Appunti Drugs-14-26 PDF

4. Human antibodies are instead complete human mAb, the first created was adalimumab (anti TNF- alpha antibody). Despite its nature as being completely human, there are reports describing the occurrence of production human anti-human antibodies (HAHAs, same name as for the humanize...

4. Human antibodies are instead complete human mAb, the first created was adalimumab (anti TNF- alpha antibody). Despite its nature as being completely human, there are reports describing the occurrence of production human anti-human antibodies (HAHAs, same name as for the humanized ones) in patients. This is due to the immune system of the receiving subject, because even if the immunogenicity is really low we could still have the activation of the immune system. Also this type is produced with recombinant technology. We have also different types of mAb that are modified from the normal mAb to make them less immunogenic or more effective: for example the primatized antibodies are genetically engineered from cynomolgus macaque monkey, they are structurally indistinguishable from human antibodies and so they may be less likely to cause adverse reactions and they are potentially suitable for long-term treatments (an example is lumiliximab, an anti-CD23 antibody). However, the production of antibodies in monkeys is not so easy, moreover it is unethical and they are not approved, at least extensively. We can mention also another type of antibody, the bi-specific antibody, which is particularly useful against tumoral cells: a characteristic of cancer cells is their ability to perform immune-escape, so they won’t be attacked by T cells anymore. If we use bi-specific antibodies, which are structured in order to have one Fab portion directed against a tumoral antigen and the other directed against an antigen of T cells (like CD3, especially in CD8+ cells), we can make T cells become closer to tumoral cells and in this way the probability that they are going to activate against them is higher. If we then consider the fact that the Fc portion of the mAb is going to recruit also NK cells (or macrophages), the final action against the tumor is going to be amplified by the contemporary activation of the T cell. This is really revolutionary because usually infiltrating T cells, once they get activated, become anergic so they cannot get activated against the tumor anymore, while circulating T cells are still able to act against the tumor but they cannot reach it and infiltrate it because the tumor can escape their action. So by bringing T cells and tumoral cells very close we can induce the activation of CD8+ cells against the tumor. Moreover, in this type of antibody, the two binding sites can significantly increase affinity or internalization rates of particular antigens on a cell's surface by binding to two different epitopes on an antigen, and they usually have a higher cytotoxic potential to bind to antigens with low expression level. With the development of antibody engineering, many types of bispecific antibodies have been designed to overcome short half-life, immunogenicity and side-effects caused by cytokine liberation. They include trifunctional antibodies, chemically linked Fabs, various types of bivalent and trivalent single chain variable fragments (scFvs), and fusion proteins mimicking the variable domains of two antibodies. The furthest developed of these newer formats are the bispecific T cell engagers (BiTEs) and mAb2's, antibodies engineered to contain an Fcab antigen-binding fragment instead of the Fc constant region. Why are scFvs so innovative? The part of the antibody which recognizes the antigen is the variable portion, which is made of light chains and heavy chains. The heavy chains have one variable domain (VH) and three constant domains (Ch1, Ch2, Ch3), while the light chains have one variable domain (VL) and one constant domain (CL). If we link together the variable domains of these two chains by a short peptide linker, we can create a molecule that only contains the light and heavy chains of the variable portion and that can recognize the antigen. This is an advantage because the resulting molecule is very small, much smaller than a whole antibody, and so these scFvs could possibly enter the cell and target intracellular antigens, like viral ones (something that an antibody cannot do because of its size; called intrabodies). So the sequence of the gene of the intrabody can be inserted in a vector (very difficult thing to do with a whole antibody, which is encoded by very long and complex sequences), which is going to be transfected inside our cell of interest, then the intrabody will be expressed inside it and it is going to target antigens found inside it (for example in the ER, which is pretty simple to induce because antibodies are normally produced there so you don’t need any particular folding or stability characteristic, or in the cytoplasm or in the nucleus if the target is located there, and in that case a correct folding of the intrabody is required). Antibodies may have different kind of actions inside the body: - one of the most important is the ADCC (antibody-dependent cellular cytotoxicity): macrophages, dendritic cells and NK cells can recognize the Fc portion of antibodies with their FC receptor, once they bind it they get activated and they also start to produce cytokines to activate T cells, so ADCC reaction is really important because it is the first reaction for any type of antibody to trigger the response against particular types of cells, especially tumoral ones. So this is how any type of anti-tumoral antibody works. The ADCC is a sort of “unspecific” reaction because it is something that happens independently from the specificity of the antibody towards the target. For this reason it is only used for anti-tumoral therapies: if we use an antibody against an autoimmune disorder we need to modify it to avoid the activation of the ADCC because we don’t need a further activation of the immune system, it is already too activated. - Then we have the complement-dependent cytotoxicity (CDC), which is also related to the recognition of the Fc portion of the antibody, so it’s again independent from the recognition of the specific antigen. - Neutralization of exotoxins and viruses is instead correlated to the ability of the antibody to recognize specific targets, in this case derived from microorganisms. - Prevention of bacterial adherence to host cells. - Membrane attack complex (MAC) resulting in cytolysis. - Agglutination of microorganisms. - Immobilization of bacteria and protozoa. - Opsonization. In general, the activity due to the specific binding is mainly of blockage of ligands and prevention of their interaction with the receptor. If we have a tumor cell we can use