Appunti Drugs - Drug Production Process PDF

everything under controlled circumstances, in order to avoid problems with different batches, however there could still be some differences between drug aliquots produced in different producing cycles, for example in quantity. The vials of the cell bank all include frozen cell lines that are able, once thawed, to start a new production of a functional protein. This system is called two- tiered cell banking system. The Master Cell Bank is created because a daily interaction with it would pose an unacceptable risk for any manufacturer (because the majority of the cell line stock will be lost if something goes wrong). For this reason, portions of the MCB are taken-off and stored separately for “day-to-day” use as the feed stock for each production run, soon after creation. These are the working cell banks, and they may be accessed every day, while the remaining stock of the MCB and of the working cell bank are very securely stored away in separate storage tanks. While the WCBs are accessed on a regular basis, the MCB is essentially left untouched as a reserve in the event the WCBs are compromised. Every vial is stored stably in liquid nitrogen at -150 degrees Celsius protecting them over the drug’s entire life cycle. How do we prepare a batch for the production? First, we have the thawing of a vial from the working cell bank, then we seed it in a starter culture, then in a production-scale starter culture, then in a production-scale bioreactor. A bioreactor is a tank that is big enough to contain several thousand or tens of thousands litres of media for the growing of a production-scale cell culture (it is also used for microbial fermentation, in fact the bioreactor can also be called fermenter). So eventually, from a small vial, we obtain litres and litres of production cell cultures. The right temperature for the growth of host cells has to be established prior to this process, and bioreactors allow the control of this parameter. These in the picture are all the steps of the upstream process. How do mammalian cell lines grow in cultures? They require very rich and expensive media, they also require antibiotics and they have to be controlled because they are more fragile than microbial cells due to absence of an outer cell wall. These cells usually secrete the recombinant protein in the medium, so it is easier to recollect it. 2. DOWNSTREAM PROCESS: it mainly consists of the purification process, so it follows the production process. First of all you have to collect the protein, and this step depends on which type of cell line you used for the production process: if you used bacteria you have to break down the cell wall, while if you used mammalian cells you can skip this step and, most of the times, the product recovery is simply performed by harvesting the extracellular fluid (because the biopharmaceuticals are secreted into the media as extracellular proteins). The cell disruption can be performed physically (sonication, homogenization, alkaline conditions exposure, …) or chemically (detergents, solvents, antibiotics and chaotropic agents). However, if you use chemical reagents, then you will find them in the final product so you will have to remove them during purification; moreover, detergents could cause protein denaturation and precipitation. To avoid it, the best choice is to use homogenization (agitation with abrasives), even if also in this case we have to make sure that we don’t use a temperature that is too high, because in that case the proteins can denature. Alternatively, glass beads are used in cellular agitation, but then beads need to be removed, so homogenization is still the best choice. In fact, an efficient cooling system minimizes protein denaturation and inactivation. Upon completion of the homogenization step, cellular debris and the remaining intact cells can be removed by centrifugation or by microfiltration (0,1-10 mm diameter of pores). Then also nucleic acids must be removed (nucleic acids significantly increase the viscosity of the cellular homogenate): nucleic acid content in the final preparation can be, at most, a few pg/therapeutic dose (they are allowed because it is impossible to remove completely nucleic acids from the homogenates). They may be removed by treatment with nucleases, an efficient, inexpensive and innocuous treatment. After the removal of nucleic acids you have to ultrafiltrate the homogenate. For ultrafiltration we use membranes with pore diameters of 1-20 nm so that we can separate molecules based on size and shape. Ultrafiltration doesn’t alter the bioreactivity of the protein, it has high recovery rates (99%) and a short time for recovery and little ancillary equipment is required. In this way, we remove any possible debris that is still contained in the homogenate, because it is going to be retained by the membrane. The proteins of the homogenate will pass the membrane driven by high pressures, they will be eluted and recollected. Instead of using the simple ultrafiltration we could use diafiltration, which is based on the same system but the volume that is inserted in the column of filtration is the same that we collect at the end, because we add the homogenate continuously (so the level of the reservoir is maintained at a constant volume). It is used to reduce or eliminate low molecular mass molecules from a solution (salts, ethanol, buffer components, amino acids, peptides, added protein stabilizers, etc.) and it is generally preceded by an ultrafiltration step. So now we obtained a solution containing only a mixture of the proteins of the cell line, which contains also our protein of interest. Now all contaminants of potential medical significance must be removed, and we do this by chromatography. We have several types of chromatography but there are some specifical types that we use for biotech drugs. Chromatography is based on the characteristics of the specific protein: each protein has a specific size, shape, overall charge, surface hydrophobic groups, … that we define, as a whole, as the chromatographic fingerprint of the protein. In fact, we can differentiate a protein from the rest of the proteic solution on the base of these characteristics. More than one chromatographic type has to be used to obtain a solution containing only our protein of interest, and so a combination of 2 to 4 different chromatographies have to be used in the downstream processing in order to have a highly purified product (at least a level of 99.9% of purity has to be achieved to be sure that, by administering the final product, we don’t have any unwanted reaction; a 100% of purity is not obtainable, so we can just get close to this value). So which types of chromatography can we use in the downstream process? ▪ Size-exclusion chromatography: it is the simplest type because it separates the proteins based on their size and shape. Larger molecules cannot enter the matrix and so they will be quickly eluted, while smaller proteins can enter the gel beads (the smaller the protein, the longer it will be retained in the matrix and so the later it will be eluted). For this reason, if we know that we have to isolate a protein with a high molecular mass we decide to use it at the beginning of the purification process. Instead, when the protein is small, size-exclusion chromatography is employed towards the end of a purification sequence when the protein relatively pure is concentrated in a small volume. Gel matrices that we can use with this technique are for example dextran, agarose, acrylamide or vinyl polymer. Although this chromatography is effective, it results in a significant dilution of the protein solution. ▪ Ion-exchange chromatography: it is the most used chromatographic technique. Some amino acids are negatively charged (aspartate and glutamate), some others are positively charged (arginine, histidine and lysine), so depending on the composition of the protein of interest the protein will have a total positive or negative charge. The charge of proteins depends on the pH of the medium in which they are in solution, and the pH value in which the protein doesn’t have any net charge (positive charges=negative charges) is called isoelectric point (if the pH is below this value, the charge is positive, if the pH is above this value, the charge is negative). The majority of proteins are negatively charged, which means that they contain a higher number of negatively charged amino acids. So if we use a matrix containing positively charged beads (anionic exchanger), we can separate the positively charged proteins, which will be eluted, from the negatively charged proteins, which will be retained by the matrix. Obviously, the electrostatic attraction that makes the negatively proteins retained in the column, is reversible, so if we change the pH of the buffer of elution (salt- containing buffer) we can change the electrostatic interactions between the proteins and the matrix, making the elution more selective. ▪ Hydrophobic interaction chromatography: some amino acids are hydrophobic, which means that they tend to aggregate together and be surrounded by water molecules. So if we use a matrix on the chromatographic column that is made with hydrophobic groups (phenyl Sepharose gel or octyl Sepharose gel for example), the hydrophobic group of hydrophobic proteins will tend to aggregate with it, while proteins with a less hydrophobic nature will be eluted first (the more hydrophobic the protein, the tighter the binding). Elution is obtained with a buffer of decreased ionic strength. ▪ Affinity chromatography: we use beads made with agarose (or resin), a spacer and a ligand, so a specific protein that is able to bind another specific protein. If for example we need to purify a cytokine (IL-2 for example), we can use, as a ligand in the matrix, a part of the cytokine receptor. Elution can be obtained by changing the pH of the buffer, changing the ionic strength of the buffer, changing the detergent or by using a competing ligand. Affinity chromatography should be used as early as possible in the purification procedure. The main advantage of this type of chromatography is that it is highly specific and selective, and thus the increase in purity is of over 1000-fold. However, many ligands that we usually use in this technique are highly expensive and exhibit a poor stability, and the ligand coupling techniques are chemically complex, hazardous and costly. Moreover, the ligand of the matrix can leach from the column and this can be a problem for the purification. Another type of affinity chromatography is the Abs affinity chromatography: instead of having a protein (like a receptor) has a ligand we use an antibody, which is even more specific for our protein of interest. However the drawbacks are the same as before. Elution is achieved by: altering the pH of the buffer, using chemical disrupting agents such as urea or guanidine, through irrigation with a glycine-HCl buffer (pH 2,2-2,8). This Abs affinity chromatography is used for purification of recombinant blood factor VIII. ▪ Dye affinity chromatography: dyes such as Cibacron blue and Procyon red have been used as adsorbent because of their binding to certain proteins (natural tendency). We know which type of proteins are able to bind to these dyes so we can exploit them for purification. The advantages are that this method is available and inexpensive, the chemical coupling to the matrix is easy and safe, resistant to chemical degradation and the leakage of the dye is easy and recognizable, however it is not possible to predict accurately which protein will bind to these dyes because as we said the binding is highly unspecific and a lot of proteins could be retained by the column. ▪ Metal chelate affinity chromatography: it is based on the principle for which a lot of proteins contain basic groups, most of the times coming from histidine residues. These groups can bind through weak bonds to metal ions. So we can bind to the matrix beads, through the use of a meta chelator, metals like zinc, nickel and copper, and proteins with basic groups will be thus