DRUG DELIVERY: FUNDAMENTALS AND APPLICATIONS

Hydrogels are polymer networks that contain a substantial amount of water. Dry polymer networks can absorb tens, hundreds, or even thousands of times their weight in water without dissolving, as illustrated in Figure 14.1. As a result, hydrogels have properties that are very similar to those of soft biological tissues such as cornea, cartilage, and intercellular matrix, and because of these characteristics, they are of great utility in pharmacy and in other biomedical fields. First, they are soft and slippery and have a very low interfacial tension when they are in contact with tissues. As a result, they often cause little irritation. Second, hydrogels can behave both like solids and like liquids. Like solids, they are elastic and return to their original shape after they are momentarily deformed. Like liquids, they permit diffusion of small molecules such as salts, metabolites, and drugs. Under proper circumstances, they also allow larger molecules such as proteins and nucleic acids to diffuse.

A hydrogel can be envisioned as a bulk of water through which polymer strands are suspended. As illustrated in Figure 14.2, these strands are connected together by cross-links, which provide elasticity and prevent the strands from coming apart. The distance between strands and the frequency of cross-linking, which determine the mesh size, can be controlled; by this means, the partitioning of molecules into the hydrogel and their rates of diffusion can be regulated. For this reason, hydrogels are often used in controlled-release drug delivery.

FIGURE 14.1 A hydrogel is a dry polymer network that is swollen with water but does not dissolve.

FIGURE 14.2 At the microscopic level, a hydrogel consists of cross-linked polymer chains immersed in water. The distance between chains and the density of cross-links determine the mesh size of the hydrogel.

As elastic materials, hydrogels can exert forces against confining or attached structures. As will be described later, the degree of swelling and shrinking of hydrogels can be affected by physical and chemical stimuli. Thus, the hydrogels can act as “artificial muscles,” which direct the motion of such structures. This property has also been utilized in controlled-release applications.

Table 14.1 provides a list of present and potential future applications of hydrogels. Soft contact lenses were initially developed in the early 1960s as alternatives to hard contacts, which are irritating and can damage the cornea. While soft contact lenses may contain only about 40% water, this is enough to confer flexibility, slip, and compatibility with the underlying cornea. Because optical properties are of foremost importance, hydrogels in contact lenses must remain constant in their swelling during use. Other applications involving hydrogels with constant swelling are listed in bold in Table 14.1.

A tantalizing property of hydrogels is their ability to alter their swelling depending on an applied stimulus, as depicted in Figure 14.3. Commonly studied stimuli include changes in temperature, pH, and ionic strength, but other possible stimuli include the concentration of specific ions such as Ca²+ or glucose and other sugars, and the presence of specific antigens or DNA sequences, electric or magnetic fields, and even light. Applications involving “stimulus-sensitive” hydrogels are listed in italics in Table 14.1. (Note: some investigators refer to stimulus-sensitive hydrogels as “smart” or “intelligent.” We eschew such anthropomorphic terminology.)

TABLE 14.1
Present and Future Applications of Hydrogels

Present	Future^a
Soft contact lenses	Burn, wound dressings
Gel permeation chromatography	Catheter, suture coatings
Ion exchangers	Fire retardants
Gel electrophoresis	Stimulus-sensitive drug delivery
Drug delivery (e.g., capsules, matrices)	(Bio)sensors
Cosmetics	Microparticulate drug and gene carriers
Food additives	Tissue culture, tissue engineering, and bioartificial organ supports
Fragrances	“Active” separations
Dessicants	Artificial muscle
Toys	Switchable chromatography
^a Applications in which size and shape are expected to be constant in time are in bold, while applications in which the hydrogel changes size or shape in response to various stimuli are in italics.

FIGURE 14.3 Stimuli that may affect hydrogel swelling.

This chapter serves as an introduction to hydrogels. More details can be found in the references listed at the end of the chapter. We first describe methods of synthesis. We then discuss factors that affect their swelling, elasticity, solute partitioning, and permeability properties. Finally, examples of present and future applications in the pharmaceutical arena are presented.

14.2 FORMATION OF HYDROGELS

14.2.1 HYDROGELS BASED ON BIOPOLYMERS

“Natural” hydrogels are produced from biological polymers. Because they are the result of biological synthesis, the structure of biopolymers comprising the hydrogels is often well controlled. Their building blocks, by nature, are often biocompatible, although care must be taken to purify the biopolymers and remove antigens and pyrogens. Examples of biopolymers used to form natural hydrogels include fibrous proteins such as collagen, gelatin (denatured collagen), and fibrin and polysaccharides such as calcium alginate, agarose, hyaluronic acid, cellulose and its derivatives, dextrans, chitosan, carrageenans, pectins, and pullulans.

Long-chain biopolymers often form physical cross-links by winding two or more chains to form helical domains. Such cross-links can be loosened by melting or by applying chemical denaturants. Calcium alginate hydrogels are produced when alginic acid chains form coordination complexes that are wrapped around calcium ions, leading to a so-called “egg case” structure. When calcium is replaced with sodium ions, these cross-linking structures are dissolved and the alginate becomes fluid. Physical cross-linking can also occur by formation of microcrystalline domains, usually driven by hydrogen-bonding interactions, as in cellulose, or by aggregation of hydrophobic domains as in pullulans.

Physical cross-links are advantageous since they can often be formed and reversed by mild processing techniques. On the other hand, physical cross-links may not be permanent, and with persistent stress, physically cross-linked hydrogels may deform permanently. Hydrogels formed from fibrous proteins can form more permanent, stress-resistant structures if they are chemically cross-linked. Here, a linker molecule such as glutaraldehyde forms covalent bonds between side chains on separate protein molecules. Companies specializing in biochemicals sell a variety of such cross-linkers, which are also used for protein conjugation. Chemical cross-linking of polysaccharides is often accomplished by chemistries that have been developed for synthetic polymers, as will be described in the following texts. Alternatively, biopolymers can be cross-linked by exposure to radiation, which works by forming free radicals along different polymer chains, which find each other and recombine, forming covalent cross-links.

14.2.2 SYNTHETIC POLYMERIZATION

Synthetic polymers and their hydrogels are usually formed from petroleum-based organic chemicals. Except for some unusual cases, these systems exhibit less intrinsic structural regularity than biopolymers. Also, only a limited set of synthetic hydrogels are biocompatible. Nevertheless, synthetic hydrogels can be produced cheaply, and their chemical compositions and network structures are easily studied and controlled. In recent decades, there have been major advances in the synthesis of organic polymers, resulting in improved quality of hydrogels. Here, we will describe several methods for preparing hydrogels, beginning with a brief review of polymer synthesis.

14.2.2.1 Step-Growth Polymerization

In step-growth polymerization, polymer-building blocks or monomer units contain end groups that react with each other to form larger molecules having the same end groups. Imagine a monomer with two end groups, say A and B, that react with each other. Two monomers of structure A–B will form a dimer of structure A–B′–A′–B. The prime symbols in the middle indicate that the joined A and B units might be modified as a result of the reaction. For example, if A is an acid group and B is a hydroxyl group, then B′–A′ will be an ester linkage. If A is an acid group and B is an amide group, then B′–A′ will be an amide linkage. In these two examples, a water molecule is liberated in a condensation reaction. However, the dimer still retains unmodified A and B end groups, which can enter into reaction with other A–B monomers. Similarly, already formed chains with A and B end groups can react with each other to form even longer chains. Thus, reactions such as

$A - {[B^{'} - A^{'}]}_{x} - B + A - {[B^{'} - A^{'}]}_{y} B \to A - {[B^{'} - A^{'}]}_{x + y} - B$

are possible, for any chain lengths x and y. If the reactions are condensations, then the step-growth polymerization can also be called a condensation polymerization. The two reactions mentioned earlier lead, respectively, to polyesters and polyamides. It should be noted that A and B need not be distinct, provided that an end group can react with an identical end group. For example, two organic acids, –COOH, can combine to form an anhydride linkage, with a water molecule liberated. It is also possible to perform step-growth polymerizations from the initial monomers, A–A and B–B, since these can react to form A–A′–B′–B, which can then go on to form longer polymer chains.

Since condensation reactions are reversible, water removal is usually required for sustained growth. Also, step-grown reactions are particularly sensitive to impurities containing only one reactive end, which will cap the polymer and prevent further growth.

The step-growth mechanism just described will only form linear chains. As with biopolymers, such chains can be cross-linked through their side groups. However, an even simpler method exists. Along with the A-B monomers, a fraction of monomers with structure R-A_f can be included, where R is a central group that is surrounded by f A groups, where f > 2. Similarly, other monomers of form R-Bf can be included. The R groups in these two monomers can be the same, but they need not be; nor do the fs, so long as they are both greater than 2. The fs are called functionalities. With these multifunctional monomers, the ends of growing linear chains can be linked together into networks, as diagrammed in Figure 14.4.

