286
`
`VOLUME 2: FACJU7Y DESIGN, STERILIZATION AND PROCESSING
`
`The AF for minimum dose, which oftentimes is the dose measured at an interior location, is
`given by
`
`AF . _ D,er
`m,n - Dmin
`
`(4)
`
`In equation (4), D,ct is the dose measured at the reference location and Dm;n is the dose
`measured at the minimum dose location.
`It is important to note that when reference location dosimetry is used to monitor dose
`during routine processiJ1g of product, the minimum dose at an interior location is not measured
`rather it is calculated on the basis of a statistical relationship given by the AF. For this reason, it is
`standard practice to measure the dose distribution in more than one product load under the same
`processing conditions with three product loads considered the minimum number to be dose
`mapped. Statistical analysis of the data from three dose maps is used to evaluate reproducibility
`in the measured dose and uncertai11ty in the statistical relationship that is used to calculate the
`minimum dose. This estimate of statistical uncertainty in the calculated value of dose can be used
`to set process paramete.rs for routme irradiation of the product.
`
`Dose Mapping Electron Beam
`Because of the much shorter radiation mean-free path of high-energy electrons in materials
`than high-energy photons and the fact that we are dealing with a beam of electrons, shielding
`and scattering effects i11troduced by localized heterogeneities within a carton of product or
`even withm a unit of product in the carton can significantly affect the dose delivered to the
`product. For example, the range of 10 MeV electrons is approximately 5 cm in water and
`polymers that commonly serve as packaging materials and closure systems for pharmaceuticaJ
`products. ln a metal such as stainless steel, the range of 10 MeV electrons is less than 1 cm.
`Therefore, localized high-density regions can result in significant dose gradients within a small
`volume and shadowing of other regions in the carton of product. These factors need to be
`taken into account iJ1 the selection of the locations of dosimeters within the product load. There
`are no standard dose map grids as is sometimes the case for gamma or X-ray irradiation. Dose
`map grids in high-energy electron beam irradiation are unique to each product type. Jn
`electron beam irradiation, it is common practice to use reference location dosimetry for
`monitori11g dose during routine processing of product. An external surface such as the surface
`where the electron beam is illcident on the product load may sometimes serve as the reference
`location or it may be at a fixed location adjacent to the product load and simply referred to as the
`monitoring location. In the case where the reference location is on an externaJ surface, it
`sometimes may also represent the minimum dose zone, which would only require use of an AF
`to calculate the maximum dose delivered to the product load. To establish the reproducibility in
`dose delivered to tJ1e prodt1ct load and estimate the uncertainly i:n the AF(s) that is used to
`calculate dose, mttltiple product loads, that is, typically three, are dose mapped. The uncertainty
`in the dose measurement process should be taken iJ.1to account when setting process parameters.
`
`RADIATION CH EMISTRY
`Radiation Interactions with Parenteral Drug Products
`As we have seen, high-energy electrons injected into a drug product from a high-power
`accelerator or generated within the medium from Compton scattering of energetic photons are
`responsible for the changes ill the properties of the drtig product and its sterilization. These
`high-energy electrons, which typically have energies in the 1 to 10 MeV range, suddenly find
`themselves embedded in the surrounding medium. Atomic electrons of the atoms ill the
`medium effectively shield the attractive force of the positive charges of the nuclei, and the
`high-energy electrons experience only the repulsive Coulombic force that is instantly
`established between them. The velocity of a 1 MeV electron is of the order of magnitude
`1010 cm/sec, which is close to the speed of light. The velocity of atomic electrons is on the
`order of 100 times less. lt takes about 10- 17 seconds for a 1 MeV electron to cross a diameter of
`an atom. During that time an atomic electron remains practically stationary and "feels" the
`risi11g and falling action of the repulsive CouJombic force created by the approaching and
`
`Regeneron Exhibit 1016.301
`
`