retained in the column. Elution is performed by lowering the buffer pH. This type of chromatography is usually used in the first steps of purification if we know that the protein that we want to isolate has a lot of histidine residues. ▪ Chromatofocusing: it is based on different pH beads positioned in order in the column, from a more acidic pH to a more basic pH (the matrix is positively charged). The separation is based on the isoelectric point of the protein: since the beads are positioned from a more acid pH to a more basic one, the proteins which have a high isoelectric point will be charged positively when the pH of the column is lower than their isoelectric point, so they will have a positive charge that is repulsed by the matrix; proteins which have a lower isoelectric point will charge negatively at pH values that are higher than their isoelectric point, so they will interact with the positive matrix. So negatively charged proteins absorb to the anion exchanger, positively charged proteins flow down until they reach a point where the pH value equals their own PI. Elution is obtained by changing the pH of the buffer. The advantage is that this technique is high resolving, however proteins precipitate more easily at their isoelectric point and the low salt concentration used in the buffer can induce aggregation (detergent addition can help). Moreover, on an industrial scale this technique is not so convenient, because of economic factors (matrix and eluent are expensive) and complexity factors (the set of the pH gradient is not easy). However, this technique is effective when used in conjunction with other chromatographies. ▪ High performance liquid chromatography (HPLC): it is not a different type of chromatography, it’s just a different way of performing traditional chromatography. So it can be applied to any type of chromatography, the difference is in the elution time because it uses a pump that allows to reach high pressure to elute proteins (enhanced separation in shorter periods of time). It is probably the most widely practiced form of quantitative, analytical chromatography today due to the wide range of molecule types and sizes which can be separated using HPLC or variants of HPLC. For example, insulin and IL-2 are purified with this technique, because it allows us to reach a level of purity that is really close to 100% (the higher is the number of times that the protein has to be administered to the patient, the higher level of purity we need to reach with the downstream process). Various chemical groups may be incorporated into the matrix beads technique such as ion exchange, gel-filtration, affinity, hydrophobic interaction and reverse-phase chromatography. The advantages are that is has a superior resolution due to the reduction in bead particle size, resulting in sharper peaks, the increased flow rates improve fractionation speeds (from hours to minutes) and it has a high degree of automation (good for industrial production scale). The drawbacks are that it is expensive and that it is employed almost exclusively in downstream processing of low-volume, highly-value proteins. Sometimes, the proteins that we have to purify, are not suitable to be isolated from the others as they are (because of their particular chromatographic fingerprint they would require more than 4 chromatographic steps, which are a lot). So, to isolate them we can add a tag: through genetic engineering techniques we induce the incorporation of protein tags on the aminoacidic sequence of the protein of interest, and then we exploit these tags and their characteristics to isolate our recombinant drug. For example, we can add a tag of poly-arginine or poly-lysine, so that we can use their positive charges to perform an ionic-exchange chromatography; or we can add a hydrophobic amino acid tag to exploit it to perform a hydrophobic interaction chromatography; or we can add a poly-histidine tag to the protein so that we can perform the metal chelate chromatography to isolate it. However at the end of this process, our protein of interest will still have the tag attached, and it needs to be removed because it could be immunogenic. How do we remove it? We could use chemical agents (like cyanogen bromide or hydroxylamine) or enzymatic agents (endopeptidases, like trypsin, factor Xa or enterokinase, or exopeptidases, for short tags). Enzymes are more effective than chemical agents but obviously we have to pay attention because enzymes could possibly cleave also our recombinant protein. After the tag is cleaved from the protein we remove it from our final product by performing another chromatography. All of this suggests that the process of production of a biotech drug must always be preceded by a step of study of our protein and of its characteristics, in order to know if we have to use a tag (that must be expressed when we produce the protein in the cell lines, so it has to be prepared during the cloning of the gene). Once our protein is purified after all the chromatographic steps, it will be in solution, in a buffer that needs to be removed. In this process we also have to be sure that the protein is sterile (sterile product means that we have removed all the possible bacteria, viruses or infectious agent that could contaminate our drug). Moreover, we know that proteins can be easily denatured and their activity can be easily altered, with chemical agents (oxidizing agents, detergents), physical agents (extreme levels of pH and elevated temperatures) or biological agents (proteolytic degradation) for example, so we also have to make sure that the protein won’t be damaged and denatured during all of these processes, and we do that by removing the possible damaging agents and avoiding extreme conditions. So the final product formulation involves the addition of various excipients, the filtration of the final product through a 0,22 micrometres absolute filter in order to generate a sterile product, followed by its aseptic filling into final product containers, the lyophilization if the product is to be marketed in a powdered format (how stable is the protein in the solution determines the decision to market the product in liquid or powder form). Before the lyophilization though, we need to stabilize the protein and its function, because different molecular mechanism can underpin the loss of biological activity of any protein. These mechanisms can be divided into two groups: covalent modification and non-covalent modifications. Non-covalent alterations are partial or complete denaturation of the protein; covalent alterations involve hydrolysis, deamidation, imine formation, racemization, oxidation, disulphide exchange, isomerization and photodecomposition. At the end of each chromatographic process we are still not at the highest level of purification that we could reach (especially in the initial steps), so this means that some other proteins, in particular enzymes, could remain in the final solution. If the enzymes are proteolytic enzymes (in particular serine proteases, cysteine proteases or metalloproteases), the protein could undergo proteolytic degradation (especially if it is denatured). This is why every step that follows the chromatographic separation must be performed at low temperatures, so that proteases do not activate. Other precautions that we could use to avoid the activation of these enzymes are for example to minimize the processing times and to use specific protease inhibitors (however some of them are toxic, so inappropriate for the biopharmaceutical processing). Another problem that could occur in the chain of a protein is the deamidation: the aminic residues of asparagine and glutamine can be deamidated (by hydrolysis) and so the two amino acids can become acids (aspartic acid and glutamic acid respectively). This could happen in insulin for example. Oxidation is another covalent alteration that could occur: in the presence of oxygen (caused by air), sulphur atoms present in methionine or cysteine can oxidize, and sometimes disulphide bonds can be formed. Oxidation of methionine is favoured under conditions of low pH, and in the presence of metal ions. For example we could have the oxidation of all three methionine residues in hGH (human growth hormone), resulting in an almost total inactivation of the molecule, while 2 oxidations out of 3 do not affect hGH activity. Oxidation can be best minimized by replacing the air in the headspace of the final product container with an inert gas such as nitrogen. Disulphide bonds can also be formed inside a chain (intrachain), changing the tertiary structure of a protein and resulting, most of the times, in the aggregation of individual molecules, while if disulphide bonds form between two different chains (interchain) we have the aggregation of different proteins. This phenomena are generally called disulphide exchange. Each one of these modifications can alter the structure, and thus the function, of our protein of interest, which must be kept in its native form. They could occur at any step of purification, so both during chromatography and after chromatography (for example oxidization could happen because we have potent oxidizers in solution but also just by contact with air, so it could happen also at the end of the process). To avoid it we perform all the steps at low temperatures or by adding chemical reagents that prevent these modifications. Once we have purified our protein and once we are sure that it was kept at its native form, we have to stabilize the protein. This is performed by adding stabilizing excipients, like human serum albumin. The human serum albumin does not alter the activity of the protein (obviously we have to use a recombinant one because if we used the one extracted from blood directly there would be the risk of transmitting infections from the blood of the donor). It is stable at low pH or at elevated temperature and it is very useful because it could work as an alternative target for proteases that are eventually still present in traces. It is also an effective cryoprotectant and once injected in blood it doesn’t cause anything wrong because we already have it in our blood (if we inject little quantities of it nothing happens in our body). Human serum albumin is the most used stabilizing agent, but we could also use amino acids, like glycine (used for IFN, EPO, factor VIII, …), which reduces the surface absorption of products, inhibits aggregated formation, stabilizes the conformation of the protein and prevents heat denaturation. We could also use polyols (like glycerol, mannitol, sorbitol, PEG and carbohydrates) or surfactants (protein denaturants which reduce the surface tension and increase the solubility of the protein, they avoid the mis-folding of the protein and they reduce the rate of protein denaturation at interfaces). The final step before the lyophilization is the sterilization: we pass the protein solution in a filter with pores of 0.22 micrometres which can get rid of bacterial cells. In this way we are 100% sure that we don’t have any bacteria in the solution. However, we could also have other types of microorganisms, like viruses, which can infect mammalian cell cultures. The filters that we use do not get rid of viruses so we have to change filters and use the ones that have 0.1 micrometres pores. Our sterilized protein then is ready to be lyophilized: lyophilization is the process through which we remove the solvent directly from a solution while in the frozen state, obtaining a powdered product (3% water). It is extremely useful to preserve the biological properties of the product, however it can promote the inactivation of some proteins, so specific excipients are added (e.g. HSA), along with amino acids (e.g. lysine, glycine). The whole process is performed while decreasing the temperature and, while it lowers, it induces the formation of ice crystals, instead the protein will aggregate and separate from the water molecules. The crystalized water is removed, while the temperature is still lowered until molecular mobility ceases (the lowering of the temperature is responsible