One characteristic of step-growth polymerization is that the individual polymer chains, or the chains connecting cross-links, are of variable length. The growth sequences of different chains vary, and the result is a distribution of chain lengths. Therefore, while the average chain length can be specified or determined experimentally, this variability must be regarded as a fundamental property.

FIGURE 14.4 Schematic of a cross-linked hydrogel with chains formed by AB units, cross-linked by R–A₃ and R–B₄ units.

14.2.2.2 Chain-Growth Polymerization

In chain-growth or addition polymerization, individual chains form by initiation at one end, where a reaction center is formed. Monomers are then added, one by one, during propagation steps, with the reactive center shifting to the last monomer added. Propagation proceeds until a termination event occurs, in which the growing chain reacts with a species that permanently blocks further monomer addition.

Free radical polymerization, the most common version of addition polymerization, is illustrated in Figure 14.5. In this class of reactions, the monomer units contain vinyl groups, denoted by ╕, which signifies a carbon double bond along with a (here unspecified) side group. When attacked by a free radical on one side, the double bond is converted to a single bond, and the free radical is shifted to the other side, where it is now available to attack a second monomer. This second monomer then attacks a third, and so on. Initiation of this chain reaction is brought about by thermal, photo, radiation, or redox decomposition of an initiator species, I, converting it to a free radical, which we denote by I·. The result is a long chain with the initiator at one end and a still reactive free radical on the other.

In most free radical polymerizations, propagation is ultimately terminated when two free radicals meet and either form a covalent bond joining the corresponding chains (combination) or having one chain transfer its unpaired electron to the other (disproportionation). Two other possibilities are as follows: (1) polymerization is terminated by stray free radicals that are present in the feed. Reactive oxygen species (ROS) are of particular concern, and care must be taken to minimize the presence of molecular oxygen, and (2) the propagating free radical extracts a hydrogen atom from another molecule, terminating the growing chain. However, a new free radical is formed on the “victim” molecule, initiates a new polymerization. This mechanism is called chain transfer and is sometimes used as a means to control molecular weight by addition of chain transfer agents. When the hydrogen atom is extracted from another polymer chain, a secondary branch may grow off the latter, creating a treelike or branched architecture.

Following initiation, which is often rate limiting, the propagation and termination steps are fast. Thus, individual chains are formed rather quickly and at different times within a polymerization run. This is in contrast to step-growth polymerization, in which all chains tend to grow at the same time.

FIGURE 14.5 Steps in vinyl polymerizations: (a) initiation and propagation; (b) termination by recombination; (c) termination by disproportionation; (d) termination by stray free radicals, e.g., reactive oxygen species (ROS); and (e) chain transfer.

Free radical polymerization, along with other addition polymerizations, possesses several advantages. It is relatively easy to carry out, and there exists a wide variety of vinylic monomers containing different side chains, which endow the polymer with desired properties. More than one monomer species can be included in a polymerization in order to fine-tune these properties or to confer two or more properties at once. Polymers containing more than one monomeric species are called copolymers. The structure of a copolymer depends on the rate that each monomer reacts with its own kind rather than with other types of monomers. When there are two types of monomers, chain propagation proceeds according to the following mechanism:

$\begin{array}{l} - M_{1} \cdot + M_{1} \overset{k_{11}}{\to} - M_{1} - M_{1} \cdot \\ - M_{1} \cdot + M_{2} \overset{k_{12}}{\to} - M_{1} - M_{2} \cdot \\ - M_{2} \cdot + M_{1} \overset{k_{21}}{\to} - M_{2} - M_{1} \cdot \\ - M_{2} \cdot + M_{2} \overset{k_{22}}{\to} - M_{2} - M_{2} \cdot \end{array}$

where k₁₁, k₁₂, k₂₁, and k₂₂ are the propagation rate constants. The reactivity ratios r₁ = k₁₁/k₁₂ and r₂ = k₂₂/k₂₁ provide information regarding the preference of monomers to react with their own kind and, hence, the tendency of monomers to be arranged (1) in blocks or (2) in alternating sequences. If r₁ and r₂ are much larger than unity, the likelihood of forming blocks of similar monomers increases, while if both are less than unity, the monomers tend to be arranged alternately. When the product r₁r₂ is close to unity, then the monomers are randomly incorporated in the copolymer chains, so long as the supply of both monomers remains adequate. It should be noted that unless special precautions are made, the monomer feed ratio will drift as the polymerization progresses, and there will be changes in monomer content.

Like step-growth polymerization, free radical polymerization produces chains whose lengths present a rather broad distribution. The reason for this is that the termination step can occur at any time, independent of the extent of propagation. Other kinds of addition polymerization can minimize such variability in chain length by suppressing the termination step. In anionic and cationic polymerizations, the growing chains have charged reaction centers at their growing ends, so termination by encounter between two chains does not occur. In ring-opening polymerization, the terminal monomer in a chain reacts with a cyclic-free monomer, which opens as it joins the polymer. Normally, there is no basis for interchain termination reactions.

In the past two decades, novel free radical polymerization schemes, including atom transfer radical polymerization (ATRP) and reversible addition-fragmentation chain transfer (RAFT) polymerization, have received much attention. In these schemes, most of the growing chain ends are “masked” by added reagents, slowing both the propagation and termination steps. Since termination is bimolecular, the likelihood of two unmasked chains finding each other and recombining or disproportionating is severely reduced. At the cost of slower polymerization, much more uniform polymer chain lengths are obtained.

Anionic, cationic, ring-opening, ATRP, and RAFT polymerizations are often called living, because the chain ends remain active even when the monomer supply is exhausted. It is therefore possible to produces block polymers, in which chains consisting of one kind of monomer serve as macroinitiators for another kind of chain. While typical applications of block polymers lie in the adhesive area, particular systems have also been shown to self-assemble into micellar and vesicular (polymersome) systems (see also Chapter 5, Section 5.5.3). As will be discussed in Section 14.2.5, properly constructed block polymers can also form physically cross-linked hydrogels.

14.2.3 STRUCTURAL CHARACTERISTICS OF LINEAR POLYMERS

We have already emphasized that the molecular weight of synthetic polymers is subject to random variation. Further structural variation can be introduced by copolymerization. Besides molecular weight and composition, another characteristic of a polymer is its tacticity. During an addition polymerization, the reactive center may be chiral with respect to the side chain, and sequential pairs of side chains may take on trans or cis configurations. Usually, this sequence of chiral additions is random, and the polymer is atactic. Chains in which all the side groups are added on the same side are called “isotactic,” whereas if the side groups alternate consistently, then the chain is considered syndiotactic. The latter two situations may require special polymerization conditions, e.g., catalysts, but there is a potential benefit insofar as the resulting polymers can form crystalline domains, or crystallites, by aligning with neighboring chains. Such crystalline domains form physical crosslinks, which may be stable even in the presence of solvent. For example, properly processed polyvinyl alcohol (PVA) can form stable hydrogels that absorb a substantial amount of water, even though there is no chemical cross-linking. Atactic polymers cannot crystallize. Similarly, copolymers must either consist of long homomonomeric blocks (blocks containing the same monomer) or long alternating sequences, in order to crystallize.

14.2.4 CHEMICALLY CROSS-LINKED NETWORKS

In Section 14.2.2.1, it was shown how step-growth polymers could be cross-linked by including multifunctional monomers. Irradiation followed by radical recombination was also mentioned as a general method for forming cross-links between polymer chains. In vinyl polymerization, the most common method of cross-linking is to include a small amount of a divinyl monomer into the polymerization reaction. As illustrated in Figure 14.6, this monomer can be incorporated into two independent growing chains. If several cross-linkers are incorporated into each chain, then the chains will have multiple connections to each other, forming a macroscopic network. Cross-linkers are incorporated randomly so that the distance between cross-links is random, the same as was pointed out for hydrogels formed by step-growth polymerization. Because of this randomness, important properties such as mesh size (Figure 14.2), which determines in part the diffusivity of molecules through the hydrogel, can only be characterized as averages.

A nonobvious characteristic of many hydrogels is their nonuniformity, which goes beyond the randomness seen in chemical structure distances between cross-links. Usually, hydrogels are synthesized in the presence of a solvent, such as water, in order to prevent overheating, which results from the exothermic nature of polymerization reactions. As monomers join to form polymers, they exhibit a tendency to phase separate from the solvent, forming small domains that are rich in polymer, alternating with regions that are rich in solvent. These variations in concentration are frozen in upon cross-linking. This phenomenon called “microsyneresis” may, if not controlled, produce a cloudy appearance. It can also lead to substantial weakening of the hydrogel structure and deviation in properties that are predicted assuming uniformity.