`

`RADIATION STERILIZATION
`
`287
`
`leaving of the high-energy electron passing by. The momentum exchanged between the two
`electrons (the product of the electrostatic force and duration of the collision) is small in
`comparison with the kinetic energy of the incident electron but may be large in comparison
`with the binding energy of the orbital electron. If the exchanged energy exceeds the energy that
`binds the electron to an atom (ionization potential), ionization of that atom will occur, whereas
`the exchange of a smaller amount of energy will result in its excitation.
`Studies have shown that the energy exchange events in liquids and solids involve energy
`packets between 6 and 100 eY, the most probable being around 25 eV. This is true in simple
`molecules such as water and cyclohexane (19), as well as in macromolecules such as DNA (20).
`Obviously all materials consisting of low-Z elements, including biological materials aJ1d APis,
`absorb energy by similar mechanisms that occur with similar probabilities. The energy of 25 eV
`is sufficient for the creation of one or two ion pairs and one or two excited molecules in liquid
`water. The small element of volume within which energy deposition occurs and within which
`newly formed species are confined for a limited time is called a spur. Occasionally, a larger
`package of energy is absorbed forming a blob (100 500 eV) or a short sidetrack (500 eV 5 keV).
`Spurs outnumber blobs by about 50:1 and short tracks by abou.t 500:1. For cobalt-60 gamma
`rays and 1 MeV electrons in water, the partition of absorbed energy is approximately spurs:
`75%, blobs: 12%, and short tracks: 13% (21). Essentially the same distribution of probabilities
`exists in water vapor and ice underscoring the random character of primary interactions,
`irrespective of the phase. This leads to the estimate that the absorption of a J MeV electron
`creates about 25,000 spurs, 500 blobs, and 50 short tracks.
`The initial volume of a spur in water may be about 1 nm3 (22), and the volumes of blobs
`and short tracks may be orders of magnitude larger, 10 and 100 nm3
`, respectively. Together
`they may occupy the volume of the order 105 nm3 containing about 106 molecules of water.
`Sterilization dose of 25 kGy is equivalent to the absorption of J .56 x 10211 eV / g requiring total
`absorption of 1.56 x 1014 1-MeY electrons in 1 g of water. The absorption of this amount of
`energy would initially affect J .56 x 10w molecules/gout of 3.3 x 1022 molecules present in 1 g
`of water, or 1 in about 200. Allowing that more than 10 water molecules may be contained
`within a J nm3 spur reduces this estimate to less than one iJ1 2000.
`The above picture is oversimpUfied: there is a distribution of spur sizes and some
`overlapping of spms. Nevertheless, it teaches us that precursors of chemical change are
`initially inhomogeneously distributed only along the tracks of fast electrons while the rest of
`the volume remains unaffected. It also teaches us that a significant fraction of small molecules
`may initially escape ionization or excitation, but that larger molecules will not be spared of
`radiation acth1g directly. lt is also obvious that in solutions, it is mostly solvent molecules that
`absorb radiation energy resulting in the creation of reactive species. The initially inhomoge(cid:173)
`neous distribution of priJnary products: electrons, positive ions, and excited molecules
`thro1.1ghout the irradiated medittm is one of the key featttres of radiation chemistry.
`Spatial inhomogeneity determines the earliest stage of radiation action, which is termed
`physical stage. It starts at 10- 17 seconds with the absorptio11 of energy and extends to
`approximately 10- 13 seconds until thermal equmbrium has been reached. The probability of
`interactions of electronic systems of atoms with photons and electrons during that stage is
`perfectly random, and nothing can be done to reduce it or to decrease the amount of ionization
`and excitation. The energy required for the creation of one ion pair in gas 0/11) is similar (25 30 e V)
`for a wide range of compounds (23), which forms the basis for the expectation that approximately
`the same munber of ion pairs would initially be created, irrespective of the chemical nature of the
`substance. However, the amounts of radiation-induced changes that become measurable at later
`stages greatly differ depending on the medium.
`
`Radiation Chemical Yield
`In an empirical approach to quantify and compare chemical effects of irradiation, the
`measured amounts of radiation-induced chemical changes have been normalized to dose. The
`quantity obtained in this way is called radiation chemical yield (G):
`
`G(X) = C(X)
`pD
`
`(5)
`
`Regeneron Exhibit 1016.302
`
`