for an increase in the viscosity of the solution, the mobility ceases around -60° of temperature). Finally, we apply a vacuum which is going to remove all water crystals (through sublimation), other solutes and salts, so that eventually only the protein will remain. After this process, a label is applied on the final product, which contains information on the name, the strength, the batch number, the date of manufacturing, the expiry date and the storage conditions of the drug. First, though, we have to perform a quality control of potency, sterility, final volume, absence of endotoxins or other toxic substances. In this phase, some mis-labelling errors could happen, so we have to pay attention. This process is performed for any lyophilized compound, with the only difference that, if the compound is intended for laboratory practice, it cannot be used for therapeutical purposes because it will not be as highly purified as biotech drugs (in fact in the vial you can read “for laboratory use only”). Once you have your final lyophilized product, before marketing and using this drug you have to perform other quality control analyses on your drug, because you have to be sure that it is in its entire form, that there are no aggregates, no shorter form of the protein and that it works. So you take one out of the thousands vials that you produce and you test it. What do you test? You test potency (potency must be analysed not only to check if the drug works but also to assess the right dosage of administration of the protein, so we have to check the potency of the drug at a specific concentration) and safety (you have to be sure if there are no potential contaminants in the final product). Which types of contaminants can we find inside our final product? We can find residues of proteins added to remove tags, proteins from the cell culture media (bovine serum/foetal calf serum), endonucleases used to degrade the DNA in microbial cells, … Why do we still find them even if we used different purification steps? There is the possibility that these proteins that we find in the final product had the same size, isoelectric point, hydrophobic level, … than our protein of interest, so even if we purified the solution we still find them in the final product. Obviously, these residual proteins need to be removed because they can be immunogenic, especially if repeatedly administered with our drug. Other impurities may be derived from the modified product, due, for example, to covalent modifications. So some proteins may have a different tertiary structure (wrong folding during the production), so they may also be biologically inactive and thus they could confer to the product a reduced potency. Some modified forms could also be biologically active, but they could have modified pharmacokinetic characteristics. Modified forms may be immunogenic too, so they must be removed. The removal of altered forms of the protein from the product happens usually in the downstream processing, exploiting the different characteristics that these modified proteins have compared to our protein of interest (for example, if the modified form is shorter than the original protein, it can be removed through the size-exclusion chromatography). Different techniques can be used to detect any protein-based impurities that may be present in the product, like non-denaturing electrophoresis, SDS gel electrophoresis, capillary electrophoresis (capillary with a diameter of 20-50 microm; this technique is highly costly so, even if it allows to analyse the product in a fast way, we tend to not use it), peptide mapping, isoelectric focusing, HPLC, mass spectrometry (most effective and precise way to determine the molecular mass of many proteins) and so on. Two-dimensional electrophoresis is really useful to detect protein-based product impurities because it separates proteins based on their molecular weight and isoelectric point, instead an SDS-PAGE only separates protein based on their molecular mass, so if contaminants have the same weight as the product they will go undetected as they will co-migrate with it. So two-dimensional electrophoresis is preferred because it allows to determine product homogeneity (guaranteed by the appearance of a single protein band) and to verify the stability of the product over the course of their shelf life. HPLC can also be used to determine the purity of high and low MM of the product; for what concerns mass spectrometry, we can use a specific type called electrospray mass spectrometry (small, highly charged droplets are formed by electrostatic dispersion of a protein solution through a glass capillary subjected to a high electric field, then protein ions are desorbed from the droplets into the gas phase, assisted by evaporation of the droplets in a stream of hot N2 gas; the last step is the analysis of the protein ions in a mass spectrometer). If we find contaminants we must perform additional purification processes, like chromatography again. How do we test the product potency? Potency is expressed as «units of activity per vial or dose» or «mg» of product. We have different approaches to determine product potency: Bioassays: you reconstitute the protein (just by adding water to the lyophilized product) and then you perform tests to see if the protein still works. For example if we have to test interferons, we can use them in a mouse model infected by a virus and see if the interferon, which works as an anti-viral protein, works. Other than animal models we can also test the efficiency of the protein in single organs, tissues or cell lines. All bioassays are comparative (which means that the product is tested against a standard). The disadvantages are the lack of precision (the test is influenced by the physiological state of the cell or of the animal), the times of performance (some tests could require weeks) and the high costs (especially if they involve whole animals) Immunoassays: radioactive assay (RIA, quite old) and ELISA are rapid way of testing the potency of the product. They are very effective for the testing for the quantification of the working protein in the final product, but we cannot calculate the quantity of the modified protein in the vial. However, if the quantity of functioning protein calculated with these assays is not the same as the product that we expect to have in the vial we can understand that something is wrong. Other disadvantages involve the fact that these tests are based on the immunological reactivity of the product, which not always correlates directly to the biological activity. The determination of the protein concentration can be assessed also with simpler methods, like absorbance. However, these techniques allow to quantify the total amount of protein present in the vial, without differentiating between working protein, correctly folded protein and modified protein. So ELISA is better because it allows to differentiate between the actual working product and the modified one. But we can use immunoassays also to detect the presence of product-unrelated contaminants (so not the modified forms). The “blank run approach” is performed: we construct a host cell identical in all the characteristics to the cells that we used for the production of the product, except that it lacks the gene coding for our desired protein. The two cell lines will undergo the same processes during the upstream and downstream processing. Finally, the protein are extracted from the cells that do not present our protein of interest are used to immunize horses or other animals. Then, we collect the serum of these animals and purified polyclonal antibodies directed against the protein of that specific cell lines are used in RIA or ELISA approaches against the final product of the cells lines that produced our protein of interest. If these antibodies recognize antigens in this test it means that in the final product we still have some contaminants coming from the proteins of the cell line in which we produced our protein. To verify the protein identity (so to be sure that the final product is actually made of the protein that we wanted to produce and that the primary sequence of the protein is conform to the licensed product specifications and is not modified) we can perform several tests: amino acid analysis, peptide mapping, N- terminal sequencing or analysis of secondary and tertiary structure. Amino acidic analysis is the analysis of the range and quantity of amino acids present in the product just by inserting the latter in an automated amino acid analyser (which is able to study peptides or small polypeptides with a molecular mass lower than 10 kDa). The disadvantages of this technique that limit its usefulness in biopharmaceutical analysis are that it has to be performed in hydrolysis conditions, which could destroy or modify certain amino acid residues (like tryptophan, tyrosine, serine), the method is half-quantitative rather than quantitative and the sensitivity is at best moderate, so low levels of contaminants may go undetected, especially if the product has high molecular mass. However, there is no test that can be considered complete to identify the protein so we still perform this one combined with other test. We can in fact perform also a peptide mapping: it takes advantage of an endopeptidase digestion, so that the protein is cut out in small pieces by enzymes like trypsin, or by chemical agents (the most used is cyanogen bromide, which cleaves the peptide bonds on the carboxyl site of methionine residues). Then the digested protein can be run in bidimensional electrophoresis so we can separate the single polypeptides based on molecular mass and number of amino acids. This allows to evaluate if there is even a single point mutation in the product’s gene that alters the primary structure of the protein, because trypsin won’t be able to digest the protein like it would with the native one so the results of the electrophoresis will be different that the ones expected. Usually single amino acid substitutions, deletions, insertions or modifications can be detected with this method. However we could have a problem with this method, because if a reagent generates only a few very large peptides, a single amino acid alteration in these peptides will be more difficult to detect than if it occurred in a much smaller peptide. Instead, a large number of very short peptides would not allow to resolve all the peptides from each other, so an average of 7-14 amino acids per peptide would be the best for this type of technique. Another test that we can use is the N-terminal sequencing: we have machineries that can sequence proteins and so they allow us to identify the presence of modified forms of the product in which one or more amino acids are missing from the N-terminus (up to the first 100 amino acids from the N- terminus). This sequencing is known as Edman degradation, it is based on the use of phenyl isothiocyanate which can bind to the N-terminal of the protein and allows the identification of the amino acids. The last method that we can use is the analysis of secondary and tertiary structure through X-ray crystallography (which is very complex and highly expensive, so we tend to avoid it for biopharmaceuticals) or through spectroscopic methods (like far-UV circular dichroism, which shows us spectral features related to the structure of the backbone of the protein and can be used to predict the secondary structure of the product). One other step must be performed before the commercialization of the drug, and this step is mandatory if we produced the protein in bacterial cells: we have to check for the presence of endotoxins, like LPS, and other pyrogenic contaminants. LPS can still be present in the final product even if we already filtered the debris of the solution, so we have to check again at the end of the process because no molecule of LPS should be present in the final formulation, since even low dosages of this endotoxin (0.5 nanog/kg of body weight) could still induce adverse medical reactions. LPS is made of a polysaccharidic core and it is bound to a lipidic molecule, the