many strategies: we can use the naked antibody, or we can add some radioactive ligands (which are in this way brought in proximity of the tumoral cell) by either using a streptavidin molecule (that binds to radioactive molecules on one side and to the Fc portion of the antibody on the other side) or a bispecific antibody (that can directly bind one Fab portion of the mAb) or directly binding it to the antibody (radioimmunoconjugate); then we could use bispecific antibodies that bind, other than to the tumoral cell, also to NK cells or CD8+ cells; or we could use scFvs bound to the surface of a liposome containing, for example, a drug; then we could use the scFv bound to an enzyme that is able to transform a pro-drug into an active drug; or the antibody could be bound to an immunotoxin, or to a cytokine (for example a pro-inflammatory one to trigger the activation of immune cells). In any case, we can increase the reactivity against the tumoral cell by bringing a toxic molecule, a drug or immune cells next to it. Out of all these strategies that we talked about, the pro-drug one is the one of the most used: it is called antibody-directed enzyme prodrug therapy (ADEPT), after tumor localization and deactivation or clearance of the enzyme from blood and other normal tissue, a prodrug is administered to the patient, and it is going to be converted into a toxic chemotherapeutic by the pre-targeted enzyme at the tumor site. Then another innovative strategy is the Potelligent technology one: we know that antibodies have carbohydrates inside, in particular they have a fucose that, if removed, can increase the ADCC activity. In fact, it was demonstrated that the mechanism behind the enhanced ADCC of a low/no-fucose antibody was its increased affinity to FcγRIIIa (CD16), the major Fc receptor for ADCC in humans. So, this technology dramatically enhances the potency and efficacy of antibodies. Clinical results with Potelligent-enhanced antibodies are expected to show higher efficacy in human patients when compared with antibodies that have not been enhanced. So by increasing Fc receptor binding it overcomes the problem of low clinical responses due to genetic differences in the Fc receptor; moreover, it lowers the effective therapeutic dose of mAb, so you can deliver much more with much less. Potelligent Technology creates antibodies that are expected to be proven safe and well tolerated with no immunogenic concerns and it also reduces the costs of production. When can we use a therapy with antibodies? First of all, since the main use of antibodies as therapeutics is to trigger an immune response (especially against cancer or against microorganisms), the patient must not be immune suppressed, so his immune system must be intact. Moreover, the antibody that we choose to use should have a high serum stability (The concentration of mAbs in serum can be measured by ELISA, electrophoresis on polyacrylamide gels (PAGE) and FACS) and it must be highly specific against a specific antigen expressed exclusively on cancer cells, so that cross-reactivity is limited to the minimum. The best solution would be to produce an antibody against a unique target, but which should also be highly expressed (for example, fibroblast activating protein-FAP-, expressed in tumor-associated fibroblasts). Moreover, the antigen must be present in a homogeneous target population and not show heterogeneity. One important phase in the pharmacokinetics of antibodies is their catabolism, which is actually different from the protein one. There is a specific receptor, called Fc receptor of neonates (because it is highly expressed in the bowel of neonates), which is almost ubiquitous and can bind the Fc portion of any antibody. Once bound, the antibody is going to be internalized in the cell inside vesicles, together with unbound molecules of the same mAb. Then these vesicles will fuse with lysosomes which will digest their content. However, the lysosome enzymes cannot digest the bound antibodies, but just the unbound ones. In this way the bound antibody can be released again in the blood stream once the vesicle is fused back to the membrane of the cell. This particular catabolism is responsible for the long half-lives of the three IgG subclasses. These receptor is mainly present in endothelial cells, however it is also expressed in the bowel, especially of neonates, to transport antibodies from the lumen of the intestine to the blood vessel (which otherwise cannot pass the intestinal mucosa easily). As we said, this binding induces a longer half-life of mAb, however since it involves a receptor, we know that it could get saturated. So, as a consequence, the higher is the serum concentration, the shorter will be the half-life of the antibody: at high concentration Fc receptors are saturated more rapidly and so there would be a lot more unbound Abs (IgG molecules) which are going to be digested. The mAb are administered through the parenteral route (intra-venously, subcutaneously, like adalimumab, and intra-muscularly, like palivizumab). From the subcutaneous and the intramuscular route they enter the lymphatic system and thanks to it they can reach the venous system. This is different from proteins, which can still pass the cells through the inclusion in vesicles or by infiltrating in the junctions and spaces in- between cells, antibodies reach the blood stream through the lymphatic circulation, because they are really big molecules. However, since the flow rate of the lymphatic system is relatively low, the resulting time of the maximum concentration (Tmax) of mAbs in the blood is after 1 to 8 days (and not after hours like for proteins!). The major limitation of the subcutaneous and intramuscular route of administration is that the maximum volume we can inject is 2.0-2.5 ml for the subcutaneous administration and 4.5-5.0 ml for the intramuscular one. This is something which is true for any kind of drug, so not only for mAb, but it is a problem specifically of monoclonal antibodies because the solubility of IgG is 100 mg/ml, so this means that the maximum subcutaneous and intramuscular doses are 200-250 and 450-500 mg/ml, respectively. For intravenous administration there is no limitation of volume, you can administer as much drug as you want (obviously, in a dose that is coherent with the therapeutical window of the drug). Oral administration as the same problems as for proteins. The distribution of mAb is very poor: they cannot pass the tissue easily due to their high molecular mass (MM) and to their hydrophilicity/polarity, so they have to use the lymphatic system. Hence, the transport happens through convention (transport of molecules within a fluid, so the mAb go from blood to the interstitial fluids of the tissues and from the interstitial flids to