`

`288
`
`VOLUME 2: FACJL/7Y DESIGN, STERILIZATION AND PROCESSING
`
`where G(X) is the radiation chemical yield of substance X created, destroyed, or altered; C(X) is
`the concentration of substance X created, destroyed, or altered; p, the density; and 0, the dose.
`The unit of C(X) is mol/J but ru1 older unit (rnolecules/100 eV) is still sometimes used
`(1 mol/J = 9.65 x 106 molecules/100 eV). The knowledge of G values allows the fraction of
`molecules affected by irradiation of 1 kg of some substance to be estimated as:
`
`(6)
`
`C(X) = 10- 3 x G(X } x D x M
`C
`where C is molar concentration of the neat substance and Mis its molecular mass. The larger
`fraction of molecules will be affected by the Larger dose and the larger is the molecule. In
`water, G(X) accounting for all interactions could be on the order of 1 ~1mol/J, which, for the
`dose of 25 kGy, gives C(X)/C = 4.5 x 10-4, or about one out of 2000 molecules, which
`fortuitously well compares with the previous estimate.
`If there were no influence of the medium on the initially produced ion pairs, C(ions) .in
`all media would be 100/W, that is 3 4/100 eV (-0.3 0.4 pmol/J). However, measured values
`of radiation chemical yields of primary species electrons, ions, aod excited molecules strongly
`depend on the time of measurement and the nature of the medimn. This means that they are
`modified by the medimn during the intervening iJ1terval of temporal evolution called
`physicochemical stage that extends from 10- n to 10- 10 seconds.
`
`Liquid Formulations-Radiolysis of Water
`The understanding of physicochemical processes occurring at early stages of radiation act.ion
`helps in devising meaningful ways to mitigate radiation-induced damage to the parenteral
`drng product. Parenteral drugs in solid form or a dry state respond rather favorably to
`radiation. However, liquid formulations particularly those aqueous in nature present more
`challenges. The pecttliarities of aqueou.s radiation chemistry are discussed iJ1 this section.
`An important reaction occmring during physicochemical stage in liquid water is the
`fastest known chemical reaction:
`
`(7)
`
`which generates the strongest known oxidizing species, hydroxyl radical. It can oxidize any
`molecitle with which it comes in contact and is ma.inly responsible for the radiation-induced
`damage of solutes in irradiated aqueous solutions. Another route for the formation of hydroxyl
`radical is the dissociation of excited water molecules that becomes possible in the same time
`window with the onset of molecular vibrations:
`
`H20'-+ H" + "OH
`
`(8)
`
`On the same timescale, the reorientation of dipolar molecules leads to the solvation of charged
`species, notably the free electron becomes hydrated in water, which, as the strongest reducing
`species known, can affect radiation sterilization of aqueous solutions of reducible substances.
`During that time frame radiation-induced species react within spurs or escape from the
`spurs by diffusion into the bulk where homogeneous distribution of reactive species is
`eventually established. The recombination of radical species gives stable molecular products:
`
`"OH+ "OH -+ H202
`
`(9)
`
`(10)
`
`which, however, are of little concern for radiation sterilization of solutions.
`During the physicochemical stage, dielectric properties of the medium have the strongest
`modifying effect on radiation chemical yields of charged species. Dielectric constant of the
`medium detennmes the critical distance at which the Coulombic attractive force of the ion pair
`equals the thermal energy that drives them apart. Only those electrons that escape the
`recombmation with the parent ion become solvated a11d eventually participate in the bulk
`reactions. ln a polar liquid like water the probability that an electron wilJ escape the
`recombination with its parent ion steeply increases with the increase of the initial electron-ion
`
`Regeneron Exhibit 1016.303
`
`