lipid A, which is highly toxic for humans because it is pyrogenic. LPS molecules are present on the surface of Gram- bacteria and they constitute the 75% of their outer membrane surface. It is heat stable so autoclaving the solution will not destroy the endotoxin. To check for the presence of this endotoxin we need animals: the first test that we could use is the pyrogen detection in rabbits. We simply inject the final product in a rabbit and then we check for the increase in body temperature. If it increases of more than 2.65°, the formulation contains some sort of pyrogenic molecule that must be removed; if the temperature increases of 1.15° or below, then the product is safe (no pyrogen inside); if the temperature is in between the two values, then the test must be repeated. This test is very easy and fast, however it is expensive, is involves the handling of rabbits (their excitation or poor handling may affect the results of the test), false positive results may be due to subclinical infections or poor animal health and different colonies can yield variable results. So we can use another test, performed on another animal, which is the Limulus (also called Horseshoe crab), which is called LAL, Limulus amoebocyte lysate. This test is used to detect the presence of LPS, not of any other pyrogenic or endotoxic molecule. This is because this animal presents a very simple blood, made of only one type of cell called amoebocyte, and once the blood is in contact with air it becomes blue, because it contains cupper in a pigment called hemocyanin, which transports oxygen in this animals just like haemoglobin in men. This animals present a very simple immune system, whose action is carried out by these single cells that we find in blood, and these cells, once in contact with LPS, are able to recognize it and eliminate it forming a coagulation of this endotoxin. So the test is performed by collecting the blood of the animal without killing it, then we obtain a lysate of amoebocytes from it in order to use its enzymes with our final product and check for the presence of LPS (if LPS is present, a coagulation cascade will start and so we will observe the formation of aggregates in the final formulation). The advantages are that this test is highly sensitive, it has low costs and it can be conducted in a short period of time (15 to 60 minutes). However, the disadvantage is that is high highly selective only for LPS, and, even if LPS is by far the most likely to be present in a pharmaceutical product, if we have to check also for the presence of other endotoxins we must use the rabbit test. Which other types of contaminants can we find in the final vail? DNA, for example, could still be present in the final product. No more than 10 picograms of DNA per dose can be tolerated by humans, so we must check for the concentration of this molecule before commercializing the drug. The technique that we use is the DNA hybridization (dot blot assay): we isolate the contaminating DNA from the product with a simple phenol/chloroform procedure, we spot this recovered DNA on a nitrocellulose filter, we extract the total DNA of the host cell that we used to produce the protein and we label it, then we hybridize it to our spotted DNA (the one recovered from the final product). The labelled DNA will anneal with any complementary DNA strand onto the nitrocellulose filter, so if we visualize a signal that means that the DNA of the cell line is still present as a contaminant in the final formulation. Then we have to make sure that we don’t have any viral contaminants: viruses can be removed with filtration with 0.1 micrometres filters. Usually this filtration is performed before lyophilization because we know that we have to remove any type of microbe inside our solution (they can cause severe infections, they may be capable of metabolizing the product because they could be able to produce extracellular proteases and microbial-derived substances secreted in the product may be toxic for the patient), however, to make sure that we don’t have any microorganism left (especially viruses, that are smaller) we repeat this process also in the last steps. Prior to that, to check for the presence of viruses, we perform viral assays: they are specific for a kind of virus or for a family of viruses and they could be immunoassays (which are sensitive and inexpensive), virus specific DNA probes (similar to dot-blot assays) or bioassays (in this case we could incubate the final product with the cell lines sensitive to a range of viruses or we could administer the final products to a test animal, and if the animal shows any symptoms then its serum is screened for a range of antibodies recognizing a range of viruses). Lastly, some contaminants may be unintentionally added to the product stream. Metal ions can leach from the product-holding tanks pipework or from chromatographic media, so finished products are often subjected to “abnormal toxicity” or “general safety” tests. These entail parenteral administration of the product to at least five healthy mice, then the animals are placed under observation for 48h and should exhibit no ill effects. The death or illness of one or more animals signals a requirement for further investigations. All of these procedures are a part of the so-called validation studies: validation can be defined as “the act of proving that any procedure, process, equipment, material, activity or system leads to the expected results”. New and older items must be routinely validated, and older ones with increased frequency. Validation of biopharmaceutical aseptic filling procedures is amongst the most critical. Contaminant-clearance validation studies entail spiking the raw material with a known level of the chosen contaminant and subjecting the contaminated material to the complete downstream processing protocol. This allows determination of the level of clearance of the contaminant after each purification step. This is mandatory for biotech drugs, for chemical drugs this complex validation step is not required because they are produced with simple chemical reactions, so the validation is far easier and quicker. In this case, all validation procedures must be carefully designed and fully documented in written format. Once we are sure of what we are dealing with as our final product we wait for the FDA approval; if the FDA approves it the product is licenced, marketed and sold. This in the picture is a recap of all the steps. Remember: different process run at different times equals to different product, even if it is performed by the same pharma company. This doesn’t mean that the protein will be different, but that the total product will be slightly diverse (because maybe the DNA vector that we use for the cloning is a little different, or the host cell has different expression levels, or the fermentation process occurs differently, or the downstream protocol is run in a different way, or the final formulation is different). This is something typical of biotech drugs, chemical drugs are always the same even if the processes are different and run at different times. The paradigm of biopharmaceutical is that the process is the product. THERAPEUTIC HORMONES 1. Insulin: insulin has a lot of functions in the body, the most important is the reduction of glucose in the blood stream. It is a very small protein (6 kDa in weight), it is composed by two chains, which are the A chain (21 amino acids in length) and the B chain (30 amino acids in length), that are kept together by disulphide bonds, which are both intrachain and interchain. The structure and the length of the protein are really important to know because through the different composition we can obtain different types of insulin. Physiologically, it is produced as a pre-proinsulin, which is a single polypeptide molecule which contains the two chains (A and B) bound together by a linking peptide called C chain. Then the disulphide bonds are produced and the C chain is digested, so that the protein is only made by the A and B chains. This physiological process can be reproduced in vitro to produce recombinant insulin. Moreover, we know that native insulin molecules are kept together by zinc atoms, forming multimers, in particular hexamers, which determine how quickly the insulin is going to be absorbed into the tissues, and this is really important for a pharmacological application. In the past, insulin was obtained by pigs, and about 70 pigs per patient per year were needed for the production of this protein for diabetic patients. Porcine insulin differs of only 1 amino acid from the human one, however it was immunogenic (the bovine insulin differs of 3 amino acids so it’s even more immunogenic than the porcine one). So in order to facilitate the treatment of diabetic patients, the recombinant insulin was first produced and marketed in 1982, becoming the first marketed biotechnological drug. The main advantages is that we can produce huge amounts of the protein in this way, with a low grade of immunogenicity, however the disadvantage is that E. Coli is not able modify the eukaryotic pre-mRNAs nor it can operate post translational modifications of the protein. Anyway, with biotechnological approaches and with recombinant DNA techniques, it is possible to produce insulins with altered pharmacokinetic properties (faster- or slower-acting) and to select insulin with higher receptor affinities (ensuring economic benefits from using smaller quantities of insulin per therapeutic dose). How do we produce insulin? The first method is the one of insulin crb: we have two cell lines of production (of E. coli), in one we clone the gene of the A chain, in the other one we clone the gene of the B chain. We make these cells express the two chains, we purify them through chromatographic steps and then we incubate both chains under appropriate oxidizing conditions to promote the formation of interchain disulphide bonds. In the second method, on the other hand, we produce the whole molecule in a single cells line (insulin prb): we insert the gene encoding for human pro-insulin in recombinant E. coli, we purify it from the molecules of the bacteria and then we remove the C peptide in vitro with a proteolytic excision. This second method is really easy and we use HPLC chromatography to purify it in the “clean-up process” (this chromatography method yields a final production of approximately 99% purity in 1 hour, keep in mind that the loaded product is 92% pure), achieving low levels of impurity to avoid any significant immunological response. The engineering of a protein allows us to modify the protein if we need it: we can change the amino acidic chain to obtain different types of insulin with different properties. The different types of insulin can be divided based on how much time they spend in the circulation: if we create a graph comparing the plasma levels of the different types of insulin we can see that there are some types of insulin that remain in blood for a longer time (multiple hours, almost a day) than others. In fact, we can have rapid-acting insulins (Humalog, novolog and apidra; they are rapid acting and they are rapidly eliminated from blood), short acting insulins (regular one), intermediate (NPH) and long-acting ones (glargine, detemir; long onset of the therapeutic effects but they last for a longer time). Regular insulin was the first one produced: it is identical to the physiological one, however its curve of duration in the blood circulation doesn’t identically overlap with the physiological insulin (it lasts for a longer time), so this means that there is a delay in the onset of insulin in the blood, in fact after injection we have the so called “lag phase”, which lasts for about half an hour/ 1 hour. After the lag phase, the insulin reaches its peak after 2-3 hours and it is cleared within 8 to 1 hours (so it doesn’t last for the whole day). This means that patients need to inject this type of insulin at least 30 minutes/60 minutes before a meal. Physiologically, zinc atoms keep together single molecules of insulin, which are called monomers, and