blood, via the lymphatic system, so they always take advantage of fluids) and through endocytosis which could be receptor-mediated (Fc receptor) and non-receptor-mediated (the endocytosis could be a phagocytosis or a pinocytosis; remember that the difference between the two is that pinocytosis is the internalization of molecules in a liquid phase, while phagocytosis is the internalization of solid material of a big molecular mass). The binding of the mAb to the target could be either specific (if it happens through the Fab portion) or non- specific (if it happens through the Fc portion), in any case we have to remember that the mAb preferentially binds to soluble antigens, while the physiological ligand of the antigen (for example a receptor, if we think about a cytokine) preferentially binds to the ligand on the cell surface. This just depends on the kinetics of the binding. Obviously, if the antigen could only be present on the surface of the cell, the antibody directed against it will still bind to it (same if the ligand is only available in a soluble form), so this high competition with the endogenous antigen ligand is only in the case in which the antigen can be present in both forms. How can the antibodies be eliminated? They are proteins, so they can be digested through proteolysis (but we have to remember that the protective mechanism of the Fc receptor against catabolism is responsible for the long terminal half-lives, for all IgGs except IgG3). However, some types of mAb are degraded in a more rapid way than others, in particular the murine mAb is degraded in few days, while the human is degraded in a few weeks (so the degradation is faster for the antibody that contains a higher percentage of murine chains: murine < chimeric < humanized < human). It also depends on the low affinity of the Fc receptor to non-human antibodies, so the longer half-lives are usually typical of fully human antibody (this is another reason why they are extremely useful). Antibody fragments (especially the intrabodies) show a very short circulation period, they are degraded and digested faster than a normal size molecule. To prevent it we PEGylate the scFv, because PEG can mask the cleavage site of the molecule so it efficiently avoids its degradation by proteolytic enzymes. This is a list of the half-lives of many antibodies (however we can see that there are some molecules reported, Alefarcept and Etanercept, which are recombinant receptors, and they are in this list because they have a similar action as antibodies, they are directed against the TNF). Adalimumab (the first fully human antibody produced) has a half-life of 14 days, while Gentuzumab has a half-life of 66 days (that can vary of 34 days). The recombinant receptors have a half-life that is tendentially shorter than the one of antibodies (for example Etanercept has a half-life of almost 3 days). The binding to the antigen with the Fab region is IRREVERSIBLE. For example when the mAb/antigen complex is located on the surface of a cell, it will be internalized as a whole and degraded. Moreover, if there is an immune reaction against the mAb, the anti-idiotype antibodies are going to be observed after 1 to 2 weeks. As we already said, the adverse effects correlated to the immunogenicity of the mAb depend on: the type of mAb (the lower the extent of the humanization, the more anti-idiotype antibodies are going to be formed), the frequency (multiple dosing often results in anti-idiotype antibodies formation), route of administration (subcutaneous administration is more immunogenic than the intramuscular or the intravenous one, or the reasons we already said for proteins), patient’s genetics (patients with an autoimmune disease are more likely to produce an immune response). Another important aspect is the drug interaction: we know that chemical drugs can interact with each other, because they are all metabolized in the liver by the same enzymes (P450) and so at that level drugs could interfere with one another. This problem does not exist with biotechnological drugs, neither with proteins nor with mAbs. This is because they are proteic molecules, so this means that they are not metabolized exclusively in the liver, but in any tissue, and by multiple types of enzymes (which are all proteases but of different types). So this means that, at low doses, you can administer both biotechnological antibodies and chemical drugs (moreover, you can also administer two biotech drugs together). In this table we can see a comparison between small molecules and mAb: small molecules can easily pass through tissues while mAb cannot; the binding between a small molecule and its target usually implies distribution (and during distribution the drug is present both in the plasma and in the tissues), instead the binding between mAb and its target implies clearance (the interaction of the antibody with the target “clears out” the drug from the circulation); the degradation of small molecules is metabolic, while the one of mAb is proteolytic; for small molecules renal clearance is really important, while for mAb it is uncommon, it rarely occurs just for small peptides; the unbound concentration (drug that has not reached the target) of small molecules could still have an effect on targets of other tissues, whereas for mAb the unbound concentration doesn’t have an effect on targets on other tissues (also because usually mAb have only one specific target), however it could still cause the activation of the immune system (so in this case the side effects are due to this activation, not to an unspecific action of the drug); finally, the pharmacokinetics is linear and independent from dynamics for small molecules, while it is non-linear (you cannot predict exactly how the pharmacokinetics will be) and often dependent on pharmacodynamics and the production of anti-idiopathic antibodies. NUCLEIC ACID BASED DRUGS The main purpose of the use of nucleic acid-based drug is to prevent the mRNA translation, to reduce or prevent the synthesis of the protein encoded by that mRNA, which could be overexpressed or with an altered function. The translation can be inhibited simply by creating a small nucleic molecule (15-25 nucleotides) that is able to bind to our target mRNA. Before the development of this technique, to inhibit the action of an altered protein, we used traditional drugs, however if we block the action of the protein directly by inhibiting its production it should be more efficient. The first nucleic acid-based drug that was developed was the ASO (antisense nucleotides), a single-strand antisense molecule of a short length (15-25 bp), usually of DNA but it could also be of RNA, that is complementary to a portion of the mRNA and binds to it blocking its translation. The reduction of target