In recent years, there have been efforts to form “regular” hydrogels with better defined distances between cross-links and fewer problems with microsyneresis. One popular technique favored by those interested in preparing hydrogels for embedding of cells, e.g., for tissue engineering, is to use polyethylene glycol (PEG) chains that are terminated on both ends by vinyl (usually acrylate) groups. These macromers, which are reasonably monodisperse (nearly uniform in length), are “zipped” together by free radical polymerization of the end groups. Because the PEG chains are of nearly uniform length, are already in a polymerized state before cross-linking, and are hydrophilic, microsyneresis is reduced. Similar approaches involving modification of biopolymers such as dextrans with acrylate groups, followed by “zippering,” have been tried. Copolymers with a small fraction of side chains that can be cross-linked in response to light (photocross-linkers) can also be used to form uniform hydrogels, although the distance between cross-links in this case is more randomized. Most recently, nearly uniform length PEGs have been end-functionalized with groups that can condense with tetrafunctional RA-f (f = 4) cross-linkers, to form the so-called diamond hydrogels. Such hydrogels have been claimed to have “perfect” structure insofar as they are essentially free of inhomogeneities, and it has been shown that they have mechanical strengths greatly exceeding those of conventionally cross-linked hydrogels.

FIGURE 14.6 Cross-linking of two growing chains by propagating their free radicals through a divinyl monomer.

Since cross-links provide cohesiveness to the hydrogel, it follows that by incorporating biodegradable cross-linkers, the hydrogel can be programmed to lose its strength over time. This may be important for injectable and implantable systems, since it obviates the need to retrieve the hydrogel from the body after its purpose (e.g., for drug delivery or cell engraftment) has been served.

14.2.5 PHYSICALLY CROSS-LINKED NETWORKS

Physically cross-linked hydrogels can be formed using block polymers. Of particular interest are systems that are fluid at room temperature but self-assemble into 3D network structures at body temperature (Figure 14.7). Here, we will discuss two kinds of triblock systems (consisting of three polymer blocks in series), one is called ABA and the other ABC. The B block in both systems is a hydrophilic polymer, while the A and C blocks are hydrophilic at room temperature, but hydrophobic at body temperature. This thermosensitivity is attributed to the tendency of water to form hydrogen-bonded “cage” structures around hydrophobic groups. These cages are broken up with increasing temperature, causing water to migrate away, followed by association of hydrophobic segments. (Incidentally, this mechanism of hydrophobicity is also responsible for the stability of globular proteins, in which hydrophobic amino acids are sequestered in the protein molecule’s “core” and are surrounded by a “shell” of hydrophilic amino acids.)

Self-assembly of ABA and ABC triblocks is depicted in Figure 14.7. The hydrophobic A and C blocks form physical cross-links by associating with their own kind. If the A and C blocks are incompatible, then they will form separate cross-linking domains. At low concentrations, ABA triblocks may form “flowers” due to A blocks associating with each other within the same chain, while this should not occur for ABC triblocks. Thus, ABC triblocks with incompatible A and C blocks are disposed to form hydrogels at lower overall polymer concentrations than ABA triblocks. Figure 14.8 shows how an increase in temperature converts a free-flowing polymer solution to a rigid hydrogel. This transformation is reversible—the fluid state is recovered upon cooling.

FIGURE 14.7 Self-assembly of (a) ABA and (b) ABC triblock polymers into networks, with coagulated A and C blocks forming physical cross-links.

FIGURE 14.8 Reversible conversion of a triblock polymer solution with thermosensitive outer blocks, from a free-flowing solution to a rigid gel (suspended on top of the inverted vial), upon heating.

“Thermogelling” systems are of considerable interest because they can be prepared and stored at room temperature and then injected through a syringe needle into a body space. Once in the body where the temperature is 37°C, the polymers form gels, which act as depots for drug release. By incorporating biodegradable elements, these systems resorb after they have carried out their task.

14.3 FUNDAMENTALS OF PARTITIONING, PERMEABILITY, AND SOLUTE-RELEASE PROPERTIES OF HYDROGELS

A primary role of hydrogels in pharmaceutical applications is to modulate the rate of drug release. Delivery can be from a reservoir, with the hydrogel serving as the rate-determining membrane. In some cases, the hydrogel should simply allow free movement of substances from one side of a hydrogel membrane to the other. For example, soft contact lenses enable corneal access of oxygen from air and nutrients from lachrymal (tear) fluid. In other cases, the hydrogel is designed to retard or control movement.

Alternatively, the hydrogel can be loaded with drug, either in the dissolved or in a solid particulate form, and then release it at the target site. The hydrogel is then called a monolith, and it acts both as a storage medium and as a matrix that controls the rate of drug release. A simple example of a monolith is a contact lens soaked in a drug solution. When the drug concentration inside the lens reaches its equilibrium level, the lens is placed on the cornea, where it releases the drug into the lachrymal fluid. The drug then diffuses through the cornea to the aqueous humor, providing therapeutic levels over several hours. Alternatively, drug can be incorporated into a hydrogel during polymerization, or the hydrogel can self-assemble around drug molecules or particles in response to an environmental change such as increase in temperature, as occurs in injectable triblock systems.

The difference between a membrane and a monolith structure is depicted in Figure 14.9. Roughly speaking, one may say that drug diffuses through a membrane and out of a monolith.

For a hydrogel membrane of uniform structure, a simple equation relates the steady-state flux, J_ss, of drug to its concentration difference ΔC, across the membrane:

$J_{s s} = (\frac{K D}{h}) Δ C$

(14.1)

FIGURE 14.9 (a) Membrane versus (b) monolith. Arrows depict drug flux.

That is, the rate of transport of molecules across the membrane is determined by (1) ΔC, which is the driving force for diffusion; (2) the molecule’s ability to enter the membrane, as determined by the partition coefficient, K; (3) the speed at which the molecules can move inside the membrane, as determined by the diffusion coefficient, D; and (4) the distance that the molecules must travel, as determined by the thickness, h, of the membrane. The combination KD/h is called the permeability of the membrane. Since the partition and diffusion coefficients are fundamental characteristics of the hydrogel and the drug molecules, we discuss them in detail here.

The partition coefficient is the ratio of solute concentrations inside and outside of the hydrogel at equilibrium. If the drug molecule is hydrophilic, then it will prefer the aqueous space inside the hydrogel, and an important factor affecting partitioning is the water content or degree of hydration of the hydrogel. When both the drug and polymer are more hydrophobic, the drug may be attracted to the polymer chains, and decreased water content may actually lead to increased partitioning. Partitioning is also favored when the molecule matches a template receptor, as discussed later.

Polymer chains must rearrange themselves to accommodate the drug molecule’s volume, and in doing so, they lose some of their freedom to move. This effect is stronger with increasing drug molecule size, leading to reduced partitioning. The presence of cross-links further limits the degree that chains can accommodate large drug molecules. At sufficiently high cross-link densities or above a critical drug molecule size, it will not be possible for a drug molecule to fit between cross-links, and no partitioning or permeation will occur. Figure 14.10 depicts how the diameter of various protein molecules affects their ability to partition into microgels of methacrylic acid, which decreases as the proportion of cross-linking monomer (N,N′-methylenbis(acrylamide)) increases and the length of polymer chains between cross-links decreases.

Similar factors affect the diffusion coefficient. Very small, hydrophilic molecules will diffuse freely through the aqueous space in a highly swollen hydrogel, hindered only by occasional obstruction and by added fluid drag due to the presence of the polymer chains. With increasing size of a hydrophilic molecule, the same factors that limit partitioning come into play. Motion of the drug molecule requires displacement of polymer chains in its path, which again will be constrained by cross-links. Hydrophobic molecules that are attracted to the polymer chains may exhibit slower transport, as they may diffuse by “inching” along the chains.

Ionized drugs may exhibit various permeability behaviors in ionized hydrogels. When the drug and hydrogel are both positively or both negatively charged, they will repel each other, leading to reduced partitioning. When the drug and the hydrogel are oppositely charged, partitioning increases due to electrostatic attraction. This is the Donnan effect, which is discussed in Section 14.4.1 with regard to swelling of polyelectrolyte hydrogels. Monovalent drug molecules that partition into the hydrogel will diffuse relatively freely, especially in the presence of physiological electrolytes. Divalent or more highly charged ions may, however, be strongly attracted to oppositely charged ionic groups and linger in their vicinity, slowing down diffusion. Hydrogels bearing ionizable acid groups are particularly useful in pharmacy since more than 75% of drugs have base groups. The formation of ionic complexes depends on the degree of ionization of both species, which, in turn, depends on the pH of the medium.

