ISOPT Clinical Hot Topic Panel Discussion on Ocular Drug Delivery

Introduction

At the ISOPT 2018 meeting held in Tel Aviv, a panel on drug delivery “hot topics” was convened and moderated by Dr. Uday B. Kompella of University of Colorado Denver and Dr. Diane Tang-Liu of DTL Biopharma Consulting. The panelists made short presentations on key ocular drug delivery topics and then addressed questions from the moderators and audience members. The panelists included Drs. Abraham Domb from Hebrew University, Ashwath Jayagopal from Roche, Arto Urtti from University of Helsinki and University of Eastern Finland, and Clive G. Wilson from University of Strathclyde, Glasgow. The topics discussed included nanospheres and physical and chemical approaches to drug delivery, stimuli-responsive drug delivery systems, drug binding to the vitreous humor, and drug delivery and health care for the aging eye. Highlights of each of these and related topics from the meeting are summarized below. Also, questions and answers from this session are included at the end, with slight modifications for this publication.

Nanotechnology and Physical and Chemical Approaches for Drug Delivery

Drug delivery systems of various sizes, including those based on polymers, are useful in improving existing drugs. Enhanced activity and safety of existing drugs can be achieved by facilitating targeting and reducing unwanted systemic drug distribution.

For the eye, since it is a closed system with limited volume that can be administered, we must find ways to deliver drugs into the eye in an efficient way using the least volume, preferably noninvasively. Since bioavailability to the eye tissues is very low after noninvasive dosing, using nanotechnology and/or physical/chemical means are potential approaches to enhance drug delivery to the eye. Table 1 summarizes key nanomedicine, physical, and chemical approaches that can potentially be used to increase ocular delivery, after ensuring appropriate risk:benefit ratio. The drug in its molecular form in solution is in the nanoscale dimension. If a drug is insoluble, it can be solubilized using surfactants that form micelles of a nanosize. Furthermore, formulations can be engineered to be nanosuspensions, nanoemulsions, nanocomplexes, or nanoconjugates, among other preparations. ¹ These formulations can enhance delivery by improving drug stability, drug entry, and drug targeting or drug retention in eye tissues (Table 1). Restasis™ (cyclosporine emulsion), Inveltys™ (loteprednol etabonate), and Cequa (cyclosporine ophthalmic micellar solution) are 3 examples of ophthalmic products that are claimed to be nanomedicines.

Physical means to enhance delivery include iontophoresis, electroporation, ultrasound, light-activation or photodynamic therapy, and use of needles/microneedles, all of which technologies can potentially be used with drugs in solution or nanoparticles. While the last approach, by design, is invasive or minimally invasive to the eye, the first 4 approaches can be applied externally, in conjunction with drug application that is noninvasive or invasive. However, some of the former approaches may also perturb the membrane at the microscopic level.

One unique way proposed by Dr. Domb to increase the effectiveness of ocular drug delivery by noninvasive intervention is to combine the delivery of nanoparticles with low electrical current or iontophoresis. ² This technology is in clinical development for the delivery of steroids (EyeGate, www.eyegatepharma.com/technology/iontophoresis-delivery-system) and in clinical use for the delivery of anti-inflammatory agents through the skin for treating arthritis. ³ Another way of delivery to the eye would be electroporation, a process by which transient pores are formed in the outer epithelium using focused electric current. This approach damages the eye surface and is not yet considered ready for clinical assessment. While iontophoresis uses low electrical current over several minutes,^2,4 electroporation uses high currents for 1 s or much shorter durations. ⁵ While iontophoresis is designed to encourage passive translocation of the barriers, electroporation is applied to create temporary pores in the membrane to drive materials into and across membranes. A third possibility for noninvasive delivery of nanoparticles into the eye surface is ultrasound, which also temporarily changes the surface of the eye to allow drug penetration. Ultrasound is a source of energy input, which disrupts membranes through cavitation to allow delivery of materials. ⁶ Photodynamic therapy is another approach, wherein a light source is used to realize drug effects. Following intravenous dosing of nanosized liposomes, a laser light is used to activate the drug and achieve photocoagulation in a clinically approved photodynamic therapy. ⁷ More traditional physical methods of drug delivery include dosing using regular length, small gauge needles, or microneedles of a similar gauge but shorter length to dose medication to either surface or intraocular tissues. ⁸ Compared with regular size needles, microneedles are considered minimally invasive.