mRNA is measured on the amount of this mRNA and confirmed by Western blot analysis of the encoded protein. The problem with this approach is that DNA and RNA molecules are digested, inside our body, by nucleases. That’s why the first attempt of producing this kind of nucleic-based drugs led soon to the generation of modified antisense oligonucleotides: they had the substitution of a sulphur atom instead of an oxygen in the phosphate group (obviously, an oxygen that doesn’t participate to the phosphodiester binding, so in a non-bridging oxygen), and this sulphur atom was able to prolong the half-life of the oligonucleotide because it avoided the action of nucleases (it also increased non-specific plasma protein binding). This type of oligonucleotide is called phosphonothioate antisense oligonucleotide (PS), however their half-life still wasn’t long enough to have a good therapeutic result. So these were the first-generation PS, and rapidly a second generation was produced: they are called 2’MOE (methoxyethyl) and they present the addition of a methoxyethyl group at 2’ level. In this way the half-life was greatly increased, without changing the ability of these molecules to bind to their target mRNA. This means that just by making some chemical modification to the molecule we can increase their stability without altering their complementarity and their affinity to the target. The prevention of the translation of the target mRNA occurs in two different manners: in the first case the binding of ASOs to the target mRNA induces the digestion by the RNase H, so the mRNA is rapidly degraded and its translation is impossible; the second way consists in a simple steric block of the translation, because the antisense oligonucleotide just obstructs the normal ribosomal activity. However, 2’MOE ASOs do not recruit RNAse H, so their antisense effect is limited to prevention of translation. In the past 7 years, over 100 ASOs have been tested in Phase I clinical trials, a quarter of which have reached Phase II/III. After the second generation, a lot of other modifications of ASOs have been produced and tested, and they all belong to the third generation ASOs. For example, the morpholinophosphoroamidate (MF) and non-ionic DNA analogues (DNA usually has a negative charge, these types of oligonucleotides have no charge) are highly resistant to nucleases. Which are the FDA-approved oligonucleotide drugs? Until now they are three: fomivirnes (1998-2001), mipomersen (2013), pegaptanib (Macugen, anti-vascular endothelial growth factor, so anti-VEFG, RNA aptamer, so it is actually a different molecule). Fomivirsen is a 21-nucleotide anti-sense molecule used to target the mRNA for the major immediate-early transcriptional unit of cytomegalovirus (CMV). CMV is in fact able to cause an infection in the eye (retinitis), especially in immune suppressed patients. So fomivirsen was used as a therapy for local treatment of CMV retinitis in AIDS patients. This drug worked, but then new drugs against AIDS were invented and these drugs were able to prevent this type of infection, the treatment wasn’t needed anymore. So, despite the initial enthusiasm and unmet clinical need in the late 1990s, the drug was withdrawn by the FDA in 2001. The EMA followed in 2002, when the manufacturer voluntarily withdrew the drug from the market due to low demand. Fomivirsen was used with a 4-week induction phase, with a single injection every other week (so two doses every induction phase), followed by a maintenance phase, in which a single injection is administered every 4 weeks. The only adverse effect that it could cause was ocular inflammation (uveitis). Nevertheless, the success of fomivirsen provided proof-of-concept of the clinical promise of treatments based on antisense oligonucleotides, which was valuable for the next wave of antisense drug approvals, beginning in 2013 with the FDA approval of mipomersen. It is a 20-nucleotide anti-sense molecule with each inter-nucleotide linkage chemically modified as a PS diester and with a 2’-O methoxyethyl sugar. It is targeted against the mRNA of ApoB-100 (apolipoprotein B), for the treatment of homozygous familial hypercholesterolemia (HoFH). Apolipoprotein B is a part of the LDL lipoprotein, so the bad cholesterol, so by reducing the expression of this protein we reduce the formation of LDL. The treatment involved the subcutaneous injection of the dug once a week: this is a real innovation because fomivirsen was used locally, instead mipomersen can be used systemically (because from the subcutaneous administration it can reach every tissue). In the skin it forms a depot that allows the slow release of the drug (which actually has a short half-life because of its nature, even if it is modified). Adverse effects involve erythema or pain at the injection site (subcutaneous), flu-like symptoms, increase in liver fat content in some patients. Pegaptanib (Macugen) is particular and different from the others because it is not a linear molecule, but it has a complex tertiary structure. It is in fact a 27-nucleotide RNA aptamer (and aptamers are oligonucleotides that share some attributes of monoclonal antibodies due to a complex 3D structure, in fact they usually recognize their target through their shape, not through complementarity), it has nucleotides annealed with hydrogen bonds and it has a 2’-O methyl modification. It targets the mRNA of VEGF; it was approved by FDA in 2004 for the treatment of age-related macular degeneration, a condition that could lead to blindness. It is injected locally (in the eye, intravitreal injection) once every 6 weeks (it is a treatment that doesn’t need to be repeated very frequently, also because of the site of injection). Adverse effects involve pain at the site of injection, sudden vision problems and headaches. This is the first type of oligonucleotide drug that was invented, but we could have also other molecules that can work to prevent the translation of mRNA, and they are the RNAi (interference). They are double stranded RNA molecules and they mainly divide into two categories: miRNA and siRNA. These two types of RNA are produced from the region of the genome called non-coding DNA, which represents the majority of the genome, it doesn’t encode for any protein but it actually encodes for RNA molecules with regulatory activities. The amount of non-coding DNA increases in more evolved species; in vertebrates, and in particular in mammals (especially in men), we have the highest amount. This is because the species that are high in the evolution scale have a fine regulation of gene transcription, which is what makes them more evolved. In the group of non-coding RNA (ncRNA) we actually have a lot of different molecules, like rRNA, tRNA, snRNA and snoRNA, other than RNAi, but this last class is the only one used for therapeutic purposes. RNAis are short non-coding RNA sequences involved in the regulation of expression, in fact both miRNA and siRNA, even if with different approaches, induce the block of the translation of specific mRNAs preventing the production of the protein for which they encode, so the purpose is gene silencing. 