usually physiologic insulin is present in groups of six monomers, so it forms hexamers. For regular insulin it is the same (because it is associated to low zinc concentrations, around 0.5-0.7%), that’s why it takes a while until it is absorbed, because for insulin to be absorbed it has to be in a monomeric form. Regular insulin exists in preparations as a mixture of monomer, dimer, tetramer and zinc-insulin hexamer and, since absorption must occur under monomeric form, the other forms slow the rate of absorption. The second insulin that was produced was NPH (neutral protamine Hagedorn): it is an intermediate acting insulin that takes about 2 to 3 hours to reach the maximal concentration in the blood and then its action lasts up to 12 hours. It contains protamine, a protein contained in fish, that is able to delay the absorption of this type of insulin so that one injection can last for a longer time compared to regular insulin. Why do we need these different types of insulin? Because we have to mimic the physiologic levels of insulin, so it’s really important to maintain also a basal level of insulin, instead of having to inject regular insulin multiple times in a day. Sometimes, the risk for a diabetic patient is the nocturnal hypoglycaemia, in fact some insulins cannot be administered before the night because if the insulin is too high in blood during sleep time the patient could suffer from hypoglycaemia. NPH is one of these: the peak of action is after 5 to 7 hours after administration, so we have the risk of nocturnal hypoglycaemia if we administer it at bedtime. Moreover, even if it lasts longer, the duration of its action is not long enough to cover insulin requirements for the whole day with a single injection. On top of that, the action profile depends on the dose and it is highly unpredictable also because we have a variability of absorption that depends on site, on exercise and on the variation in mixing of suspension. For this reason, some devices have been invented to inject amounts of insulin repeatedly and directly in the body of the patient, and they are called insulin pumps. For these reasons we can understand why we need insulin analogues even if we have insulins with improved characteristics: insulin therapy is supposed to mimic physiological insulin secretion (both basal and bolus), so preparations providing bolus (meal-related) insulin requirements should have a rapid onset, rapid peak and short duration, while those providing basal requirements should have low basal concentration, without a peak, for a long duration. Regular insulins don’t mimic physiological insulin secretion, they often result in mismatches between requirement and availability and this results in an inadequate glycaemic control and late hypoglycaemia. So different mixtures of short and long- acting insulins are now marketed as products. Insulin analogues are called analogues because they are slightly different from the normal insulin, some amino acids have been changed in order to alter its pharmacokinetic profile but not its pharmacodynamic one (in the previous cases the sequence of the insulin was the same but we used additional proteins to prolong the action). The analogues can be divided into short acting ones and long actin ones: in the short acting ones we have Lispro, Aspart and Glulisine. Their onset is after 5 to 10 minutes, their peak is reached after 30 to 90 minutes and the elimination is usually in 4-6 hours (curve of duration in the blood stream really similar to the normal physiological insulin). This means that these analogues can be injected 5 minutes before a meal, and this makes the life of the diabetic patient much easier. Lispro is an analogue where the second last and the third last amino acids on the B chain have been switched (they are a proline and a lysine, that’s why it is called Lispro) and this is enough to induce a very rapid absorption, because this alteration decreases the propensity of the individual insulin molecules to self-associate, so preventing aggregation and reducing the formation of hexamers. Lispro is produced as pro-insulin in E. coli. In Aspart, we have the substitution of proline with an asparagine in the B chain (position B28). This type of analogue is produced in a modified Saccharomyces cerevisiae strain, so in yeast. Again, the modification induces a lower tendency to aggregate in hexamers and so the peak time is reached I less time compared to regular insulin. The Glulisine has a substitution with a lysine instead of an asparagine and glutamic acid instead of a lysine. In this way we introduce two charges, one positive and on negative, and in this way we avoid the association in hexamers (slight decrease in the isoelectric point compared to normal human insulin). This happens again in B chains because B chain is really important for the aggregation of the monomers. So with these small changes we change the pharmacokinetics of the protein, ensuring a quick absorption from the subcutaneous site to the bloodstream. There have been other attempts to produce short-acting insulin analogues, however most of them were not approved because it induced tumors in the mammary glands in rats. Instead, the long-acting insulin analogues are used to guarantee a basal level of insulin for a very long time. The first one produced was the Glargine, which presents a substitution in position 29 on the A chain of an asparagine with a glycine and in chain B it has the addition of two arginine residues which raise the isoelectric point of the protein (because the arginine is positively charged), lowering its solubility. It is produced as pro-insulin in E. coli, the purified product is formulated at pH4 (it is soluble at pH 4, so it will be in the form of monomers), but upon sub-cutaneous injection the pH increases (up to 7) so that the protein becomes un-soluble and it aggregates and precipitates in the tissue. In this way we have a prolonged duration of release (the longer the absorption time, the longer will last the peak of concentration in the blood). After subcutaneous injection, plasma insulin levels rise slowly to a peak after 8 to 10 hours, with a duration of action of about 20 up to 30 hours (whole day long duration, so Glargine can be administered also once a day). Another long-acting insulin is the insulin detemir (Levemir): in this case we are in the B chain, B30 threonine is removed, and it is replaced by a chain of fatty acids, which confers the property to the protein to bind to the serum albumin (the pH of the protein remains neutral though). In this way, the release is very slow, so in this case the long duration depends on the release kinetics and not on the absorption. It is produced in Saccharomyces cerevisiae, and it has a shorter duration compared to Glargine, however it is still used a lot (onset in 0.8-2 hours and mean duration of 16-18 hours). The last one is the degludec: it has a fatty acid chain attached again to the lysin at B29 (so again we have the removal of the B30 residue). Thanks to the presence of zinc atoms, it assembles in multi- hexamers and so it takes a lot for the monomers to be absorbed. The degludec’s day to day glucose lowering effect is usually constant during the day, while the glargine one usually varies at different hours, and degludec’s half-life is actually longer (25.4 hours vs 12.1 hours), but they are both really used. Even if these analogues have some particular modifications in the amino acidic chains, the binding affinity to the insulin receptor is always the same, otherwise they could not be used as drugs. New attempts of producing different types of insulins that could be administered through other types of routes, other than the parenteral one, have been made, with the intent of making the life of diabetic patients easier. In fact, insulin must be administered to these patients multiple times during the day, and the sub-cutaneous route makes this process harder because of how impractical it is if used in recurrent deliveries. So, researchers have tried to come up with new formulations for an easier distribution of the drug, that could be compatible with normal life. One example is Exubera, an inhaled insulin, that was however withdrawn from the market in USA in 2007, after one year of marketing. This retirement was due to the fact that the device used for this type of formulation was not comfortable and it was expensive, and some cases of lung cancer have been recorded in ex-smokers after the use of this type of insulin. Moreover, this type of formulation needed patient’s training and pulmonary function tests, which are highly costly. However, the idea of an inhalable insulin was not abandoned, also because we have an history of protein delivery to the lungs routinely without adverse events, of which the main example is Pumozyme (Dornase alpha or DNase, therapy for cystic fibrosis). In 2015 Afrezza, an inhaled insulin for type I diabetes, has been approved in the USA (it is an insulin which must be used in combination with a long-acting insulin; it is not recommended for the treatment of diabetic ketoacidosis, and it cannot be used by patients who smoke), and now many other types of insulins that can be administered by the pulmonary route are in the pipeline. Another type of insulin that was developed is Oralin, but in this case the administration is by oral route (actually, it is buccal, because it is absorbed by the oral mucosa, in the mouth and throat). It has been developed by Generex and it has been approved in Ecuador and India. This type of insulin is formulated with a variety of additives and stabilizers to prevent denaturation on aerosolization and to stabilize aerosol particles (however, the dosage is still different compared to the sub-cutaneous type because we actually don’t know how much of the insulin is absorbed by the mucosa). It hasn’t been approved by the FDA yet, but researchers are trying to formulate it also for diabetes type II. 2. Glucagon: glucagon is synthesized by the alpha-cells of the islet of Langerhans of the pancreas. The major biological actions of this hormone tend to oppose those of insulin, in fact it is only used as a therapy to the major and most frequent complication of insulin administration in diabetic patients, which is hypoglycemia. Glucagon preparations utilized therapeutically were chromatographically purified from bovine or porcine pancreatic tissue (their structures are identical to human), but recently, recombinant glucagon has been produced by Novo Nordisk in Saccharomyces cerevisiae and by Eli Lilly in E. coli. 3. Growth hormone: it is used to treat dwarfism. It was the second recombinant protein produced and marketed after insulin (1985). Before the production of its recombinant counterpart, growth hormone was purified from the pituitary glands of corpses because this protein, in order to be used as a therapy, must be human. However, extracting it from corpses had a lot of disadvantages: first of all we didn’t have a lot of corpses to purify it from, so the amount of hormone that we could obtain was vey low, moreover we extracted the protein from neuronal tissue, and this could cause the insurgence of infections in the patients, such as the Creutzfeldt-Jakob disease (mad cow disease). So the production of a recombinant growth hormone in E. coli was a really great innovation, because the native and recombinant forms of the molecule are identical. The growth hormone is used for the treatment of dwarfism but also of short stature in general, for the induction of lactation, for the counteraction of ageing, the treatment of obesity, body building and induction of ovulation. Other than the recombinant growth hormone, also an analogue that acts as an antagonist has been produced: it is called Somavert and it is used for the treatment of selected patients suffering from acromegaly. This molecule, commercially called Pegvisomant (Peg stands for pegylated), is able to bind to the receptor for the growth hormone but it does not induce receptor dimerization and thus the JAK/STAT pathway, preventing the expression of the genes regulated by the activation of this receptor (like IGF-1). The main adverse effect of this drug is that it may raise endogenous GH circulating levels because we don’t have any negative feedback regulating its synthesis in the pituitary gland or hypothalamus. The main difference between this analogue and the growth hormone molecule is that it has a substitution of 9 amino acids and it is pegylated, so that we lower the immunogenicity and we raise the half-life of the molecule. The aminoacidic substitutions include the G120K mutation, which is able to increase the affinity of the molecule with the GH receptor, but at the same time it prevents its activation. 