FIGURE 14.10 Effect of the proportion of cross-linker agent (feed ratio 0.10–0.25) on the hydrogel mesh size, which, in turn, determine the molecular size of proteins that can enter in the hydrogel (GAPDH, glyceraldehyde-3-phosphate dehydrogenase). (Reprinted with permission from Eichenbaum, G.M., Kiser, P.F., Dobrynin, A.V. et al., Investigation of the swelling response and loading of ionic microgels with drugs and proteins: The dependence on cross-link density, Macromolecules, 32(15), 4867–4878, 1999. Copyright 1999 American Chemical Society.)

When the drug and the polymer attract each other, then high partitioning and reduced diffusivity effects may sometimes cancel each other out. In addition to hydrophobicity and strong ionic attraction, there may be specific interactions. For example, hydrogels have been produced with sites that strongly bind heparin, and heparin partitions strongly but diffuses very slowly through these hydrogels. Templated hydrogels (described in Section 14.4.2.5) may also have this feature.

In monolithic systems, partitioning is not an issue per se, but drug solubility may limit the amount of freely diffusible drug inside the hydrogel. It may be convenient to have a dispersion of undissolved drug particles in the hydrogel, which act as miniature drug reservoirs. Solubility and, hence, dissolution rate of drug from these particles into the hydrogel matrix will be affected by the same factors that affect partitioning, such as hydrophilicity/hydrophobicity, volume fraction of water, and cross-link density. Also, diffusivity and, hence, release rate will be affected by these structural factors, as before.

While a hydrogel is formally defined as a water-swollen network, it can be prepared and administered in the dry state, and it is possible to control the rate of water uptake into an initially dry polymer and, hence, its rate of swelling and release properties. A useful control parameter is the polymer’s glass transition temperature, or T_g, which is the temperature above which the polymer is soft and rubbery, but below which it is hard and glassy. Drug diffuses through the polymer extremely slowly when it is in its glassy state. When water enters the polymer, it plasticizes it, converting it to the wet, rubbery state that permits more rapid diffusion. The rate of plasticization and hence swelling decreases with increasing T_g, in general. The polymer must also be intrinsically hydrophilic for water to enter the polymer and swell it. For example, poly(methyl methacrylate) (PMMA) and poly(hydroxyethyl methacrylate) (PHEMA) both are glassy at room temperature, but only the latter, which has pendant hydroxyl groups, will absorb sufficient water to become rubbery. By combining HEMA and MMA at specified ratios into cross-linked networks, a series of monoliths with different swelling and drug-release rates can be formed. Typically, a moving front separating a glassy core from a rubbery periphery is observed, and drug release is determined by the rate of advancement of this front, as depicted in Figure 14.11 (see also Chapter 7, Figure 7.2).

FIGURE 14.11 Drug release from a glassy polymer network. The network swells to form a rubbery hydrogel periphery surrounding a glassy core. Drug is released through the outer periphery to the external fluid as the core recedes. Eventually, the core vanishes and drug release ceases.

Biodegradability is another property that, when incorporated into a hydrogel, can influence its drug-release properties. As chains and cross-links are degraded either by water or by enzymatic reaction, the hydrogel loosens and the diffusion of drug molecules can proceed more quickly. Alternatively, degradation can be designed to occur after the release process is finished, with the degradation products resorbed or cleared from the body.

14.4 STIMULUS-SENSITIVE HYDROGELS

14.4.1 A BALANCE OF FORCES DETERMINES THE DEGREE OF SWELLING

We have already noted that polymer synthesis often leads to considerable structural variability in chain length, tacticity, sequence (for copolymers), and distances between cross-link points. Almost all synthetic and many natural polymer chains also manifest randomness in their individual geometrical configurations. Naively, one might assume that polymer chains will have a linear, rodlike configuration in space. Although polymer chains that are crystallized may contain linear runs that are aligned with each other, noncrystalline polymer chains take on a “random walk” configuration. An analogy is a metal-linked chain. When tossed on the ground, the chain will take on a randomized shape, and each time it is tossed, a different shape results. (A linked chain would also adopt a randomized configuration, now in 3D, in a satellite or space station.) On the microscopic scale, thermal energy is very important, and individual chain segments are constantly being bombarded by each other and by solvent molecules, leading to randomized configurations. Because of incessant thermal bombardment, individual chain configurations are forever changing with time, and they can be characterized only statistically.

When chains are not cross-linked, this constant thermal motion leads to diffusional migration, and when sufficiently hydrophilic, the chains will dissolve in water. When cross-linked, the polymer chains cannot migrate, but they nevertheless are in a constant state of undulation. As will be explained later, these motions are important for permeation of drugs or proteins, since chains need to remove themselves from the diffusing molecule’s path. The larger the molecule, or the shorter the distance between cross-links, the less likely that a whole polymer chain can “get out of the way” by random motion.

FIGURE 14.12 “Forces” acting on a hydrogel. Swelling equilibrium occurs when these forces are in balance. For nonionic hydrogels, only the mixing and elasticity forces are present.

We are now in a position to describe, in general, the forces that determine the degree of swelling of a hydrogel. These forces are depicted in Figure 14.12. We consider a cross-linked polymer that is immersed in a uniform aqueous medium at a fixed temperature. Generally, water will attempt to enter any material due to its thermal motion, i.e., its kinetic energy. However, the degree to which it can do so is determined by the hydrophilicity or hydrophobicity of the polymer, which are governed by the polymer’s molecular structure. Polymers containing hydrogen bond donor (e.g., —OH and —COOH) and acceptor groups (—NH₂, —O—, and >C=O) or ionized groups are generally hydrophilic, while those lacking such groups (hydrocarbons, silicones) tend to be hydrophobic. The combination of water’s thermal motion and its degree of affinity with the polymer chains corresponds to the “mixing” force contribution to swelling.

As water enters the hydrogel, it stretches the chains, pushing the cross-links farther apart. As this happens, the randomly configured chains straighten. However, as we already noted, thermal energy counteracts straightening, and this leads to a retractive force. Similarly, if too much water is withdrawn from the hydrogel, the chains crumple below their most randomized configuration, and now the thermal forces resist this tendency. Thus, the polymer chains act as miniature “springs” that resist deformation. This force is called “polymer elasticity” since it is identical to the force underlying the elastic action of rubber. The strength of the elastic force in a hydrogel is proportional to the cross-link density, and hence it is inversely related to the polymer chain length between cross-links.

For many hydrogels, the balance between the mixing and polymer elasticity forces determines the degree of swelling. This balance contains both enthalpic and entropic components, so swelling can be temperature dependent. Molecules that adsorb to the polymer and alter its hydrophilic/hydrophobic balance (e.g., proteins, lipids) can also affect swelling.

When the hydrogel contains ionic groups, usually ionized acids or bases residing on the polymer backbone or on the side chains, a third swelling force is present. These so-called fixed charges on the polymer network are balanced electrically by mobile counterions, and the total concentration of mobile ions inside the hydrogel is greater than the concentration of mobile ions outside. This distribution is analogous to the Donnan equilibrium that results when charged colloidal solutions are separated from colloid-free solutions by a semipermeable membrane. Due to the relative abundance of mobile ions inside the hydrogel, an ion osmotic force drives water into the network. Swelling equilibrium is now determined by the balance of the mixing, elastic, and ion osmotic forces. The ion osmotic force increases with the fixed charge density inside the hydrogel and decreases with increasing ionic strength of the external solution. It may also depend on the concentrations of specific ions in the external solution, which partition into the hydrogel according to their charge and their affinity for water. For example, calcium ions (valence +2) are more effective at reducing swelling of negatively charged hydrogels than sodium ions (valence +1). At a fixed valence, ions that are considered “structure breaking” tend to reduce swelling compared to “structure-making” ions.

14.4.2 STIMULUS-SENSITIVE HYDROGELS FOR DRUG RELEASE

Each of the three forces governing hydrogel swelling is susceptible to external stimuli, provided the right chemistry is built into the hydrogel. By taking advantage of this fact, hydrogels can be designed to release their contents in an environmentally sensitive manner, either by altering the drug’s diffusion coefficient in the hydrogel network or by changing the drug’s affinity to the hydrogel’s chains.