1. siRNA (19-24 nucleotides): they are transcribed from non-coding regions from RNA polymerase II, they are produced in the form of dsRNA or shRNA (short hairpin RNA). This precursor is then released in the cytoplasm, where it is digested by an enzyme called DICER, which produces a double stranded siRNA (also called siRNA duplex or siRNA-siRNA). Of this duplex, only the antisense strand is used as a guide strand, in fact it is incorporated in the enzymatic complex RISC-AGO2, while the sense strand (also called passenger strand) will be discarded. The guide strand of siRNA, once incorporated in the RISC complex, will bind to the target mRNA because of its complementarity with it, and thanks to this binding the enzymatic complex is going to induce a cut in the target mRNA, in a region called seed region, inducing its cleavage and its subsequent degradation. 2. miRNA (20-23 nucleotides): miRNAs are actually transcribed from introns (so they are still encoded by non- coding RNA, but from the one contained inside genes!) by RNA polymerase II. They are transcribed first in the form of pri-miRNA, a precursor, which is then transformed in pre- miRNA from an enzyme called  Drosha-DGC8, inside the nucleus. Then the pre-miRNA is exported in the cytoplasm (thanks to exportin 5), where the DICER is going to transform it into a double stranded miRNA molecule (miRNA duplex or miRNA-miRNA). Even if the process is quite the same of the siRNA one, we have a difference between these two molecules: siRNA have a complete complementarity with the target mRNA, so also the duplex will be a linear double strand molecule, while miRNA do not have a precise complementarity, and we can understand it from the fact that their duplex contains a sort of “hole” inside it (region in which the nucleotides do not pair with each other). Again, only one strand will be the one used for the gene silencing (the other one is digested), and, incorporated in the RISC-AGO2 complex, it will bind either to the regulatory regions of the mRNA molecule (5’-UTR and 3’-UTR) or to the ORF (open reading frame, so the encoding region). This is because the miRNA doesn’t have a specific complementary to the coding region of the mRNA, so they can target a lot of parts of the mRNA molecule. Finally, the miRNA could also go back inside the nucleus and, always bound to the RISC-AGO2 complex, bind to the promoter region of genes (so of DNA). More than 500 miRNAs are encoded in the human genome, so they constitute the largest gene family; it has been estimated that more than 30% of protein-coding genes can be regulated by these miRNAs. Both miRNA and siRNA are molecules that are normally produced by the cell but they could also be injected inside a cell (or they could be expressed by it), like ASO molecules. What is the seed region? It is a short sequence present on the target mRNA that is usually perfectly complementary to the siRNA, while we don’t have the same level of complementarity between this region and miRNAs. So, if the molecule that binds to the seed region is perfectly complementary (usually siRNA, full match), the mRNA will be degraded through the action of the RNase H; if the molecule that binds the seed region is not perfectly complementary (usually miRNA, partial match) the mRNA will not be degraded, but there will just be a physical block of the translation. In both cases, the result is the repression of the translation and the mRNA will not be translated. However, these different levels of complementarity are index of the fact that siRNA will eventually have a single target (they are complementary to only one mRNA because they have the same sequence), while miRNA have multiple targets (because they are not fully complementary to any of them, so they could potentially target a lot of different mRNAs). We already said that these RNAi are physiologically produced by our cells, in particular miRNAs are the most represented ones. They have a lot of physiological roles, that they carry out through their regulatory action, for example they are involved in organ and tissues development, in stem cell differentiation and maturation, in cell growth and survival, in metabolic homeostasis, oncogenic malignancies and tumor formation (because it can regulate the expression of genes involved in the growth of the cell, so they are unfortunately also involved in these pathologic activities), viral infection and epigenetic modification. But we can use them also as therapeutics. In this sense we can consider them as both New Chemical Entities (NCEs), because they can be chemically synthetized, and New Biological Entities (NBEs), because they have a highly specific target and they have a complex modality of action. So they are different from the other biotech drugs that we talked about, which are proteins and cannot be chemically synthetized. Another big difference is that one molecule of RNAi can target one mRNA but then it can be recycled and target another mRNA molecule, so this means that RNAi based compounds can be used for several rounds of cleavage (obviously it also depends on the concentration because the lower will it be, the less frequently this phenomenon will happen, but at the same time the lower the dosed the less the side effects). So which type of RNAi have been used in the therapeutic field? Systemic administration: this type of administration is generally preferred for any type of drug because in this way it can easily reach every tissue. However, nucleic drugs should be modified prior to a systemic administration because their stability inside our body is very low. The first example is the one of Patisiran: it is a siRNA that targets the mRNA of a protein called transthyretin, that, when aberrantly produced, causes the accumulation of amyloid fibrils in different organs, determining the appearance of a pathology called Familial Amyloidotic  Plyneuropathy (FAP). The siRNA is encapsulated in a lipid nanoparticle formulation (LNP) to be protected from degradation so that the drug can be administered systemically. This drug is currently in phase III trial and it’s being used in infusions every three weeks. The second example is the ALN-PCS02, used for the treatment of coronary heart disease, in particular for the reduction of LDL. In fact this drug targets the mRNA of the pro-protein convertase subtilisin/kexin type 9 (PCSK9), but it is still in phase I. Another siRNA used for the same purpose is