4. Gonadotrophins (or gonadotropins): they are really important also in the veterinary field. They are a group of hormones constituted by the gonadotropin releasing hormone, released by the hypothalamus, the follicle stimulating hormone (FSH) and luteinizing hormone (LH), produced by the pituitary gland, they have the gonads as target organs (ovaries and testis), and human chorionic gonadotrophins, released in the placenta. They bind to different receptors but these receptors all share the same subunit. They must be glycosylated, otherwise they cannot function, so they cannot be produced by E. coli. Human gonadotropin is excreted in the urine during pregnancy (used for pregnancy tests), so before being reconstituted in vitro (recombinant gonadotropin), it was extracted from the urine of pregnant women through ion exchange chromatography and gel filtration. This process had a lot of disadvantages, because dealing with urine is very complicated (urine is toxic) and also the quantity of hormone that was extracted was really low. However, even nowadays it’s still available as an extracted protein because the process is cheaper. Menotrophin (human menopausal gonadotrophin) is the name given to FSH-enriched extracts from human urine (2,5 l/dose). Gonadotropins could also be extracted from pituitary glands (especially FSH and LH which are produced from this gland). However, in 1985 the extraction of gonadotrophins from the human pituitary was abandoned because of the risk of the Creutzfeldt-Jacob disease. Now the human chorionic gonadotropin is produced as a recombinant protein, it has the name of Ovitrelle and it was approved in 2001; it is used for the treatment of female infertility due to anovulation and for patients undergoing assistive reproductive technology, in fact it is used to trigger final follicle maturation and luteinization after follicle stimulation. It is produced in CHO cell lines because it has to be glycosylated, it is identical to the human protein and it is purified by multistep chromatography. The number of oocytes retrieved with recombinant hCG was similar to the number retrieved when urinary-derived hCG was used. Sometimes it is used also for the treatment of male infertility in assisted reproductive techniques. For veterinary use, the use of gonadotropins is pretty similar: we use it to obtain an identical offspring from surrogate dams (animals that receive multiple embryos coming from the same donor). So a cow is inseminated from the sperm of a superior bull (selected species of bull, with some particular characteristics), then we recover the embryos produced, we split them and we reimplant them in other cows, which are surrogate mothers, for differentiation, growth and birth. This technique is used to amplify the offspring of a genetically selected sperm donor and it is called embryo transfer; the aims are to increase the productivity of the superior donors, maximize the use of their valuable semen, transport genetics across long distances and produce identical offspring by embryo splitting (potentially valuable as research animals). If we select females in the offspring we can use them for the production of milk, for the high growth rate (if we use a superior donor with an enhanced growth characteristic) or for the outstanding reproductive capacity. The limitations of this embryo transfer technique is that it is expensive, it takes time and you need to have experienced and trained people to perform it. For the success of this technique we have to induce superovulation in the surrogate mothers, and for this purpose we use recombinant gonadotrophins. Both the donor and the recipients surrogate mothers are synchronized with a treatment with recombinant prostaglandins (to obtain a synchronized the oestrus cycle in all the animals and prepare their uterus to support embryogenesis). The donor is artificially inseminated, then the embryos are recovered (after 68 days a catheter is inserted into the uterus and inflated to prevent retrograde flow of flushing medium) and transferred to the recipient animals (using a transfer pipette). Each of the recipient animal will become pregnant and they will give birth to the offspring that will be identical to the donor. The superovulation is induced in the donor through FSH (it has a short half-life, of about 2 hours; sometimes also another hormone, called PMSG, which stands for pregnant mare serum gonadotrophin, could be used in Europe, which has a long half-life, of about 2 to 4 days); the purpose of the superovulation is to hyper-stimulate ovaries with gonadotropins and provide a higher number of follicles than normal. A typical treatment response would be of 8 to 10 ovulations. 5. Three additional recombinant hormones: they are thyroid stimulating hormone (TSH), parathyroid hormone (PTH) and calcitonin. Thyroid stimulating hormone (TSH) is structurally similar to gonadotrophins, although it functionally targets the thyroid gland as opposed to the gonads. It is produced in CHO cells and it has been approved for medical use as a diagnostic aid in the detection of thyroid cancer/thyroid remnants in post-thyroidectomy patients, so we don’t use it for therapy purposes. Parathyroid hormone (PTH) is a primary regulator of calcium and phosphate metabolism in bones; a truncated version of PTH (of 4 kDa) is produced, it is identical in sequence to the N-terminal residues (1-34 AA) of endogenous hPTH and triggers the same effects. It is produced in E. coli and it is used for the treatment of osteoporosis in postmenopausal women. Finally, calcitonin is a hormone that lowers serum Ca2+ and inorganic phosphate (Pi), it is used clinically to treat hypercalcemia (condition that can occur in come tumors of the bone), associated with some forms of malignancy and Paget’s disease. Salmon calcitonin is 100-fold more potent than the human native hormone (it differs from the human hormone by 9 amino acids), so the recombinant protein that we produce is not human but it is derived from fishes (it could be immunogenic but the treatment with calcitonin is very short so it’s not so problematic). It is reduced in E. coli and the amination is carried out in vitro. CYTOKINES 1. The main cytokine used in therapy is interferon: it is a key molecule in the first line defence against viral infections, it belong to the innate immunity and it is released by the cells upon a specific stimulus. We can have two types of interferon: type I (mainly represented by interferon alpha, produced by virus- infected leukocytes, and beta, produced by virus-infected fibroblasts or epithelial cells) and type II (represented by gamma interferon, produced by T cells and activated NK cells). When a virus infects a cell, interferon is produced: it binds to its receptor on NK or T cells and they will activate an immune reaction. In particular, interferon alpha and beta induce the synthesis of inhibitory proteins, while interferon gamma increase the expression of MHC II molecules on APCs and increase the ability of macrophages to resist to viral infections and to kill other infected cells. All types of interferon increase the expression of MHC II molecules and increase the NK cell activity. Interferon is used in therapy against infection (like from hepatitis virus), cancers or multiple sclerosis. Up until the 1970s, interferon was sourced directly from human leukocytes obtained from transfused blood supplies (species- specific), but only 1% pure (they were extracted as a mixture of IFNs). Moreover, the extraction from blood was really dangerous because of the possible transfer of infection from infected blood to the patient. In the late 1970s, the production of interferon became possible in significant quantities: various cell lines were found to secrete substantial amounts of IFNs (for example Namalwa cell line was able to produce 8 subtypes of IFN-). However, it was still produced in mixtures, so we still weren’t able to obtain single proteins. DNA recombinant technology led to the production of a single IFN subtype in the 1980s. Most interferons have been produced in E. coli, yeasts, and some mammalian cell lines. Most IFNs currently in medical use are produced in E. coli because the majority of human IFNs are not normally glycosylated (mostly IFN-a and IFN-b). IFN-g is glycosylated but the non-glycosylated form displays a biological activity similar to the native protein, so the glycosylation is not mandatory for this molecule to be use in therapeutical applications. However, the production of IFN in recombinant microbial systems means that any final product contaminants will be microbial in nature. So a high degree of purification is required: up to 99% of purity can be achieved with extensive chromatographic purification. The first recombinant INF was produced in 1980 (INF-alpha2a), in 1986 it was approved for the treatment of hairy cell leukemia (Intron A), but it can be used also against viral infections like in hepatitis. PEGylated interferons are now available: the PEG molecule is able to prevent the degradation of the molecule, shielding it from enzymatic degradation, so we have a prolonged half-life (13 hours to 25 hours compared to the 4 hours of the non-pegylated protein) with the same intrinsic biological activity as non-PEGylated INFs. Moreover, the PEGylation lowers the immunogenicity of the protein, allowing multiple and frequent dosing and achieving a higher and sustained interferon concentration in blood. If we administer the interferon with an antiviral chemical drug, we can see that the combination of this molecule with the non-PEGylated molecule reaches the 44% of remission of the HCV infection, while the combination with the PEGylated molecule allows us to reach a remission of 56%, so we obtain better results (both these combination work even better than immunotherapy). We can see that another main advantage is that, with better results, the dosage can be decreased, and this is very important for the adverse reaction profile (the less drug we use the less side effects we could have). We could have two different types of PEGylated interferon: PEG interferon alfa-2b (linear molecule, with a weight of 12 kDa) and PEG interferon alpha-2a (larger and branched molecule, with a weight of 40 kDa). This changes their pharmacokinetics: the volume of distribution of the larger PEGylated interferon is lower than the smaller one (8L vs 20L), in fact, being a larger molecule, it is harder for this interferon to pass through tissues, in fact also the absorption half-life is longer, just like the mean elimination half-life. If we change and we modify the molecule by adding a different type of PEG molecules we can obtain different molecules with different pharmacokinetics characteristics (for example, if we want to have a longer half-life, we will use a bigger PEG molecule, but we have to remember that in this way also the time of absorption will be prolonged, while the volume of distribution will be lower, so some characteristics will have to be “sacrificed”). When talking about recombinant interferons we have to mention Infergen (interferon alphacon-1): it is not a real natural interferon molecule, it is “invented”. This protein is in fact an engineered INF approved for the treatment of hepatitis C, produced in E. coli, with a higher antiviral, anti-proliferative and cytokine-inducing activity than native type I interferons: the aminoacidic sequence presents the most frequently occurring aminoacidic residues in native IFNs-a. So it