Hydrogels that modulate drug release as a function of specific stimuli can work in open or closed loop. Open-loop systems control drug release by responding to an externally applied stimulus such as ultrasound, electric or magnetic fields, or light irradiation, as illustrated in Figure 14.3. Release rate is not critically determined by conditions of the biological environment. In contrast, closed-loop or self-regulated systems detect changes in the physiological environment and release drug in order to restore the body’s function to a desired set point. Thus, closed-loop systems can be viewed as artificial means for enhancing homeostasis in the body. Perhaps the most familiar example of a closed-loop system is an insulin delivery device that senses increases in blood glucose levels and responds by releasing insulin, leading to a recovery of normal glucose levels.

Imprinted hydrogels, formed by coagulating monomers around the template when the polymer is in the unswollen state, provide another example of stimulus sensitivity. If the imprinted hydrogel swells due to a stimulus, the monomers forming part of the receptors separate from each other and the attraction for the drug is lost. Drug is then released.

Although sustained release of the drug often leads to adequate therapeutic responses, discontinuous release as a function of specific signals may be desirable in many situations. Regarding systemic administration, the efficacy of a number of treatments has been shown to improve if the release occurs in phase with certain biorhythms; such drugs include reproductive hormones, growth hormone, gastric acid inhibitors, β-blockers, antiarrhythmics, and antiasthmatics. Since many of these molecules are associated with conditions that require repeated and often chronic administration, implantable hydrogel networks that can switch release on and off at the adequate time of the day would be beneficial, provided that the drugs are sufficiently potent that only small amounts need to be implanted.

In the following subsections, we discuss how the principles that were introduced regarding control of hydrogel swelling can be applied to modulate drug release through and out of hydrogels, in either open or closed loop. We pay particular attention to means by which release can be modified according to external conditions, such as temperature, and the chemical or biochemical environment.

14.4.2.1 Temperature-Sensitive Systems

In recent decades, there has been considerable interest in strongly thermosensitive hydrogels, which alter their swelling properties over a narrow range of temperatures. Thermosensitive-swelling controlled-release systems can be activated when exposed to body temperature (37°C), which is higher than typical ambient temperatures (20°C–25°C). As will be discussed later, temperature can also be controlled locally by directed delivery of different forms of energy to the hydrogel, such as ultrasound or electromagnetic fields.

To date, the most widely investigated thermosensitive hydrogels are those that contain the monomer N-isopropyl acrylamide (NIPAm). These hydrogels, whose structure is illustrated in Figure 14.13, typically shrink between room and body temperature. As discussed previously, shrinkage is due to the release of bound “clathrate” water surrounding the NIPAm side chains, followed by attraction of NIPAm groups to each other. The “critical” temperature, T_c, associated with shrinkage of pure poly(NIPAm) networks, is about 33°C. This temperature can be raised by coincorporating more hydrophilic monomers such as acrylamide or lowered by adding more hydrophobic monomers such as butyl methacrylate.

FIGURE 14.13 Structure- and temperature-dependent swelling of poly(N-isopropyl acrylamide) hydrogels. A sharp transition in swelling occurs at a “critical” temperature, T_c.

Recently, hydrogels containing oligo(ethylene glycol) (OEG) side chains have also been shown to have favorable thermoshrinking properties. Because they are closely related to PEG, a biocompatible polymeric material, some researchers believe that these OEG-based hydrogels will be more readily accepted than NIPAm-based hydrogels.

While shrinkage of a hydrogel might normally be expected to slow down drug release due to tightening of the hydrogel mesh (Figure 14.13), the opposite trend is sometimes observed. Increased drug release upon shrinkage has been attributed to a squeezing effect, in which drug is carried out of the hydrogel along with the water that is squeezed out, i.e., by convection.

14.4.2.2 pH-Sensitive Systems

When the ionizable groups in the hydrogel are weak acids or bases, then swelling is controlled by external pH, as depicted in Figure 14.14. In polyacid hydrogels, an increase in pH over a range near the pKa of the pendant acid groups leads to an increase in fixed charge density and, hence, swelling. Common polyacid hydrogels contain acrylic acid and its alkyl derivatives, or monomers containing sulfonated side chains. In polybase hydrogels, fixed charge density and swelling increase with decreasing pH near the pKa of the pendant base groups. Such hydrogels are usually based on imines or tertiary amines. (Quaternary amines have a positive charge regardless of pH.) It is also possible to form mixed polyacid/polybase hydrogels that exhibit increased swelling at the extremes of pH but are relatively shrunken at intermediate pH values. Shrinking is enhanced by electrostatic attractions between the oppositely charged acid and base groups; this is the so-called polyampholyte effect. Polyampholyte hydrogels can also be formed using monomers containing zwitterionic side chains such as sulfobetaines. Interestingly, the shrinking behavior of polyampholyte hydrogels is weakened and the hydrogels swell when ionic strength increases, the opposite of what occurs when the hydrogel is negatively charged.

Thus, pH can affect hydrogel’s hydrophilicity by ionizing acid or base groups. Considering these factors, initially glassy systems have been studied for oral delivery of drugs in which controlled release is desired to a specific part of the GI tract. Depending on whether the polymer contains basic or acidic side chains, it can be programmed to swell and release its contents when it hits either the acidic environment of the stomach or the alkaline environment of the small intestine (see also Chapter 7, Figure 7.7). Hydrogels that swell and release their contents in the alkaline gut region are reminiscent of enteric coatings, which remain intact in the stomach but dissolve in the small intestine, thereby preventing drug release in the acidic and enzyme-rich gastric fluids.

FIGURE 14.14 Structure- and pH-dependent swelling of (a) polyacid, (b) polybase, and (c) polyampholyte hydrogels. Swelling/shrinking transitions occur near the pKa values for the acidic (—COOH: pKa₁) and basic (—NH₂: pKa₂) groups.

14.4.2.3 Glucose-Sensitive Systems

While temperature and pH provide relatively simple means to control hydrogel, there are many examples in which the decision regarding release rate should be determined by the level of a more specific biomarker, providing closed-loop control. Here we discuss glucose-sensitive swelling of hydrogels, which have been thoroughly studied due to the interest in using them to treat diabetes. We will then turn to more general biomarker-responsive systems that are under investigation.

Hydrogels containing phenylboronic acid (PBA) side chains alter their ionization and swelling in the presence of molecules containing cis-diol groups, most notably sugars. Attempts have been made to use such hydrogels to provide glucose-triggered release of insulin. The mechanism of sensitivity of PBAs to glucose is illustrated in Figure 14.15a. In the absence of glucose, PBA acts as a Lewis acid. At low pH values (low OH^- concentrations), the boron atom in PBA is coordinated in an uncharged, trigonal configuration, with two pendant hydroxyl groups and an empty orbital. As pH increases, the increasingly abundant OH^- ions fill the empty orbital, and the boron atom is now tetragonally coordinated, with three pendant hydroxyls and a net negative charge. The pKa of this Lewis acid–base reaction is controlled by the electron density distribution of the phenyl group attached to the boron atom.

For a typical monomer, methacrylamidophenylboronic acid (MPBA), pKa ≈ 8.5, the boron group is mostly uncharged. Remarkably, however, the tetragonal boronate is able to undergo a bidentate condensation reaction with cis-diol groups on the glucose molecule. This reaction stabilizes the charged boron atom. Hence, increasing glucose concentrations effectively lowers the pKa of MPBA, leading to a higher fixed charge density and greater swelling in hydrogels containing PBA side chains. Both the Lewis acid–base and the di-condensation reactions, while forming covalent bonds, are completely reversible, and when glucose is withdrawn, the hydrogel shrinks back to its original size.

FIGURE 14.15 (a) Mechanism by which sugars such as glucose can affect the ionization of phenylboronic acids (PBA). The uncharged trigonal form (1) reacts with OH⁻ to form a charged tetragonal form and (2) is further stabilized by a cis-diol on the sugar molecule, through reversible bidentate condensation. (b) A complex formed between chains containing –PBA and polyvinyl alcohol forms cross-links, which break up in the presence of free sugar molecules, causing the hydrogel to swell. (c) When the pKa of the –PBA moiety is well below physiological pH, then almost all –PBA groups are charged and will bind sugars. Glucose, having two cis-diols, will bind to two chains containing –PBA groups, cross-linking them. At high glucose concentrations, cross-linking is reduced, however, due to occupation of each –PBA group by a glucose molecule (not shown). (Reprinted from J. Control. Rel., 190, Siegel, R.A., Stimuli sensitive polymers and self regulated drug delivery systems: A very partial review, 337–351, copyright 2014, with permission from Elsevier.)