A chemical approach to improve penetration of a drug into the eye is to modify the drug by conjugating to a macromolecule or by derivatization with small molecular motif so that the properties of the drug are altered to allow better penetration through the eye surface into the interior. During the transfer of the modified drug, it degrades and releases the active pharmaceutical ingredient once the primary barrier is crossed. While macromolecule derivatives are referred to as drug conjugates, these small-molecule derivatives that revert to the parent drug are referred to as prodrugs. Another approach to improve a drug, particularly its safety, is to design a molecule that is not a prodrug but active and metabolized to inactive agents in the body to avoid systemic or other side effects associated with metabolites. For instance, loteprednol etabonate forms metabolites rapidly in the system, minimizing systemic side effects. ⁹ These agents are called soft drugs. Using a chemical approach, the nanoparticle surface can be modified to improve cell entry. For instance, ligands capable of entering cells through cell surface receptors can be covalently attached to nanoparticles to enhance delivery of nanoparticles. ¹⁰ These particles are referred to as functionalized nanoparticles. All the strategies mentioned so far can be considered when developing new formulations or delivery approaches for the eye.

The interest in improving delivery to the eye is not that high as compared with cancer and other types of diseases that are much more feared and of greater commercial value. Still, there is an opportunity here for nanomedicines, physical approaches, as well as chemical approaches, since drug delivery to the eye is extremely challenging and critical to maintain vision under various pathological conditions. ¹¹

Personalized Medicine and Stimuli-Responsive Delivery Systems

Many pharmaceutical companies and academic scientists are focusing on personalized medicine.^12,13 Personalized medicine involves tailoring the therapy to the patient. A modern avenue of research in this field is to develop stimuli-responsive drug delivery systems.^1,7

Stimuli-responsive therapeutics and drug delivery systems allow us better control of how, when, and where the therapy is delivered to the patient. ¹⁴ Table 2 summarizes a few stimuli-responsive drug delivery systems. There are many other examples, either in the ophthalmology field or in other fields, which are well described in several reviews. There are several stimuli that are relevant for the field of ophthalmology. These might be instigated in the clinic: a physician-triggered stimulus, with a drug formulation perhaps based on a hydrogel-type system. ¹⁵ An example is a drug-loaded, hydrophilic, interconnected polymer system. In this scenario, the physician can use a light-triggered stimulus to crosslink or disrupt the hydrogel network in several different configurations to deliver the therapy on a temporally controlled basis. Other stimuli might be disease specific, such as pH/redox changes or based on environmental conditions such as shear stress or temperature gradients (Table 2).

Photodynamic therapy, whereby a physician uses laser light to activate a liposomally entrapped drug (an example being Visudyne™; see product literature) in the ocular vessels was approved by the FDA in 2000. Visudyne indications include predominantly classic subfoveal choroidal neovascularization (CNV) due to age-related macular degeneration, pathological myopia or presumed ocular histoplasmosis. Liposomes allow drug targeting to the CNV lesion, with minimal quantities of drug in the surrounding tissue. The liposomes contain verteporfin, a photoenhancer drug. The liposome formulation is reconstituted and dosed as an intravenous infusion over 10 min. Verteporfin is activated by a physician using a nonthermal diode laser at 15 min after initiation of the infusion, over a period of 83 s. For CNV, the recommended light dose is 50 J/cm² at an intensity of 600 mW/cm². The activated drug occludes leaky blood vessels by the generation of short-lived free radicals, confining the action of the drug.

Other systems that may be engineered to respond to stimuli include molecularly imprinted drug delivery systems.^16,17 In this case, a solid material, most typically a polymer, is engineered with a number of cavities that bind to one or more different targets or analytes for either biosensing or drug delivery, or for ligand-stimulated drug release. For example, there could be a drug which has low affinity for the pocket relative to a disease-specific ligand, and when the ligand displaces the drug, we can achieve a type of disease-triggered drug delivery. In addition, drug release could be programmed in response to other biological changes. Using timolol maleate and molecularly imprinted contact lenses, it was shown that imprinted contact lenses provide 8.7-fold higher AUC in rabbit tear fluid when compared with an eye drop. ¹⁸ For ketotifen fumarate, the mean residence time of the drug in rabbit tears after topical application was 0.25, 3.36, and 12.60 h, respectively, for eye drop, nonimprinted contact lens delivery system, and molecularly imprinted contact lens delivery system. ¹⁹ In the future, contact lenses and other material forms can be engineered to exploit the potential for stimuli-responsive drug delivery. A key requirement for these advanced systems is a high predictability of dosing in response to stimuli among heterogeneous patient groups, to limit variable clinical response due to insufficient or hyperactive drug release in response to varying intensities of stimuli.

