saporin, Ricin A chain and Thricosanthin) the first residue (Y88 in Saporin) is always present and hold in position by another tyrosine (Y76 in Saporin) conserved in all RIPs (see Fig.2). Thus according to what observed in Trichosanthin, it is reasonable to suppose that this conserved motif may be mutagenized in RIPs with alteration of immunological epitope(s) without alteration of catalytic activity, thus representing a good starting point for the de-immunization of these toxins.
Prediction of immunogenic epitopes is a complex issue, since the presence of allelic variants of MHCII in humans render this analysis anyhow very imprecise. However if we restrict the analysis to the most diffuse allelic forms of MHCII (according to Greenboum et al.) we can observe that in the case of Thricosanthin (Fig. 5) the program PropRed predicts some peptides containing the YFF and among those (DRB_0101, DRB_0301, DRB_0401 and DRB_0405) the one recognized by the allele DRB_0101 has Y82 as the critical residue for the binding site recognition. Looking at Saporin (Fig. 6) there are some peptides in the region 85-100 that are predicted to be recognized by the MHCII alleles DRB_0401, DRB_0405, DRB_0802 and DRB_1101 containing the YYF sequence analysed above. Particularly in the case of DRB_0802 and DRB_1101, Y88 is reported as the critical residue for the binding site recognition. The tyrosine residue (Y76 in Saporin and Y75 in Trichosanthin) contacting Y88 in Saporin and Y82 in Tricosanthin, is furthermore in both cases included in antigenic peptides predicted to be recognized by many of the allelic MHCII. Furthermore, it can be seen that among the residues involved in catalytic activity of this RIP, only W208 is included in a potential MHCII binding site. Since the catalytic site of Saporin is deeply embedded in the protein matrix, it is unlikely that this peptide may be available for the recognition by antibodies raised against this sequence in the immune response to the circulating toxin.
It can be underlined furthermore that, notwithstanding a diffecrence in aminoacidic sequence between Saporin and Bouganin, the critical antigenic regions of Bouganin correspond (on the basis of the prediction made by PropRed) to peptides predicted to be antigenic also in Saporin (see fig. 6, boxes in green).
Thus on the basis of the analysis of the experimental data available on Trichisantin and Bouganin, and looking at both the structural position and the predictions made on MHCII binding sites, we could hypotesize that in Saporin the antigenic critical regions are quite overlapping those of this two typeI RIPs.
If we furthermore compare the prediction for antigenic sites in a larger number of RIPs (among those whose three dimensional structure is known) possibly some common futures can be evidenced. As shown in Fig. 7, if we analyze the position of the first aminoacid of each predicted immunogenic peptide, some clusters can be evidenced (numbered from 1 to 7). Among those, if we exclude those containing catalytic residues and those represented by a single aminoacid, we can focus on clusters 2, 4 and 5. Looking at the position of peptides orginating from the first aminoacid in the cluster (typically 9 aminoacid peptides), for saporin (see fig. 8) we can observe that both cluster 2 and 4 contains residues facing loops that are exposed to the external part of the toxin, while cluster 5 is embedded in an alph-helix that spans over an internal surface. Thus it could be suggested that residues present in those two loops may be studied as potential candidates for immunological de-potentiation of this toxin.
Vascular leak syndrome: how could we act on Saporin structure?