`

`RADIATION STERILIZA710N
`
`289
`
`separation distance. Therefore, free ion yield is high in water and polar liquids and low in
`nonpolar liquids.
`At the beginning of the chemical stage radiation chemical yields (in ~1mol/J) are as
`follows: G(. OH) = 0.28, G(.H ) = 0.06 and G(eaq - ) = 0.27. Until this moment, the only
`modifying action on these yields was that of the medium itself, and no additives could have
`altered them. As it now comes to chemical reactions with the components of the medium, the
`complex interplay of ionization potentials, electron affinities, bond dissociation energies, and
`chemical reactivities of the involved species finally determine the outcome of the chemical
`stage on nanosecond to micro- and millisecond timescales.
`The extremely high rate constant of the reaction given by equation (7) and the high
`molarity of neat water even in concentrated solutions make the reactions given by equations (7)
`and (8) unavoidable. Any attempts to mitigate in advance ill effects of hydroxyl radical induced
`oxidations must admit the impossibility to prevent its formation and recognize that the first
`opportunity to convert it into a more innocuous species occurs only after it has been already
`formed.
`The hydroxyl radical can oxidize any molecule with which it comes in contact and is
`mainly responsible for radiation-induced damage of solutes in irradiated aqueous solutions. lf
`the substance of interest, an AP!, reacts with ·oH radical with the rate constant kAPJ giving an
`unwanted product P, it is possible to find a compound S with a preferably higher reactivity
`with ·oH (rate constant ks), which acts as a scavenger and which does not give P. The hydroxyl
`radical is thus given two channels to react:
`AP[+ •o H ~ P
`
`(11)
`
`s+ ·oH _, no P
`Radiation chemical yield of unwanted product P, G(P) is given by the ratio of _probabilities of
`·o H reacting in the channel giving P to the overall probability of ·oH reaction:
`G(P) = G(. OB)kAPi!AP11/(kAPifAPI] + kslSl)
`G(P) will be at minimum the higher the product k5[S], that is, the more reactive scavenger and
`the higher its concentration. The same formalism is applicable to all other reactive species.
`The hydrated electron and hydrogen atom may be considered a basic and an acidic form,
`respectively, of a reducing species in the radiol.ysis of water. Their interconve.rsi.on is possible
`because the respective chemical equilibria are strongly shifted to the right. In acidic media,
`hydrated electrons are converted into H· atoms:
`eaq - + H30+ - H· + H20
`whereas in basic media all H· become eaq- :
`H• +OH- _. e.q- + H20
`
`(12)
`
`(13)
`
`(14)
`
`(15)
`
`Using scavengers that specifically react only with the oxidizing or the reducing radicals, it is
`possible to achieve the presence of only one kind of radicals. In a reducing medium hydroxyl
`radicals are converted into H• atoms:
`
`(16)
`while in an aqueous solution satmated with N20 (0.02 mole/ L), eaq - are converted into ·ott:
`e.<J- + N20 + H20 ~ ·o H +OH- + N2
`(17)
`Tertiary butanol efficiently removes •oH and slowly reacts with H•, while other alcohols (e.g.,
`isopropanol) remove both tt• and ·oH. At the same time alcohols do not react with e,q_- ·
`
`Aqueous (Liquid and Frozen) Parenterals
`The absorption of radiation energy in a crystalline solid is 11ot focused on a single atom, but a
`collective excitation involving many electxons spread throughout the c1ystal lattice is induced.
`
`Regeneron Exhibit 1016.304
`
`