is not the copy of any other existing interferon, but it still works in a very good way. For what concerns INF beta, it is used for the treatment of multiple sclerosis, it inhibits the production of pro-inflammatory cytokines (like TNFalpha and INF-gamma) and it can be produced in either E. coli or CHO cell lines, because even if it is not glycosylated it yields the same therapeutic efficacy as the glycosylated protein. The recombinant INF-gamma is a particular interferon because it is used against the CGD (Chronic Granulomatous Disease), a rare genetic disorder where there is a defect in the granulocyte response to infections, so granulocytes do not activate properly. This pathology is responsible for the development of a lot of infections and granulomas in the patients and sometimes they could be lethal. It is produced in E. coli, hence it lacks the carbohydrate component, but it still displays the identical biological activity as the glycosylated protein. It is a potent activator or phagocytes, in particular of granulocytes, that’s why it is used against CGD. So interferons are really important for their therapeutic effect against infections, tumors and other types of disorders, however they could be toxic: INF elicit a number of unwanted effects, such as flu- like symptoms (we naturally produce interferons during an infection, that’s why we have this type of toxicity), which is typical of all types of interferon, but then we could have also more serious adverse effects, like anorexia, insomnia, cardiovascular complications, autoimmune reactions and hepatic decompression (typical of interferon alpha use, but they are rare); hypersensitivity reactions, menstrual disorders, anxiety and depression (typical of the use of interferon beta); heart failure, CNS complications and metabolic alterations (typical of interferon gamma). Generally speaking, we consider interferons safe molecules despite these adverse reactions, because the more serious side effects are usually rare and the flu-like symptoms are not so dangerous, since they usually appear within a few hours (8 for interferon alpha) from administration, but then tolerance is developed within the first week of treatment. We could also have additional interferons, and they are: - Interferon tau: trophoblastin, has similar effects to type I INF, with significantly lower toxicity. It is found in sheep and cattle and it is used for the synchronization of oestrus cycle in the veterinary field. - Interferon omega: it is found in human and pigs but not in rodents, so it cannot be studied in these animals. It is a type I interferon; a recombinant form of feline INF-omega is used for the reduction of mortality and clinical symptoms of parvoviral infections in young dogs. 2. Another type of cytokines that are used for therapeutic purposes are interleukins. They are a group of protein constituted by 33 members, produced by a number of different cells. Most interleukins exhibit paracrine activity, and some displays autocrine activity, through which they regulate a variety of physiological and pathological conditions. Thus far, only three interleukin-based products have gained approval for general medical use, IL-2, IL-11 and IL-1. IL-2 is a single chained polypeptide containing 133 amino acids. It is a glycoprotein, but glycosylation is not required for its biological activity. It is an autocrine growth factor for T cells and it has proven to have beneficial effects on the treatment of some cancer types (through the activation of T cells, whose activation is normally suppressed when the cancer develops). Recombinant IL-2 is produced in E. coli and differs from native human IL-2, since that it is non-glycosylated, lacks an N-terminal alanine and has a replacement of one amino acid. Despite these modification, this product displays the same biological activity as native IL-2. IL-2 is able to activate LAKs (lymphocyte activated killer cells), which are NK cells activated by the binding of this cytokine with its receptor on the surface of these cells. Clinical studies have shown that if we collect NK cells from the patient (ex vivo), we activate them in vitro with the use of IL-2, generating activated LAK cells, these were able to induce a complete tumor regression in 10% of patients suffering from melanoma or renal cancer. This percentage is too low, so in another experimental setting we extract a tumoral biopsy, we isolate TILs from it (tumor infiltrating lymphocytes), we activate these TILs with IL-2 in vitro (these cells are already directed against tumoral antigens) and then we reintroduce them in the patient. In this case 50% of patients presented the regression of the tumor, so the technique worked better. However, this result is still not so encouraging. So they thought about administering IL-2 directly to the patient: ovarian, bladder cancer, non-Hodgkin’s lymphoma and acute myeloid leukemia are at least partially responsive to IL-2 treatment. However, several trials yielded conflicting results, because IL-2 could cause the so-called cytokine storm, causing very serious side effects. For this reason it cannot be administered directly to the patient. This application of IL-2 was extended also to the treatment of infectious diseases, because the IL-2 capacity of stimulating T cell responses against intracellular pathogens can be useful for the treatment of a wide range of infections (like mycobacteria, Listeria and Legionella). However, this applications yielded the same problems are the one against tumors, which are mainly related to the side effects of IL-2. These unwanted reactions are usually related, as we said, to cytokine storms, because usually IL-2 stimulated the production and release also of other cytokines (IL-3, IL-4, IL-5, IL-6, TNF and INFgamma) by the cells that it activates; they are dose-limiting and they may cause cardiovascular, hepatic and pulmonary complications, which usually induce immediate termination of the treatment. However, IL-2 has been used to produce a fusion protein with an antibody (short chain variable fragments, so only the portion of the antibody that recognizes the antigen). Some examples are the L19-IL2 (darleukin) and the F16-IL2 (teleukin): in this way, IL-2 is bound to the variable portion of an antibody that is able to recognize a tumoral antigen in a specific way, so that the delivery of IL-2 will be specific (in this cases, these Abs are vascular targeting Abs) and it will induce an immune modulation on immune effector cells and stromal cells only in the site of the tumor. This expedient is much better than administering simple IL-2 without a specific target, which can cause really dangerous side effects, as we saw. Darleukin has already been tested in clinical studies in melanoma and pancreatic cancer. More recently, combination of Darleukin with other immune-cytokines or with chemotherapeutic drugs has shown potent synergistic activity in pre-clinical studies, and therefore the current clinical trial strategy is based on this concept. On the other hand, Teleukin has previously been tested in clinical studies in various solid cancers. More recently, Teleukin has shown the potential to enhance the therapeutic performance of other agents, including taxanes. We saw that this is the only way through which we can use IL-2 has a therapeutic agent without it causing multiple serious side effects. But if we change our point of view we could think of using IL-2 as a target, not as a therapy. In fact, when too much IL-2 causes damage, prevention of IL-2 is necessary. A high amount of IL-2 can be present in autoimmune diseases (rheumatoid arthritis, inflammatory bowel disease, lupus erythematosus,...), because we have an over-activation of the immune system so also IL- 2 will be produced in higher quantities, but we could also have a high level of IL-2 in organ transplantation, and we can block its production to avoid host vs graft complications. IL-2 can be blocked by administering soluble forms of the IL-2 receptor (IL-2R), we can administer a mAb towards the IL-2 receptor, or against IL-2 itself, or we could administer IL-2 variants that fail to initiate the signal transduction, or we could deliver IL-2 coupled to bacterial toxins. The last method that we mentioned can be explained through the example of Ontak (denileukin diftitox): it is a really effective drug against IL-2, it is a fusion protein produced in E. coli formed by A and T fragments of the diphtheria toxin and by IL-2. In this way, the toxin is brought in close proximity to the cell and it will bind to the cell-surface IL-2 receptor through the IL-2 portion; in this way it will be internalized through endocytosis and it will cause the death of the cell. In fact, the acidic pH in the endocytic compartment causes a conformational change that enables the translocation of the A chain of DT to the cytosol. Once in this compartment, DT modifies elongation factor 2 (EF2) by adenosine diphosphate (ADP)-ribosylation, which leads to inhibition of protein synthesis and cell death. Another important cytokine is IL-1, a family of proteins characterized by an array of 12 beta strands and by the absence of a classical secretory N-terminus peptide sequence. The two main components of this family are the IL-1alpha and the IL-1beta: the first one is generally associated with the plasma membrane of the producing cell and so it acts locally, it has a widespread production (in endothelial cells and keratinocytes), it is important in priming T cells during contact hypersensitivity and for the induction of high levels of serum IgE and its pro-domain has a nuclear localization sequence (nuclear IL- 1a has transcriptional transactivating activity); the second one is generally secreted and it circulates systemically, it is mainly produced by monocytes and macrophages and since it can circulate to the brain, it is important for the induction of fever. They are both non-glycosylated, so they could be potentially really easy to produce directly in E. coli, however therapeutically they cannot be used because they give the same problems of IL-2, so they are mainly used as targets. The two proteins signal through the same receptor complex and have identical biological activities in solution. Because of their potency and extensive functions, their activity is tightly regulated (through mRNA induction, regulation of processing and secretion, expression of a receptor antagonist and of a decoy receptor). IL- 1 proteins are usually produced as longer peptides and then they undergo maturation through digestion with caspases (caspase 1, also called ICE, IL-1-converting enzyme). Stimulation of cells for the production of IL-1 leads to the production of inactive ICE which, when activated, cleaves inactive pro-IL- 1, which is released across the cell membrane in an activated mature form. In the case of IL-1alpha, calpain cleavage of active pro-IL-1 stimulates release of mature IL-1 across the cell membrane. Some clinical studies demonstrated that no significant anti-tumoral response was observed in the treatment of various cancer with IL-1, so they are actually more harmful than useful. So instead of studying them as biotechnological drugs they have been identified as targets: there are a lot of diseases where we have a major production of IL-1 and so by reducing its amount we can ameliorate the clinical severity of these disease, which are typically acute/chronic inflammatory diseases. Which are the functions of IL-1? It promotes the synthesis of eicosanoids and other inflammatory mediators, it activates B lymphocytes, and thus T cells, along with IL-6 it induces synthesis of acute- phase proteins in the liver, it induces the production of adhesion molecules in the endothelium and in fibroblasts, it acts as a co-stimulator of haematopoietic cell growth and differentiation and it stimulates the production of collagen from chondrocytes. Moreover, in the brain, it induces the raise of the body temperature. All of these activities correlate to an inflammatory state. To block these functions e could use the IL-1 receptor antagonist (IL-IRA), a protein that is able to bind