There are alternative ways that glucose can affect the swelling of PBA-containing hydrogels. For example, hydrogels have been synthesized with both PBA-containing chains and PVA chains. Closely spaced –OH side groups on PVA can arrange to form bidentate condensation complexes with PBA, cross-linking the two chains together, as depicted in Figure 14.15b. Glucose introduced into such hydrogels competes with the PVA for tetragonal PBA, breaking some of these cross-links and reducing the elastic retractive force, leading to increased swelling. Again, these reactions are reversible, and swelling in the presence of glucose is reversed upon withdrawal of glucose.

When screening hydrogels for sensitivity to molecules such as glucose, specificity against other similar molecules or conditions that might interfere or produce unwanted responses must be checked. For example, the Lewis acid–base reaction that is involved in the swelling of PBA hydrogels (Figure 14.15a) is affected by changes in blood pH from its normal physiological value, pH 7.4. Acidosis, a common symptom of diabetes, will lead to a reduced hydrogel swelling at a given glucose concentration. Also, other sugars such as fructose and nonsugar diols such as lactate bind to PBAs. Fortunately, these interfering species are at relatively low concentration compared to glucose. Nevertheless, efforts have been made using molecular design of PBA derivatives to minimize both pH sensitivity and interference from molecules other than glucose.

When the PBA side chain is derivatized such that its pKa lies below physiologic pH, then binding by OH^- to –PBA is essentially complete and independent of pH. The binding propensity of cis-diols is maximal in this case. Interestingly, glucose has two diols that can bind and cross-link PBAs residing on different polymer chains, as illustrated in Figure 14.15c. At low glucose concentrations, this leads to shrinkage of hydrogels with increasing glucose concentration. However, hydrogel volume eventually reaches a minimum and starts to increase again as glucose concentration continues to increase, due to “flooding” of PBAs by glucose. Interestingly, this cross-linking mechanism seems to be specific to glucose.

An alternative method to confer glucose sensitivity is to immobilize the enzymes glucose oxidase, catalase, and gluconolactonase inside a pH-sensitive hydrogel. Together, these enzymes catalyze the net reaction: glucose + O₂ → gluconic acid + ½O₂. The gluconic acid dissociates readily, releasing H⁺ and lowering pH inside the hydrogel, thus affecting the hydrogel’s fixed charge density and swelling and, hence, its permeability to insulin. Since glucose oxidase is specific to glucose, interference from other sugars is minimal.

14.4.2.4 Other Biomarkers

Are there methods to provide high binding specificity for other kinds of molecules, with possible control of swelling? Recall that the body generates antibodies to specific molecules or antigens. Suppose now that one constructs a hydrogel with some side chains containing an antigen and others containing the corresponding antibody. Then, we expect that the tethered antigens and antibodies will form reversible cross-linking complexes (Figure 14.16). Suppose now that the level of free antigen fluctuates in the blood. When free antigen enters the hydrogel, it will compete with the tethered antigen for the tethered antibody and break the cross-linking complex, causing the hydrogel to swell. Again, this swelling will be reversed when antigen is withdrawn from the hydrogel as its blood level falls, since the intra-hydrogel complexes will reform. This strategy has been pursued—in fact, it is not necessary to include the whole antibody—only the Fab chain containing the antigen recognition site is needed.

Thus, it is possible to incorporate into the hydrogel recognition elements (e.g., antibodies or Fab fragments) in such a way that the hydrogel undergoes changes in volume in response to a target biomolecule, because their cross-linking density changes by formation or dissociation of the biomolecular complexes. Nowadays, it is possible to prepare antibodies for many biomarkers, and reversible antigen-sensitive hydrogels are promising biomaterials for constructing self-regulated drug-release systems. In the absence of the antigen of interest (i.e., the biomarker), the hydrogel does not release the drug because of the high cross-linking density (thus, low mesh size) of the network due to the internal antigen–antibody interactions. Drug release starts when the antigen biomarker appears in the surrounding medium.

FIGURE 14.16 Antigen-responsive reversible swelling of hydrogels prepared by copolymerization of complementary antigen and antibody monomers that can act as reversible cross-links. Interaction of free antigen with the immobilized antibodies leads to the rupture of the initial cross-links, which cause the hydrogel to swell and thus trigger drug release. If the free antigen vanishes, the cross-links can be reformed and the release stops.

Other hydrogels modeled on biological recognition have been studied. Monomers such as acrylamide can be functionalized with single-strand DNA (ssDNA) oligonucleotides and their complements. Upon polymerization, hybrid pairs are formed, which act as cross-links. These hybrids are broken, reversibly, upon exposure to a free ssDNA containing one of the complementary sequences. DNA aptamers to a specific target molecule can also be generated, for example, using the SELEX process, starting with a library of ssDNAs to many target molecules. Connecting the target and the aptamer to the hydrogel, cross-links will form that will be broken, again reversibly, when exposed to the free target molecule.

14.4.2.5 Templated Hydrogels

Since antibodies and oligonucleotides have limited physical and chemical stability and may lead to allergenic reactions, the design of artificial receptors is gaining increasing interest. In molecular-imprinting technology, the spatial distribution of the monomers is optimized to achieve the maximum efficiency of interaction between the substance of interest and the polymer network. This technology was born in the context of analytical chemistry, with the goal of preparing synthetic materials with receptors able to specifically bind and separate analytes. Thus, tailor-made receptors are created by synthesizing the polymer network in the presence of the target substance, which acts as a template, as depicted in Figure 14.17. Incorporation of the target in the monomer solution, which may contain several comonomers, is expected to cause the spatial arrangement of the monomers as a function of the strength of their interactions. Monomers with high affinity are drawn close to the template molecules, while those with less affinity tend to be excluded from contact. Different monomers can also “cooperate” due to their affinity to different parts of the target molecule. The arrangement of the monomers is “frozen in” during polymerization and cross-linking.

Once the molecularly imprinted network has been formed, the template molecules are removed, revealing imprinted cavities, sometimes called receptors, in the polymer network. These cavities are complementary in size, shape, and functional groups to those of the target molecules. If the polymer network again enters into contact with a solution containing the target molecules, then the latter will be absorbed and bind more strongly to the network than other, indifferent molecules.

The success of molecular imprinting relies on the strength of interactions between the template molecules and the monomers responsible for creating the imprinted cavities, which are called functional monomers. These interactions should occur favorably throughout the polymerization process and during the rebinding step. Ideally, all receptors would be uniform and optimally configured around the template molecule. However, this can only occur if the template is covalently bound to the functional monomers during polymerization, with the covalent bonds broken after polymerization. Unfortunately, this approach requires difficult or sometimes impossible chemical derivatizations.

FIGURE 14.17 Steps of the preparation of a molecularly imprinted network. Drug molecules can act as templates to drive the arrangement of functional monomers.

Hence, the most common approach to molecular imprinting is noncovalent self-assembly, depicted in Figure 14.17. Multipoint interactions between a template molecule and various functional monomers are required to form strong complexes, and the strengths of the monomer–monomer, template–template, solvent–monomer, and solvent–template interactions during polymerization determine the extent and the quality of the imprinted cavities, which are not identical to each other. If the molar ratios between components in the complex are not appropriate or if the assemblies dissociate to some extent during polymerization, then the functional monomers will not correctly arrange around the template molecule, resulting in a small difference between imprinted and non-imprinted (conventional) networks. Thus, the simple presence of target molecules in the monomer solution does not ensure an imprinting effect if the interactions are not favorable. The interactions will determine the number of receptors that can be formed and their heterogeneity.

Traditionally, imprinted networks have been prepared with a high proportion of cross-linking agent (>50% mol/mol) in order to obtain rigid structures with very limited swelling. This rigidity confers physical stability to the receptors. For a variety of applications, such as isolation or analysis of target substances, imprinted systems are prepared as nondeformable particles or rigid plastics or surfaces, using a solvent that acts as porogen to facilitate the removal and rebinding of the target molecules. Such an imprinted receptor resembles a permanently engraved hole. The adaptation of the molecular-imprinting principles to loosely cross-linked networks (as those that are responsive to stimuli) is quite challenging since the receptor structure must be “memorized” in order to maintain the recognition ability after several swelling/collapse cycles. The undulations of polymer chains tend to reduce the probability that monomers attached to them will remain in the optimal receptor configuration. At particular degrees of swelling, the receptors should maintain some memory of their structure and ability to bind the target molecule, while, at other degrees of swelling, this memory should be reduced, as will be the affinity for the target molecule.

Thus, when the imprinted cavities are located in a stimulus-sensitive hydrogel, the conformation of the receptors can be deformed and reconstituted as a function of an external or a physiological signal. Stimulus-responsive imprinted hydrogels are particularly suitable for developing advanced intelligent drug delivery systems able to (1) selectively and effectively load a certain drug, (2) release the drug at a rate modulated by a stimulus, and (3) uptake again the released drug from the environment if the drug remains around the hydrogel when the stimulus stops or diminishes its intensity, at which point the cavities are reformed.