Vascular leak syndrome is due to a nonspecific binding of the plant toxic domains to vascular endothelial cells (VLS) [179–181] which is characterized by interstitial edema, hypoalbuminemia, weight gain, and in most severe cases, hypotension and pulmonary edema. Vitetta and colleagues identified in ricin A chain a consensus aminoacid sequence “X-Asp-Y”, also found in Interleukin-2, where “X “ could be the non-polar aminoacids Leu, Ile, Gly or Val and “Y” could be Val, Leu or Ser constituting an oligopeptide sequence which may induce vascular damage to human endothelial cells in vitro by binding to integrin receptors: proteins such as RTA and some type I RIPs [refs.] indeed carry this consensus sequence. In case of RTA, molecular modeling showed that these motifs were mainly exposed on the surface of the molecule . A very similar motif is also shared by the viral disintegrins which disrupt the function of integrin receptors . Vitetta and coworkers produced a series of RTA mutants, and identified Asn 97 change to an Ala mutation, in a region flanking the VLS-responsible motif, as displaying significant less VLS in mice. (Smallshaw et al 2003; Baluna et al 1996; Vitetta 2000; Janosi 2013)
Vitetta’s group addressed first this problem arisen during the clinical treatment of cancer patients with RTA-based immunotoxins,(Vascular Leak Syndrome, #8). They used native deglycosylated ricin A chain and observed that dgRTA ITxs often caused a marked increase in vascular permeability leading to diffused edemas and organ failures (refs 4,5 #8). This syndrome was then also recognized to be dependent on the use toxins both of plant and bacterial origin and of the cytokine IL2 (ref. 5-7 #8). All these proteins were shown to interfere with fibronectin adhesion and hampered cell-cell and cell-matrix interactions (#8).Vitetta and co-workers showed that peptides derived from RTA or from the bacterial toxin PE38, if linked to a monoclonal antibody (anti-CD22, #8), induce HUVEC cell toxicity. Peptides with altered or deleted “LDV” sequence of RTA in which the motif (x)D(y), where (x) can be L, I, G or V, and (y) can be V, L or S was mutated (#8were instead unable to induce damages to HUVEC cells. Further experimental work then identified also other flanking structural aminoacids as potentially cooperating to this toxicity, and a mutagenesis strategy has been adopted to reduce VLS in dgRTA by changing Arg48 and Asn97 that are nearby the LDV sequence in the crystal 3D-structure of RTA (#9). The aspartate D75 (the central aminoacid of the LDV sequence) instead appeared to be critical for RTA toxicity and couldn’t be mutated, thus R48A, L74M, V76A, V76M and N97A mutants were considered. The results obtained with these single mutants showed that R48A and N97A mutants retained RTA full catalytic activity but didn’t elicit pulmonary leak syndrome (in a mouse model). These mutants were used to construct an anti-CD22 immunotoxin that was then tested in a mouse lymphoma model (#9). The immunotoxin containing R48A had very similar pharmaco-kinetics properties as the dgRTA-based immunotoxin and its therapeutic index was even better than the original dgRTA-antiCD22. Thus it was demonstrated that acting on LDV and its surrounding aminoacidic residues it was possible to reduce VLS and thus increase the potential application of RTA-based immunotoxins.
From the aminoacidic sequences comparison shown below, it is clear that many type I RIPs share this (x)D(y) motif, with a major group of them strikingly maintaining them at the same position as in RTA. Other plant RIPs show a somehow different localization along the protein with some other also showing a potential double VLS-eliciting peptide. In particular seed-extracted saporin SO6 has a “LDL” sequence at the N-terminal and a “IDL” sequence more close to its C-terminal region. On the basis of our alignments, we may speculate to cluster the analysed RIPs’s sequences in three different clusters, the first one (including Ricin A chain (RTA), Trichosanthin (TC) and the RIP from Momordica balsamina, see Fig. 2 A) sharing LDV (or similar) sequences with Ricin A chain. The two residues that have been mutated in RTA to eliminate VLS (i.e. R48 and N97) however, are not found conserved among RIPs of this first group (see also Fig. 1) and thus we could not easily infer that mutants at the corresponding positions in other plant RIPs of the same group could be effective in reducing the potential VLS problem. What can be clearly noted in Fig. 2A, is that instead residue R56 is well conserved and is positioned to maintain D57 in RTA, thus suggesting that this position could be explored as a potential common residue of this cluster of RIPs to be mutagenized.
If we look at the crystal 3D-structures of other plant RIPs belonging to cluster 2 (including Trichosanthin and Momordin) and Saporin (see fig. 2B), we can notice that the LDS motif is located on alpha helix 4 which is well exposed to the external surface and thus, potentially able to bind to integrins and elicit VLS. Saporin has a peculiar “IDL” motif located just one helix-turn before the L,D,S sequence. Saporin shows potentially another L,D,L motif located at the N-terminus of the mature protein, therefore we may speculate on the presence in saporin of two potential binding sites for integrin. No any further conserved residue within this region could be detected, thus any prediction of potentially mutagenic sites available to reduce saporin-derived VLS is speculative. The only conclusive observation we may drive is that the location of VLS sequence in this second cluster is away from the key catalytic residues, therefore it would be predictable that their mutagenesis may not be detrimental to the catalytic activity.
A third cluster of RIPs (including PAP, PD-L1 and the RIP from Iris hollandica, see Fig. 2C) shows the presence of an L,D,L motif in the αhelix 6. Also in this case it s not possible to infer any other conserved residue around the motif on which eventually try to make mutagenesis to reduce VLS. Also in this case a forst approach could be directed to directly mutagenize the L,D,L motif.
Saporin shows peculiar structural motifs, if compared to other plant RIPs, demanding a specific mutagenesis strategy to by able to explore the real contribution of thsee two sequences in the genesis of VLS.