the TMK-ApoB, which targets the apolipoprotein B (ApoB) to reduce the production of LDL. However, this drug was halted in 2010 due to unexpected immune stimulation (so not every nucleic drug, even if it works, eventually reaches the approval because there could also be some problems). Both these drugs are encapsulated in LNP like in the first example. Another example is the one of siRNA used against cancer, in particular for the treatment of hepatocellular carcinoma (TKM-PKL1, encapsulated in LNP, targets polo-kinase1, so PLK1) and of neuroendocrine tumors (Atu027, encapsulated in lipoplex nanoparticles, targets protein kinase N3, so PNK3). They are both in phase II. We have then another siRNA, the naked siRNA 2’-O methylated QPI-1002, with a modification that is usually applied to ASOs but that prevents the digestion of the molecule without the need for liposome nanoparticle encapsulation. This siRNA targets p53 in the treatment of acute renal failure, it is in phase II currently. Local administration: the local administration is preferred when the tissue that the drug needs to target is easily accessible from the outside. This is because the systemic administration, even if it is generally always preferred, cannot always be achieved with biotech drugs (especially with molecules so unstable like RNAs), so in that case we prefer to use local administration. One example of this administration is Bevasiranib, a siRNA targeting VEGF for the treatment of the age-related macular degeneration; however, it was discontinued in phase III clinical trials. Other local administration examples could be the ones of siRNA for the treatment of diabetic retinopathy, targeting fibronectin, laminin and collagen IV, of ocular neovascularization, targeting VEGF and of glaucoma (targeting myocilin). We can see how all of these examples involve a retinal administration, and we already saw another nucleic drug used for the treatment of macular degeneration that was applied by this route, so in general we can say that this route of administration is optimal for this type of drugs. Could miRNA and siRNA be toxic? We know that RNA and DNA molecules cannot be administered naked, at least not as they are, they need to be at least chemically modified. Another expedient could be to encapsule them in lipid nanoparticles and polymeric nanoparticles. These complexes are then internalized by the cells through endocytosis. If the endosomal vesicle doesn’t break (no endosomal escape), once the nanoparticles disrupt, the RNAi is released outside of the nanoparticle complex but still inside of the vesicle, where we can usually find toll-like receptors (which were previously on the part of membrane that produced the endosomal vesicles). The binding of the RNAi molecule to the toll- like receptors (especially TLR7 and 8) activates them because the molecule is recognized as exogenous. In this way we could have an unwanted immune response. Otherwise, if the endosomal vesicle breaks (endosomal escape), the RNAi, let out from the nanoparticle complex, is released in the cytoplasm of the cell, but the same will happen with the residual nanoparticles, and especially if they are polymer nanoparticles, they induce the release of cathepsin B and of IL-1b (responsible for fever) and the triggering of inflammation and apoptosis of the cell. So RNAi could be toxic for our body: even if the benefits that these drugs give are still much higher than the risks, some side effects could still manifest. Another aspect to consider is that the toxicity can also derive from the hybridization on the target: if miRNA could have multiple targets, how are we sure that the miRNA that we are using as a therapeutic agent is not actually silencing genes which are not our target? This is the risk of having off-target effects. This risk could be avoided by using a drug with a higher specificity and selectivity for our target, for example siRNA have a higher complementarity for their target compared to miRNA, so using the short interference instead of the microRNA could be a solution. But if our final goal is just to reduce the expression of a protein, and not of totally silencing a gene, how can we know that our RNAi is just decreasing the expression and not preventing it completely? This is a risk called exaggerated pharmacology, and it should be avoided. So in conclusion, the toxicity of RNAi could be hybridization-dependent, and in that case it depends on the molecule itself, which could cause exaggerated pharmacology (the result is an effect much bigger than the one we wanted, so it has a negative impact to the cell) or off-target effects; or it could be hybridization- independent, and in that case it is caused by the chemical modification that we make on the drug to improve its stability or by the delivery agents. Lipid based vehicles (like LNP), cause infusion-related reactions, activation of the complement, and other reactions that we already saw. A lot of RNAi are currently in clinical trials (especially for the treatment of cancer), a lot of them are in phase III trials and one has recently been approved (for the treatment of transthyretin mediated amyloidosis). Also an aptamer called “aptamer-antidote” has been approved against factor IX (coagulation cascade) for the treatment of percutaneous coronary intervention. Other than RNAi and ASOs, other nucleic-based molecules are being tested right now (like shRNA). Moreover, exogenous miRNA can be used also to replace their physiological counterpart: if one pathological condition is given by the alteration on the expression or on the function of a particular miRNA, we administer an exogenous miRNA with the same sequence and function and we can counteract the pathologic condition. In some other cases, especially in cancer, miRNA have been found to have an altered action, so they are used as targets instead of drugs. What about the pharmacokinetics of RNA/DNA-based drugs? For what concerns the administration, they are mostly delivered locally but they could also be delivered systemically, in particular by intravenous and subcutaneous route. The non-parenteral administration could be only made possible with the aid of new formulations. They are rapidly distributed, they have a short half-life because they are degraded by nucleases, but we can modify them to avoid this problem (for example the 2’-MOE can make them last for 10 to 30 days). They usually have a high binding affinity for plasma proteins. After intravenous administration, the greatest accumulation of ASOs occurs in the liver and kidneys (so this could also be used to possibly target these organs), however it is difficult to find them in urine or in the faeces because they are degraded in liver and kidney cells (so you could find traces of the drug); if they are