`

`290
`
`VOLUME 2: FACJLJ7Y DESIGN, STERILIZATION AND PROCESSING
`
`The energy that would have been localized on an individtrnl chemical bond in an isolated
`molecule in gas or in a molecule in solution is distributed over many bonds in a crystal.
`Consequently, radiation chemical yield of decomposition in a crystalline matrix is lower than
`in solution, which is in turn lower than that in gas, Ggas > Gliquid > Gsolid·
`The buildup of free radicals in solids at low doses proceeds proportionally to dose, then
`the rate of their accumulation decreases until the concentration reaches the limiting value. The
`limiting concentration is reached when sufficient free radicals are produced within each
`other's migration volume so that they can recombine. The upper value of the recombination
`radius critical for permanent trapping in a solid is considered to be about 1 nm (24).
`The uptake of radiation energy by a medium is essentially proportional to the total
`number of electrons (valence and bottnd) present in a unit volume, that is, proportional to the
`mass of material exposed to irradiation. On irradiation of solutions most energy is deposited in
`the solvent. In irradiated aqueous solutions, reactive species e. q- , H•, and · oH prodttced by
`radiolysis of water react with any dissolved substances that act as their scavengers and
`consequently sttffer chemical changes. Radiation-induced effects that occur as a consequence
`of the absorption of energy in the target compound are termed direct effects, whereas those
`that occur in the reactions between a target compound and reactive species produced in a
`solvent are termed indirect effects.
`
`Effect of Temperature
`Direct effects are not expected to depend on temperature. The effects of elevated temperature
`on chemical reactions of reactive species in solution that are responsible for the indirect effect
`can be described by the Arrhenius equation. As the activation energies are rather small
`(6 30 kJ/ moD, the effects on reaction rate constants are also not large. The effects of reduced
`temperature are more dramatic because a significant increase of solution viscosity impedes the
`diffusion of reactive species, which leads to their spending more time close to their respective
`places of origin and ultimately, to their enhanced recombination. For example, radiation
`chemical yield of e0q - is reduced by a factor of 10 on reducing the temperature from 5 to
`55°C (25) and that of the hydroxyl radical by a factor of 60 on reducing the temperature from
`20 to 40°C (26). The yields of products derived from electron or hydroxyl radical attack at
`these temperatures in ice would be reduced by about 90% and 99.7%, respectively, compared
`to fluid solutions. Because of the redticed mobility at low-temperature reactions, damaging to
`solute would be possible only at solute concentrations high enough to have solute molecules in
`a region of reactive species formation, which we have estimated to be one in 2000 water
`molecules. However, even at low temperature, larger molecules such as proteins cannot escape
`direct effects.
`
`Effect of Oxygen
`Oxygen normally does not react with stable compounds at room temperatme, but its
`paramagnetic properties make it reactive with free radicals, which are also paramagnetic
`species created by irradiation of AP[s, excipients, or solvents:
`R· +02 ....... Roo·
`The most simple rnute for creating free radicals directly is the dissociation of an excited
`molecule R-H yielding a hydrogen atom and a free radical residue R·:
`(R Hr --+ R° + H°
`(19)
`In an indirect radiation action, the abstraction of a hydrogen atom by H• or ·oH radicals
`formed in the radiolysis of water or dissociative electron attachment by a molecule R X,
`containfog a strongly electronegative substituent X, also yield free radicals:
`
`(18)
`
`R X + eaq - __. R° + x-
`
`(20)
`
`(21)
`
`Regeneron Exhibit 1016.305
`
`

`

`RADIATION STERILIZATION
`
`291
`
`Doubly allylic hydrogen atoms, such as found in polyunsaturated fatty acids, are
`particularly weakly bound to the backbone of a molecule, which makes these locations
`especially vulnerable to oxidation. Peroxyl-free radicals fonned by the reaction given by
`equation (18) propagate a chain reaction:
`Roo· + R H -' ROOH + R°
`which continue to produce damage of an oxidizable substance as long as there is a steady
`supply of oxygen.
`Oxidation is one of the major causes of dmg instability, even without radiation. The ill
`effects of oxidation can be avoided by the exclusion of oxygen that underscores the importance
`of packaging and closure systems. It can also be prevented by the use of compounds that
`interfere with the propagation of radical chains by competing with the reaction given by
`equation (22), which are known as antioxidants. An antioxidant molecule A H itself possesses
`a weakly bou11d hydrogen atom, the abstraction of whicl1 produces free radical A•, that is more
`stable (less reactive) than R· and that therefore cannot further propagate the chain reaction:
`Roo· + A H -+ ROOH + A.
`More detailed aspects of stabilization of pharmaceuticals to oxidative degradation can be
`found in (27).
`
`(22)
`
`(23)
`
`RADIATION EFFECTS
`When considering the effects of radiation on a parenteral dmg product, it is important to take
`into account all elements of the drug product that may be exposed to the radiation
`environment. This includes the container, closure systems, and packaging materials. If the
`drug product was previously sterilized using a modality other than radiation, some materials
`that were selected because of physical-chemical features or tribological attributes may not be
`radiation compatible, whicl1 would entail selection of different materials for the radiation
`sterilization process. Therefore, whenever possible it is important to select the modality of
`sterilization early in the development of a new drug product.
`
`Container/Closure Systems and Packaging
`Most materials that are found in container/ closure systems and packaging consist of different
`types of polymers and glass. ln the evaluation of the effects of radiation on these materials, it is
`important to take into account possible changes in mechanical properties, radiation-induced
`discoloration, and biocompatibility. Because glass is amorphous, its mechanical properties are
`tmchanged when exposed to radiation. However, most glass materials discolor in varying
`degrees when exposed to radiation, which may not be acceptable from the standpoint of
`aesthetics or possibly functional reasons. The degree of discoloration depends on the type and
`amount of impurities in the glass, which are a source for radiation-indttced stable conjugated
`chromophores. Some types of glass such as cerium oxide glass show less discoloration than
`borosilicate glass when exposed to radiation (8). A very high purity glass material such as
`synthetic fused silica also will not discolor when irradiated. Polymers fall into three general
`classes that include thermoplastics, thermosets, and elastomers. Thermoplastics are the class of
`polymers that are commonly selected for containment of a drug product, and closure systems
`are usually elastomeric in nature. A large compendium of information on the effects of
`radiation on these classes of polymers can be found in published references and irom the
`manufacturers of the polymers themselves (28,29). Only a few polymers are not radiation
`compatible and shottld not be used if radiation is the choice for sterilization. Polyacetals, for
`example, Delrin and Celcon, polytetrafluoroethylene, that is, Teflon, and natural polypropy(cid:173)
`lene are not radiation tolerant and should be avoided. Polypropylene auto-oxidizes and will
`continue to degrade following irradiation. A radiation~stabilized polypropylene with
`antioxidants may be used in some applications. Two elastomers that are not radiation tolerant
`and should be avoided are butyl rubber and a fluoroelastomer. For example, butyl rubber is
`friable and will shed particulates. It is important to note that a poor choice in the selection of
`the polymer is not the only reason a part may fail when it is exposed to radiation. Improper
`
`Regeneron Exhibit 1016.306
`
`