to the same receptor as IL-1 but it doesn’t elicit any biological activity. A lot of studies have demonstrated that IL-1 is involved in a lot of pathologies, so we can consider it a sort of pathogenic molecule, because a lot of diseases with an inflammatory basis produce inflammatory mediators, of IL-1 is one of the main effectors. Some examples of pathologies in which IL- 1 is massively produced and so in which we can target it are sepsis syndrome, rheumatoid arthritis, inflammatory bowel disease, insulin-dependent diabetes mellitus, acute and chronic myelogenous leukemia and atherosclerosis. Other diseases include transplant rejection, graft vs host disease, psoriasis, asthma, osteoporosis, periodontal disease, autoimmune thyroiditis, alcoholic hepatitis, sleep disorder and premature labour secondary to uterine infection. Also neurological disorders can be included in the diseases in which IL-1 is implied: Alzheimer’s disease, Parkinson disease, epilepsy, stroke, … they all could present an increased expression of IL-1. In general, any pathology which presents an inflammatory state will present the production of IL-1. One inhibitor of IL-1 is called Anakinra: it is a human IL-1 receptor antagonist and is produced by recombinant DNA technology. It is non-glycosylated and is made up of 153 amino acids. With the exception of an additional methionine residue, it is similar to native human IL-1 receptor antagonist. This endogenous IL-1 receptor antagonist is a 17-kDa protein which competes with IL-1 for receptor binding and blocks the activity of IL-1. Anakinra (commercially called Kineret) is recommended for the treatment of severely active rheumatoid arthritis for patients 18 years of age or older. It is recommended for patients who have not responded well previously to the disease-modifying antirheumatic drugs. It reduces inflammation, decreases bone and cartilage damage and attacks active rheumatoid arthritis. The most serious side effects of Kineret are infections and neutropenia; other side effects may include headache, nausea, diarrhea, flu-like symptoms and abdominal pain. An increased risk of malignancies has also been observed. IL-1 in rheumatoid arthritis activates monocytes and macrophages causing the activation of the inflammatory process, it induces the fibroblast proliferation causing synovial pannus formation, it activates chondrocytes causing cartilage breakdown and it activates osteoclasts inducing bone resorption. Other IL-1 antagonist molecules for the treatment of rheumatoid arthritis are under development. IL-11 is the only interleukin which is produced with a therapeutic purpose, as it is useful in the bone marrow for the formation of platelets. It is not involved in inflammatory processes, so it does not cause the serious side effects that the other interleukins induce. It is usually produces by IL-1-actiavted bone marrow stromal cells and fibroblasts and it stimulates thrombopoiesis by inducing growth and differentiation of bone marrow cells. Commercially it is available under the name of Oprelvekin and it is mainly used in patients undergoing chemotherapy to stimulate platelet synthesis. Another drug used against interleukins is the Ustekinumab, used against IL-12/23. These cytokines are involved in the production of specific T cells that are implied in the inflammatory process of psoriasis. They usually bind to their specific receptor, activating T cells and inducing their differentiation into Th17 cells: these types of cells are a subtype of T cells that activate in some autoimmune diseases, like in psoriasis, inflammatory bowel diseases and colitis ulcers. The drug that we use is, as the name suggests, a monoclonal antibody and it is currently used in the treatment of psoriasis under the tradename of Stelara because of its ability to prevent the binding of the natural ligand to its receptor by attaching to the interleukins. Anther mAb that acts against an interleukin is tocilizumab, which isa humanized antibody able to bind to the IL-6 receptor and thus to prevent the binding of this interleukin to it. It is used in rheumatoid arthritis (especially in cases of RA resistant to other therapies), where IL-6 has a pathogenic role: it is involved in the control of the balance between autoimmunity and self-tolerance, favouring autoimmunity, in fact it usually stimulates the production of Th17 cells over Tregs (sentinel cells that prevent the hyperactivation of T cells, especially regulating Th17 cells, so these two populations of cells must always be balanced in number, if that doesn’t happen Tregs are not able to control Th17 cells anymore and so we have the development of autoimmune diseases). The side effects of this mAb therapy are infusion related reactions (flushing, headaches, fever, nausea and fatigue). The last cytokine against which we can act with biotechnological drugs is TNF (tumor necrosis factor), which is involved in a lot of autoimmune pathologies. We have two types of TNF, the alpha and the beta; it has been discovered in vitro as an anti-tumor molecule, that’s why it is called tumor necrosis factor, in fact it was also used for the treatment of some tumors. It is a pro-inflammatory cytokine involved in the activation of elements of both non-specific and specific immunity (macrophages, neutrophils and granulocytes), in the induction and regulation of inflammation (synthesis of IL-1, IL-6, IL-8; activation of neutrophils, expression of adhesion molecules, chemotaxis, synthesis of PGE and proteins of the acute phase), but it also presents a selective cytotoxicity against a range of tumoral cells and it mediates various pathological conditions like septic shock, cachexia and anorexia. TNF is known to mediate the symptoms of many diseases, like cancer, septic shock, rheumatoid arthritis and diabetes. However, the use of TNF as an anti-tumoral agent has failed, it can be used only in vitro because in vivo it causes also important side effect because its anti-tumoral action is not its only function. The only drug that is based on TNF that can be used in vivo is Beromun, a human TNF-alpha produced in E. coli identical to the native human protein that is used against soft tissue sarcoma to prevent necrosis and amputation of limbs. Side effects include nausea, liver toxicity, and locally (limb) oedema and infection. In every other case, TNF is used as a target, and we have come up with two strategies: first, we know that TNF has its own receptor (actually, we have one for TNF-alpha and one for TNF-beta), so we can produce its external portion as a recombinant soluble protein able to bind TNF and prevent its binding to the actual receptor, otherwise we can use mAbs. The mAbs against TNF are some of the oldest mAbs produced as biotechnological drugs. The soluble receptor drug is called Etanercept and it is a recombinant fusion protein made of a soluble version of TNFII bound to the Fc region of an IgG1 (actually it is made of two molecules of receptor because naturally the TNF receptor dimerizes to start the signal so in order to bind TNF it has to be dimerized), so the structure is actually pretty similar to an antibody but it is a recombinant protein. Infliximab (also called Remicade) is instead a mAb, a chimeric one (mouse-human) directed against TNF, it was the first chimeric antibody ever produced and it is still used in therapy; adalimumab (also called Humira) is a fully human mAb anti-TNF, it was the first totally human antibody ever produced; certolizumab pegol (also called Cimzia) is a PEGylated molecule constituted by Fab fragments, kept together by a linker peptide, of a fully human anti-TNF Ab; finally we have golimumab, a fully human mAb anti-TNF. How do we choose which drug to use? It depends on how the patient responds to the therapy, on their half-life (infliximab has an half-life of 7 to 9 days, adalimumab of 10 to 20 days, certolizumab pegol of 2 weeks) and also on how many times they need to be injected per month (for example infliximab can be injected every 2 months while etanercept every month and adalimumab every week: this could be a problem for the patient, the lower is the frequency of injection per month the more compatible the therapy is for the life-style of the patients). Certolizumab pegol is a singular antibody: first of all it is PEGylated, to prolong its half-life (which, without PEGylation, would be much lower because it is a smaller antibody compared, for example, to adalimumab, so it would be digested in a faster way) and reduce its antigenicity, moreover recent studies have assessed that this mAb is not able to pass the placenta so it can also be used on pregnant women. In vitro, this antibody didn’t induce monocytes or lymphocytes apoptosis, complement activation or ADCC. In addition, along adalimumab, it is administered sub-cutaneously, which is a much easier administration route compared to the intravenous one of infliximab, because the patient can perform it even by himself without having to go to the hospital. Another particular aspect of mAb compared to the recombinant protein (Etanercept) is that antibodies (with the exception of certolizumab pegol) can form complexes because they have two Fab portions that can bind to two different molecules of TNF, meanwhile the recombinant protein only binds to a single molecule: the formation of immunocomplexes induces a higher immunogenicity, a higher clearance (through phagocytosis) and higher probability to have Fc-mediated effects. However, antibodies are able to bind to both the soluble and the membrane-bound TNF-alpha, while the recombinant proteins only binds to the soluble TNF. The ant-TNF drugs are used in inflammatory bowel diseases (like Chron’s disease), spondyloarthropathies (like psoriasis), juvenile rheumatoid arthritis and Adult Still’s disease (diseases against which they have a confirmed efficacy); we are now trying to use them also against vasculitis, scleroderma, graft vs host disease, inflammatory myositis, interstitial lung disease, Sjögren’s syndrome, inflammatory eye and ear disease, asthma, hepatitis C and sarcoidosis. Which are the disadvantages of these therapies? If TNF is removed from our body in pathogenic conditions, the pathology is reduced, but if we remove it in physiological conditions we could have some complications, like the reactivation of some infectious diseases (like Hepatitis B and tuberculosis), because TNF is produced in our body against infectious agents so if we have latent infections, we could make them reactivate if we deprive our body of TNF. Sometimes we could also have the reactivation not only of infectious diseases but also of tumors, like skin cancer. We could also have the development of autoimmune diseases (Lupus-like syndrome), demyelinating disorders, congestive heart failure and liver toxicity. Other side effects could be related to the immunogenicity of the drug that we use, especially with mAbs, and with the fact that if we interrupt the therapy, we could have the reoccurrence of the symptoms. Moreover, other disadvantages are related to the costs of the therapy (etanercept costs 15 000 euros per year per patient). [N.B. in this paragraph we talked about drugs used against autoimmune diseases but we have to remember that they cannot cure the disease, instead they are just able to ameliorate the symptoms, because autoimmune diseases are chronic: this is because they are due to genetic abnormalities which cannot be changed and so we cannot fully recover from an autoimmune disease] GROWTH FACTORS A lot of growth factors (GFs) are produced by our body (gamma-INF is also considered a growth factor because it promotes the clonal expansion of T cells), of these we have two specific haemopoietic growth factors which are erythropoietin and thrombopoietin, all the other growth factors have -GF as a suffix. We have a lot of growth factors that have been approved for medical use, of these the most important are G-CSF (granulocyte growth factors), GM-CSF (granulocyte-monocyte growth factor),

Appunti Drugs - Drug Production Process PDF

Document Details

Tags

Related

Summary

Full Transcript

Upgrade to continue