injected subcutaneously they form a depot and they are accumulated in the skin and they are very slowly released to the blood stream and also slowly digested. It would be possible to create new formulations for the gastrointestinal delivery of these drugs (like suppositories, which cover less surface area, enemas, gels and foams, which can reach the sigmoid colon, for the rectal delivery, and capsules and tables for the oral delivery). However, this is just for a local use in the gastrointestinal tract (for example the use of capsules and tables could allow a regional release of the drug in the colon, like in the treatment of Chron’s disease). The ocular delivery is still the most efficient for the local administration until now (with intravitreal injections, but an alternative is low current iontophoresis). Another delivery could be the pulmonary one, through single-dose nebulizers, metered- dose inhalers, and dry powder inhalers, but anyway the aim is always to locally treat pulmonary pathologies. The delivery to the brain is extremely difficult because of the blood-brain barrier, however, since the BBB is only permeable to lipophilic molecules of a molecular weight lower than 600 Da, a possible way of administration could be the use of lipidic nanoparticles, or a continuous intracerebral infusions with a mini-osmotic pump, or we simply create conjugates of streptavidin as carrier to transport ASOs after systemic administration. Another way of delivery could be through the skin, for example by using some chemical modifications of ASOs (alteration in the thermodynamic properties of ASO through a hydrophobic counter cation, like benzalkonium; chemical elimination of the anionic backbone charges of the ASO, to increase the penetration in the skin), or by the use of ultrasound-induced sonophoretion and iontophoresis or by gene-gun, or finally by topical formulations containing permeation agents (however, all of these are still under study, because we are trying to administer biotech drugs with the techniques that we use for chemical drugs). For what concerns the parenteral administration, with the use of liposomes we can prevent the degradation and prolong the half-life of the nucleic molecule. Given the excellent solution stability and solubility, simple aqueous formulations are generally administered by slow infusion, rather than intravenous bolus to control pick plasma levels below those associated with acute toxicity. Lately, we are assisting to the rise of the use of novel formulations, such as functional liposomes (charge-based liposomes; long-circulating Stealth liposomes, in which a fraction of the lipids presents a polyethylene glycol polymer bound to their head-groups so that it can bind a lot of water molecules creating a water cloud around the liposome, hiding it from the immune system; cell-surface targeted liposomes), cationic complexes and microspheres. We could also use PEG in the final formulation, in fact grafting PEG onto liposomes has demonstrated several biological and technological advantages. The most significant properties of PEGylated vesicles are their strongly reduced mononuclear phagocyte system (MPS) uptake and their prolonged blood circulation and thus improved distribution in perfused tissues. Moreover, the PEG chains on the liposome surface avoid the vesicle aggregation, improving stability of formulations. Parenteral administration could be preferably intravenously or subcutaneously for this type of drugs, and a subcutaneous administration is preferable to the continuous intravenous infusion (it is easier, less invasive for the patient, and it allows the gradual release in the blood stream prolonging the half-life of the protein). However, we still don’t have any drug that could be delivered through this route. Oral delivery for systemic administration is still not available for this type of drugs, new techniques are being tested right now because this delivery would be the best for the patient (much easier modality). It is not easy to make nucleic drugs pass through the mucosa of the gastrointestinal tract without having them digested from the nucleases present at this level, so the formulations that are the most effective right now are, as we already saw, mainly intended for a local application. One strategy would be to insert the nucleic based drugs into specific carriers that could allow the absorption from the intestinal mucosa. The issues to consider when designing oral dosages forms for these molecules, like for the other biotech drugs, are their chemical stability and precipitation at gastric pH, metabolic instability within the intestinal environment, low intestinal permeability, high protein binding and first-pass metabolism. Chemical modifications could be an aid against these issues, for example the 2’-MOE derivates have greatly improved nuclease resistance, however we have to avoid that the chemical modifications that we perform will interfere with the ability of the molecule to hybridize to its target. One possibility for the absorption in the gut is the paracellular route (naked ASOs passing through the tight junctions), because certain surfactants, such as bile salts and fatty acids, appear to facilitate this absorption modality, however it is not so easy to achieve and the problem with the stability of the molecule remains. When ASOs are delivered intravenously they directly pass through the tissues, rapidly, and then they slowly return back to the systemic circulation. The elimination occurs predominantly in the tissues by the nucleases. The passage through the endothelial cells of the vessels depends on the type of endothelium that we have in a specific organ: for example in some organs, like the liver, we have discontinuous capillaries, which allow an easy passage of the drug; in muscles we have fenestrated capillaries, which still allow a rapid passage, even if in a lower way compared to the discontinuous ones; however we could have also continuous capillaries (like in the brain, with the BBB) that allow only the passage to lipophilic molecules or non-charged molecules, so DNA cannot pass this type of endothelium. In some inflammatory conditions however the capillaries are “leakier”, so they can allow an easier passage through their endothelium, which becomes more permeable. So in some conditions, like tumor or other pathophysiological conditions in which we could have an inflammation, the structure of the vasculature can change and the passage of charged molecules could be easier. So a variety of delivery systems have been evaluated to transverse the biological barriers, especially in order to reach the nucleus in target cells, thus delivering drugs. So for example the delivery of naked plasmid DNA could be either intravenous, intra-muscular or intranasal. The