`

`292
`
`VOLUME 2: FACJL/7Y DESIGN, STERILIZATION AND PROCESSING
`
`processing of a polymer or incorrect design may lead to failure of a part that is .irradiated even
`though the polymer is considered radiation compatible. For example, thermoplastics are often
`fabricated using ru1 injection molding process. lf the conditions for fabrication are not
`optimum, for example, temperature during the mold process, the final part may contain
`residual tensile stresses. Irradiation leads to breakage of molecular bonds in the polymer.
`Because of the presence of residual tensile stresses, crazing and microcracking of the polymer
`may occur. In the design of a part, stress raisers should also be avoided, for example, avoid
`sharp corners in design of the part.
`
`Radiation Effects-Excipients, Biopolymers, and APls
`Excipients are used to promote pharmacological action of an APT by formulation of the drug
`product in a viable delivery system. Examples of excipients, some of which may appear in
`parenteral medications, include gum Arabic, talc, starch, and paraffin. The principal effects of
`radiation that need to be taken into account are change in color, change in pH, and lowering of
`viscosity. Past studies have shown that excipients should respond favorably up to doses
`required to sterilize the drug product, that is, 25 kGy or less (30). Loss of viscosity may be of
`some concern in some cases. lJ1 particular, some thickening agents may suffer a significant loss
`in viscosity at relatively low doses of radiation. Radiation-induced chain scissions in the
`aliphatic molecular structure of the cellulose component significantly lowers its molecular
`weight with a concomitant decrease in the viscosity of the thickening agent. Addition of a
`radical scavenger may significantly improve the radiation stability of the thickening agent.
`Biopolymers are used for controlled drug release (CDR) and controlled drug delivery
`(CDD) of APls following parentral administration (31). Biopolymers react to radiation in a
`manner similar to other polymers. There is a possibility of chain scissions, cross-linking, and
`formation of free radicals. The principal changes of concern from irradiation of biopolymers
`include change in color and physical properties, which may lead to a change in the drug
`release characteristics of the biopolymer. For example, polyester polymers such as poly(lactic
`acid) (PLA) and copolymer poly(lactic acid-co-glycolic acid) (PLGA) are routinely used in
`CDR/CDD applications. Radiation will reduce the molecular weight of these polymers, with
`the percentage reduction increasing with increase in absorbed dose. For drug products that
`have low levels of microbiological contamination, it is possible to set an acceptable minimum
`dose that satisfies the desired SAL while maintaining a maximum dose that keeps the
`reduction in molecular weight within acceptable limits.
`The principal effects of radiation on an APT are formation of small amounts of
`degradation by-products and possible changes in the chemical-physical properties of the API
`including pH, color, and viscosity. The radiation-induced degradation by-products may
`produce toxic extractables that need to be taken into account in the evaluation of the
`biocompatibility of the APL Changes in the chemical-physical properties of the APl cottld
`affect the efficacy of the drug product, that is, its potency. Because a vast variety of chemical
`entities may appear as the AP!s, it is almost impossible to accurately predict radiation
`sensitivity of individual compounds. Previous work on particular or related molecules may
`inform and guide the assessment of radiation stability of an API.
`The effects of irradiation on drugs have been attracting the attention of researchers over
`the past 60 years. Bibliometric count finds about 1400 references until the year 2000, peaking
`in the seventies. This literature has been periodically reviewed and a compilation of results
`from the selection of 217 papers on some 380 APl.s has recently been published in form of an
`encyclopedia (32). Most of the included drngs and excipients are used in sterile prodt1ct
`formulations suitable for parenteral administration. The material included in another more
`recent review (33) is partially overlapping with the former one giving, in addition, an insight
`into the more recent work, mainly originating from the authors' group. These data may
`provide clues to the parameters affecting the radiation stability of a drug, the types of possible
`radiolytic damage, and radiation chemical yields of stable radiolytic products under a variety
`of irradiation conditions. Together with radiation chemistry principles expounded in the
`previous section, these data can help the optimization of key parameters to reduce the
`radiolytic degradation of water-based parenteral dmg products. API's in a dry formulation, for
`
`Regeneron Exhibit 1016.307
`
`