14.4.2.6 External Stimulus-Sensitive Systems

For some pathophysiological processes, there is no biomarker that can be readily used to alter the drug-release properties of a hydrogel. To overcome this limitation, hydrogels can be endowed with responsiveness to external stimuli, such as light, electric or magnetic fields, and ultrasound, which can be applied to the target area when the hydrogel is either physical placed there, or accumulates there following injection as a nanocarrier. Externally triggered hydrogels switch the release on/off as a function of the intensity and the time that the external stimulus is applied. These latter systems have the advantage of providing on-demand (open loop) regulation of the release process, which can be adapted to the progress of the illness and the patient conditions.

Ultraviolet (UV), visible, and near-infrared (NIR) radiation sources can be used to trigger drug release at a specific site of the body, offering a very precise control of the release site. Light responsiveness is most commonly provided by photoactive groups such as azobenzene, cynnamonyl, spirobenzopyran, or triphenylmethane, which undergo reversible structural changes under UV–visible light. However, UV light can only trigger release from hydrogels placed on the skin or against mucosa, since radiation of wavelengths below 600 nm does not penetrate more than 1 cm into the body. NIR light (650–900 nm) can penetrate deeper because hemoglobin (the principal absorber of visible light), and water and lipids (the main absorbers of infrared light), have their lowest absorption coefficient in the NIR region.

FIGURE 14.18 By incorporating metal nanoparticles into microgels or nanogels that are temperature sensitive, delivery localized in time and space can be achieved by application of electromagnetic radiation.

Metals (e.g., gold nanoparticles) can absorb NIR light and efficiently transform the radiant energy into local heating, increasing the temperature at the surroundings several degrees above body temperature. This phenomenon can be exploited to trigger drug release from temperature-sensitive networks, and to improve the efficiency of tumor treatments by means of thermal ablation, as depicted in Figure 14.18 (see also Chapter 18, Section 18.4.1). Tumor cells are less resilient than normal cells when exposed to elevated temperatures. Heating alone at 43°C helps to kill tumor cells, but this temperature might be reduced if anticancer drugs are simultaneously released.

It is also possible to incorporate superparamagnetic iron oxide nanoparticles (SPIONs) into nanohydrogels. Magnetic drug carriers containing temperature-responsive polymers possess three unique features: (1) SPIONs can be visualized by means of magnetic resonance imaging, (2) tissue distribution can be increased by placing an external magnet near the target site, and (3) drug release can be triggered when an alternating magnetic field is applied, locally, either due to squeezing of the hydrogel in a constant magnetic field or due to local heating in a radio frequency magnetic field. In a similar manner, localized heating of hydrogels followed by release of their contents can be triggered by focused ultrasound.

Electrical stimuli can be generated using commercially available equipment for transdermal delivery, which enables precise control of the intensity and amount of current, the duration of the pulses, and the intervals between successive pulses. Electrically sensitive networks are commonly made of polymers with a high density of ionizable groups (polyelectrolytes), similar to those used for preparing pH-responsive hydrogels. Electrically responsive networks can be used in the form of injectable drug-loaded microparticles or as subcutaneous implants. An electric field can be applied through an electroconducting patch placed on the skin over the polyelectrolyte network. The changes in pH that occur near the electrodes due to the movement of the protons to the cathode cause shrinking of the network and squeezing of the drug out of the hydrogel. The intensity of the electrical field and the time of application regulate drug-release rate and duration. When the electrical field is switched off, the hydrogel swells again. Therefore, alternating shrinking (release on) and swelling (release off) can be achieved by applying repetitive pulses of electricity.

An alternative to the polyelectrolytes is the use of intrinsically conducting polymers (ICPs), such as polypyrrole (PPy) or polyaniline, which possess the electrical, electronic, magnetic, and optical properties of a metal. Electrical conductivity is due to an uninterrupted and ordered π-conjugated backbone. ICPs are electrochemically formed as a continuous film on the surface of a working electrode. When a current is applied, the ICP undergoes reversible redox reactions that cause swelling of the polymer network and, as a consequence, can be useful to trigger drug release. ICPs have been tested in vivo as coatings of cochlear implants, which consist of an electrode array implanted into the scala tympani of the cochlea to electrically stimulate spiral ganglion neurons (SGNs), providing auditory signals to individuals with hearing loss. However, the continued electrical discharges may cause apoptosis of SGNs. To overcome this problem, the electrode array is coated with a layer of PPy containing therapeutic neurotrophins, which are able to prevent the loss of SGNs. The electrode array provides electrical stimulation and thus regulates the release of the neurotrophins to the SGNs.

FIGURE 14.19 Pressure-sensitive hydrogels can provide on-demand drug release through a mechanical stimulus generated intentionally by the patient. In the scheme, a hydrogel composed of a β-cyclodextrin derivative and alginate hosts the antiemetic drug ondansetron (ODN). The drug is released in response to mechanical compressions by changing the inclusion ability of the cyclodextrin moieties. (From Izawa, H., Kawakami, K., Sumita, M. et al., β-Cyclodextrin-crosslinked alginate gel for patient-controlled drug delivery systems: Regulation of host–guest interactions with mechanical stimuli, J. Mater. Chem. B, 1, 2155–2161, 2013. Reproduced by permission of The Royal Society of Chemistry.)

Hydrogels can be also designed to release drug in response to physical pressure. As explained before, cross-linked polymer chains behave as small springs. One can design the hydrogel with drug receptors (template hydrogels) or binding moieties (for example, cyclodextrins, which can form inclusion complexes with the drug) that exhibit binding affinity for the drug when the network is at rest, but when a mechanical pressure deforms the hydrogel, the receptors distort and the drug is released. This principle is being implemented to produce transdermal hydrogel–based formulations with drug release controlled by mild mechanical compression exerted by the patient’s hand (Figure 14.19).

14.4.2.7 Multiresponsive Systems

Stimulus-responsive systems described previously rely on polymers that undergo reversible phase transitions when activated by the stimulus. An alternative approach is the utilization of dynamic covalent bonds, i.e., covalent chemical bonds that can be formed and broken under equilibrium control. These events can be made sensitive to changes in physiological parameters, such as pH or redox conditions. For example, dual pH- and redox-responsive nanogels have been prepared by cross-linking polymer chains through both imine and disulfide bonds. If only one stimulus (low pH or the presence of reducing agent) acts on the hydrogel, only one type of bond is broken and a minor increase in drug-release rate occurs because there is still adequate cross-linking. When both stimuli are present, the hydrogel, which is in a particle form (see the next section), disintegrates.

14.5 RECENT DEVELOPMENTS

14.5.1 MlCROGELS AND NANOGELS

Thus far, there has been little discussion of the size of hydrogel-based drug delivery systems. While millimeter- or centimeter-sized systems can function as membranes or monoliths, hydrogels with smaller dimensions have their advantages. In the following discussion, we include both already swollen hydrogels, and networks that swell into hydrogels upon exposure to body fluids.

Microgels, of size between one and several hundreds of microns, are typically formed by suspension polymerization, wherein small droplets of water containing the monomers are suspended in a continuous oil phase, the suspension being stabilized by stirring and adding surfactants. Particle size is controlled by stirring conditions, with increasing stirring rate generally leading to a decreased particle size. Microgels created by this method tend to be polydisperse, and the particle size range may need to be trimmed by filtration. Microgels are small enough to be injected subcutaneously or intramuscularly through hypodermic needles, and they can establish a drug depot at the site of injection. They may be subject to a localized foreign body response, which can sometimes be attenuated by surface modification of the particles. Microgel depots should be programmed to disintegrate after their release function has been fulfilled.

Nanogels, which typically are of diameter less than 300 nm, can either be synthesized by polymerization and coagulation in solvents with low monomer concentrations or in microemulsions, in which water droplets containing the monomers are smaller than 1 micron. Nanogels are appropriate as circulating depots, as they can access the full circulatory system, including capillary beds. Even smaller nanogels may show enhanced extravasation into tumors, due to the fenestrated, leaky anatomy of tumor vasculature. This means localized release into tumors does not require direct injection, and it might be useful for targeting otherwise difficult-to-access primary solid tumors and their metastases. By incorporating pH sensitivity, nanogels may be able to withhold drug release until they are localized in the acidic environment that is typical of many solid tumors. By functionalizing their surfaces with tumor cell–specific ligands, nanogels may also be taken up specifically by those cells. By proper chemistry, they might be programmed to release their cargo (drug) selectively in the low-pH, high-glutathione environment of tumor cells. Uptake of nanogels by cells prior to drug release may be especially beneficial in tumors that are drug resistant due to the presence of membrane efflux pumps. Similar strategies may apply in using environmentally sensitive nanogels to deliver drug selectively to infected or inflamed tissues that have altered high temperature, acidic or alkaline pH, high ROS levels, or high glutathione levels.