half-life of the molecule depends on its tertiary folding: if it is linear it increases (median of 11 minutes), if it is supercoiled it decreases (1,2 minutes). This is because the higher the complexity of the tertiary structure the more rapidly it will be degraded. However a circular molecule of DNA usually has the longest half-life (21 minutes). Introduced DNA is largely degraded in most tissues by 24 hours. Non-naked nucleic acids could be administered by either liposomes or polymers and they are both polycationic carriers. Once the DNA or RNA molecules complex with these cationic molecules, they enter inside the cells through endocytosis, forming an endosomal vesicle. These polycationic carries are also used for transfections. From these point of view, these biological complexes are able to transport molecules of any sizes, from the smaller ones to the bigger ones, so they have a high loading capacity, which is an aspect that for the delivery of nucleic drugs is not so important (they are all very small), but for the delivery of plasmids for the gene therapy it is crucial, because for example viruses (for transduction) are not able to do it. Moreover, they can enter the cell easily (especially liposomes, because they are made of lipids) and they are biocompatible (so they are not toxic for the cells). 1. Polymer-based vectors: they are polycationic carriers which could naturally occur or be chemically synthesized. Examples are histones, protamine, cationized human serum albumins, chitosan, cationic peptides (e.g. polylysina, polyhistidine, polyarginine), polyamines like polyethyleneimine (PEI), and polyamidomide dendrimers (PAMAM). Of these, the most used is PEI. Cationic polymers spontaneously form complexes, called polyplexes, with nucleic molecules via electrostatic interaction, because they are negatively charged. Moreover, the cell surface membrane is negatively charged due to the presence of glycoproteins and glycolipids, so the cationic polymer complex spontaneously interacts with the membrane too. That’s why we need polycationic molecules to transport nucleic-based drugs in the cell. However, such formed complexes do not have any cell-specificity, which means that they could possibly deliver the nucleic-based drug to any cell in our body because they interact with the cells just because of their charges. This event could cause off-target reactions, so we need to avoid it: we can add a specific ligand on the surface of the polymer complex and the ligand will specifically bind to its receptor on the specific cell that we want to target. In this way there will be no unwanted or unspecific deliveries to cells that are not the target of the therapy. There are some problems with cationic polymers, first of all they could aggregate with erythrocytes, causing an event of acute toxicity. This generally happens in the small capillaries of the lungs once the polymer complex is administered intravenously: at this level, polyplexes usually interact with erythrocytes causing their aggregation. Moreover, since they have a net positive charge, they can also bind negatively charged plasma proteins, for example albumin: if the polyplexes bind to albumin, they are not free to reach the tissues (just like any other chemical drug that stays bound to albumin, it is released slowly from the blood stream and they slowly reach the tissues). How do we solve these problems? Just masking the positive charge of the cationic polymer complex with PEG, reducing the surface charge and so increasing the circulation time (less albumin interaction and reduced uptake by the liver) and stabilizing the particles against aggregation with erythrocytes.  Schematic illustration of shielded polyplex formation and intracellular de-shielding: (A) DNA is condensed with a mixture of polycations PC (condensing agent), PC–PEG (shielding conjugate), and PC–ligand (for receptor-mediated uptake, for example transferrin, specific for muscle cells) to form a shielded polyplex for systemic circulation. (B) Receptor-mediated uptake of the polyplex into the endosomal vesicles. (C) Acid-triggered de-shielding and exposure of the positive inner polyplex core. So PEGylation and cell-targeting approaches (attachment of an antigen for a specific receptor, specific delivery approach) can be combined. How are these complexes of polymers and DNA distributed in our body? All the data that we have on distribution of polymers come from studies on animals: once in the circulation, the complexes of polymers with nucleic-based drugs (DNA/RNA-PEI) can be found first in the lungs (after minutes) and then in the liver. In the lungs, as we already said, aggregates with erythrocytes could get formed and get trapped in the capillaries, causing acute toxicity (possible lung embolism). However, once into the body, the cationic polymer will last only a few hours (short half-life), so for a therapy to be effective you need multiple injections per day. This is something which should be avoided: the higher is the number of administrations the higher is the probability to have toxic effects. To avoid frequent administrations we should find a way to prolong the half-life of the drugs, but it is still under development. 2. Lipid-based vectors (liposomes): cationic complexes made of cationic lipids (the positive charge must always be present to allow the spontaneous complexing with the nucleic molecule and the interaction with the cell). The lipids have a cationic head linked, through a short linking molecule, to a hydrophobic tail. The shapes of the complex of liposomes with the nucleic molecule could be multiple: we could have multilamellar structures, spaghetti structure and beads on a string structure. The mechanism of entering a cell is the same for liposomes and polymers (endocytosis), and again to avoid unspecific delivery we have to add ligands to the complex. The problems that we may encounter are again the same as polymers, especially with plasma proteins. Their distribution is a little bit different compared to the polymers: from the blood stream they reach the lungs, then the liver, the spleen, kidneys, heart and then again the blood. Their half-life is short, they usually reach the lungs within the first hour. Keep in mind that to deliver nucleic-based molecules we could also use viral vectors, but they are only used for gene therapy, not for ASOs and RNAi technologies. The most used virus is the recombinant adeno- associated virus, they induce a stable transgene expression in diving post-mitotic cells, they lack human pathogenicity, they can possibly target multiple tissues (lungs, neurons, eye, liver, muscle, haemopoietic progenitors and endothelial cells) and they are specific for the target cell (viral entrance of the cell happens through specific receptors).