`

`RADIATION STERILIZATION
`
`293
`
`example, powder or freeze dried, are being success.fully terminally sterilized on a commercial
`level using radiation. Parenteral medications in a liquid form present a greater challenge.
`
`IRRADIATION OF SPECIFIC DRUG PRODUCTS
`Vaccines
`The use of radiation to inactivate a pathogen in the preparation of a vaccine was explored at an
`early point in the evolution of the radiation sterilization industry (34). These early studies were
`typically conducted at relatively high doses of radiation, that is, >25 kGy, which was
`considered necessary to inactivate the pathogen. Even so, some successes were observed
`wherein sterility was achieved while the antigenic properties of the vaccine were preserved.
`Most of these studies appear to only have advanced to a preclinical stage. Over the past several
`years, there has been a renewed interest in the use of radiation in the preparation of vaccines.
`The reemergence of certain infectious diseases such as tuberculosis may have stimulated this
`renewed interest in vaccines that are prepared using irradiation. Dependent on the
`microorganism, the dose of radiation to inactivate the pathogen may be relatively low. For
`example, researchers at the University of California, San Diego, have shown that Listeria
`111011ocytoge11es, a bacterial pathogen, was inactivated at doses as low as 6 kGy and the
`irradiated vaccine still triggered long-term immunity in the vaccinated animals (35). However,
`viral pathogens, which typically have significantly higher D10 values than bacterial pathogens,
`may reqttire much higher doses of radiation, that is, greater than 25 kGy, to inactivate the
`pathogen. On the basis of studies that have been conducted over the past several years, a
`significant advantage of radiation in the preparation of vaccines may reside in the possible
`formulation of vaccines in a dry state, for example, freeze dried (36). A vaccine that is prepared
`in this manner could possibly be stored for long periods of time in an unrefrigerated state,
`shipped world wide to a location of need, ru1d reconstituted on site.
`
`Proteins
`Protein drugs are specific, exert their effects at low concentrations, and their virtually limitless
`number enables their use to influence a large variety of biological processes. Therapeutic
`proteins include monoclonal antibodies, growth factors, cytokines, soluble receptors,
`hormones, and proteins that block the hmction of a variety of infectious agents. Specific
`hmctions of proteins in the body strongly depend on their structu.res.
`Proteins are characterized by four levels of structural organization. Primary structure of
`proteins is defined by the amino acid sequence. The ability of antigenic structures to elicit
`immune response is mostly a sequence-dependent property. At this (primary) level of
`structu