Nanogels of poly(N-isopropylacrylamide-co-acrylic acid) bearing doxorubicin conjugated through a pH-labile bond exhibit dual temperature/pH-dependent cellular uptake and cytotoxicity. These nanogels shrink when temperature rises from 37°C to 43°C favoring cellular internalization, while the pH responsiveness enables a fine control of drug-release rate. Thus, localized thermal stimulation of a tumor may improve drug delivery in a site-specific manner.

Dually pH- and temperature-responsive hydrogels that combine NIPAm and ionizable components can adjust the release of thrombolytic agents such as heparin or streptokinase to the small changes of pH and temperature that accompany the formation of the thrombi. They can also be prepared to coat the vaginal tissue and to release an anti-HIV microbicide in response to semeninduced pH (see also Chapter 12, Section 12.7.6).

Nanogels are susceptible to opsonization and removal by the reticuloendothelial system, and surface modifications may be needed to minimize these effects. As with microgels, attention must be paid to their clearance and potential toxicities.

14.5.2 MICROFLUIDIC CHIPS

We conclude this section by discussing alternative ways by which the swelling and shrinking of hydrogels can be used to modulate drug delivery, focusing on microfabricated systems. In these systems, methods that were originally developed for electronic microchips and integrated circuits are used to create small solid structures that respond to their environment. Because of their origin in the electronics industry, these systems are often called “microelectromechanical systems” or MEMS. While initial work in this field was carried out with silicon, it has been expanded to include many other hard and soft materials, such as metals, glass, plastics, and silicone rubber, the latter of which is also referred to as polydimethylsiloxane or PDMS. In fact, micromachined silicon masters can be used to form microstructures in PDMS by molding. Microfluidic systems, related to MEMS, include small channels through which liquids, including drug solutions or the blood, can flow. Integrating microfluidics with microsensors, one obtains microscopic total analytical systems or lab-on-a-chip.

How can hydrogels be integrated into such microsystems? One example is to totally confine a thin hydrogel film between a solid but porous membrane, which allows small solutes to pass through it, and a slightly deformable diaphragm. When the hydrogel is exposed to external conditions (pH, glucose concentration, etc.) that would normally cause it to swell, it will exert a swelling pressure on the diaphragm, causing the diaphragm to distort slightly. By coupling this distortion to either a piezoresistive or capacitive microsensor, the concentration of the target analyte can be sensed. Alternatively, if the hydrogel film is bonded on one side of a microinductor but is otherwise free to swell and if SPIONs are incorporated inside the film, then swelling and, hence, analyte concentration is reported by a change in inductance, which can be monitored continuously at radio frequencies. These techniques can be used to aid in decisions regarding drug dosing, e.g., for insulin.

In an even more advanced scenario, the hydrogel controls the flow of a drug solution in a microfluidic valve. Suppose the hydrogel is pinned in a microchannel. In its swollen state, the hydrogel blocks fluid flow, but when it shrinks, fluid flow is permitted. In one proposed system, a thermoshrinking hydrogel containing gold nanoshells is synthesized in a microchannel. When exposed to infrared radiation, the nanoshell heats up, causing the hydrogel to shrink and permit flow of drug solution. Attempts have also been made to stop and start the flow of insulin solutions in microchannels using the swelling and shrinking of the glucose-sensitive hydrogels. However, the latter system may be problematic in practice, since its response is “hard wired,” while flexibility is needed in insulin dosing between patients and within patients, due to circadian and other variabilities in response to insulin.

14.6 CONCLUSIONS

Hydrogels are soft materials that can function either as membranes controlling diffusion of drugs or as monoliths that both store drug and control its rate of release. They can be loaded with a variety of therapeutic substances, which makes them appealing platforms for drug delivery. As controlled-release systems that provide a variety of release rates, they can be produced in various sizes, depending on the desired application.

For hydrogels that do not suffer significant degradation or changes in volume after administration, drug release occurs via diffusion through the polymer network. If the network experiences swelling due to hydration or the action of certain stimuli, modifies its affinity for the drug, or undergoes hydrolytic or enzymatic degradation, then these processes also influence the rate of drug release. Figure 14.20 summarizes these processes.

For a given drug, the diffusion time decreases as the water content of the hydrogel increases, and for a given hydrogel, the lower the molecular weight of the drug, the shorter the release time. Thus, drug release can be regulated by tuning the swelling degree of the hydrogel, which, in turn, regulates the mesh size and the volume of water available for movement of the drug. A shrunken or a glassy network can act as a cage and is able to completely prevent drug release. When water enters the network and swells it, the chains loosen up, enabling drug release.

FIGURE 14.20 Release mechanisms of a drug, with hydrodynamic diameter d_h, from a hydrogel with mesh size ξ that can remain constant or that can be modified as a function of the water uptake, the action of a stimulus, or an erosion process due to hydrolytic or enzymatic degradation.

This chapter has provided a review of some, but not all, of the possibilities involved in combining drugs and hydrogels. While we have cast a rather wide net, it is certain that new ideas and applications will arise as hydrogel chemistry and materials science advance. It should be recognized that, however, like any biomaterial, hydrogels must pass rigorous efficacy, biocompatibility, and toxicity tests in their development and approval. These conditions become more rigorous as the complexity of the systems increases, which may include a hydrogel, a drug, and auxiliary materials (e.g., gold on iron nanoparticles). Currently, there are already a number of stimulus-responsive products under clinical evaluation, and some of them are based on hydrogels, mainly as monoliths for controlled release of active substances on the skin during wound healing or as nanogels suitable for diabetes and cancer treatment. The information generated in the design and evaluation of these pioneering products should pave the way to the development and approval of more stimulus-responsive hydrogels.

FURTHER READING

Alvarez-Lorenzo, C. and A. Concheiro. 2014. Smart drug delivery systems: From fundamentals to the clinic. Chemical Communications 50:7743–7765.

Annabi, N., A. Tamayol, J.A. Uquillas et al. 2014. 25th anniversary article: Rational design and applications of hydrogels in regenerative medicine. Advanced Materials 26:85–124.

Davis, K.A. and K.S. Anseth. 2002. Controlled release from crosslinked degradable networks. Critical Reviews in Therapeutic Drug Carrier Systems 19:385–423.

Eichenbaum, G.M., P.F. Kiser, A.V. Dobrynin et al. 1999. Investigation of the swelling response and loading of ionic microgels with drugs and proteins: The dependence on cross-link density. Macromolecules 32(15):4867–4878.

Izawa, H., K. Kawakami, M. Sumita et al. 2013. β-Cyclodextrin-crosslinked alginate gel for patient-controlled drug delivery systems: Regulation of host-guest interactions with mechanical stimuli. Journal of Materials Chemistry B 1:2155–2161.

Kabanov, A.V. and S.V. Vinogradov. 2009. Nanogels as pharmaceutical carriers: Finite networks of infinite capabilities. Angewandte Chemie–International Edition 48:5418–5429.

Kopecek, J. 2009. Hydrogels: From soft contact lenses and implants to self-assembled nanomaterials. Journal of Polymer Science Part A–Polymer Chemistry 47:5929–5946.

Omdian, H. and K. Park. 2012. Hydrogels. In Fundamentals and Applications of Controlled Release Drug Delivery, eds. J. Siepmann, R.A. Siegel, and M.J. Rathbone, pp. 75–105. New York: Springer.

Peppas, N.A. ed. 1986. Hydrogels in Medicine and Pharmacy, Vols. 1–3. Boca Raton, FL: CRC Press.

Peppas, N.A., J.Z. Hilt, A. Khademhosseini et al. 2006. Hydrogels in biology and medicine: From molecular principles to bionanotechnology. Advanced Materials 18:1345–1360.

Qiu, Y. and K. Park. 2012. Environment-sensitive hydrogels for drug delivery. Advanced Drug Delivery Reviews 64:49–60.

Siegel, R.A. 2014. Stimuli sensitive polymers and self regulated drug delivery systems: A very partial review. Journal of Controlled Release 190:337–351.

Siegel, R.A., Y.D. Gu, M. Lei et al. 2010. Hard and soft micro- and nanofabrication: An integrated approach to hydrogel-based biosensing and drug delivery. Journal of Controlled Release 141:303–313.