A third example is a disease-triggered therapeutic based on the molecular engineering of an antibody. ²⁰ In this case, a proantibody, similar to a prodrug is manufactured. The proantibody has no activity by itself but is activated by endogenous enzymes at the site of disease. An example of this technology is a proantibody designed to be activated by matrix metalloproteinases, as demonstrated for anti-EGFR and anti-TNFα-targeted antibodies. ²¹ In this system, the antigen-binding activity is restored in a disease-triggered manner. A possibility that has yet to be realized, is the design of antibodies that could be responsive to, for instance, hypoxia, inflammation, or reactive oxygen species, thereby providing a disease-triggered antibody-binding mechanism. All the examples quoted show the immense scope for stimuli-responsive therapeutics.

Drug Binding to the Vitreous Humor

In systemic drug delivery, plasma binding of drugs is very well known, very well characterized, modeled, and its impact on general pharmacokinetics is well understood. ²² In general, the higher the drug binding, the lesser the available drug for target tissues. Furthermore, one drug can displace another, if they both have an affinity for a binding protein in plasma. Additionally, when the drug concentrations vary, the extent of binding may also differ. In effect, the binding may be saturated for some drugs at the concentrations used. However, in the eye, regarding vitreal binding of drugs, there is very limited literature. Only a few studies have reported drug binding to vitreous, especially for antibiotics, which was determined to be low.^23,24

Dr. Arto Urtti presented data from his laboratory on drug binding to vitreous humor. ²⁵ Dr. Urtti's group selected some 30 or so ophthalmic clinical compounds and determined drug binding to porcine vitreous humor using equilibrium dialysis. Furthermore, drug binding to vitreous humor was compared with the binding of those compounds to plasma proteins. In general, it was evident that the vitreous binding is less than plasma binding. Also, the maximum binding or the range in the plasma is well known, ranging from 0% to 100%. This huge range at the higher end can have a profound impact on drug pharmacokinetics in systemic circulation. In the case of vitreous, the maximum binding was observed with acyclovir, which is only 74%. Other molecules exhibited binding of about 50% or less. Table 3 shows porcine vitreous binding of some drug molecules and their corresponding human plasma protein-binding values (Table 3; based on Rimpela et al.). ²⁵

Using mathematical simulations, the impact of vitreal binding on drug pharmacokinetics was further assessed by Dr. Urtti's group. Since the free drug fraction in the plasma is well known, the simulations were performed using these typical values.^25–27 At the extreme binding to vitreous humor of 74%, observed for acyclovir, the half-life in the vitreous would be about 3-fold higher, relative to the drug binding at 0%. This is a moderate change, given the need to prolong vitreal half-life for several days or months for many drugs and the short half-life of small-molecule drugs. This is particularly so, since vitreal injections are currently indicated at once a month or lower frequencies for approved drug products. The impact of vitreal binding on the half-life of most other drugs is expected to be far lower. To have a really high impact on the half-life prolongation, probably a drug should have a binding of 95% or higher. Thus, the impact of vitreal protein binding would be moderate for the known clinical drugs in ophthalmic use. However, there might be other compounds relevant for eye diseases, which could be binding at much higher levels than acyclovir.

Interestingly, the correlation between vitreous binding and plasma binding is poor, and strikingly, there are a few compounds like acyclovir that show higher binding to the vitreous humor as compared with the plasma (Table 3). The poor correlation indicates that the binding sites in the vitreous may be different from plasma, where albumin and alpha-1-acid glycoprotein are the major binding targets. Potential binding sites in the vitreous include collagen, hyaluronic acid, and glycosaminoglycans. Future studies should investigate these possibilities.

Drug Delivery for the Aging Eye

In preclinical research, scientists tend to work on juvenile animal eyes, which are very different from the clinical population that we would be treating. For example, in animal eyes the lens occupies much of the eye. ²⁸ This anatomic ratio for lens/vitreous dimensions is unlike that in man and it may result in differences in drug distribution. As an added complication, in the elderly eye, there are ongoing structural and biochemical changes which would also impact drug distribution.^29–31 The situation could be complex, given the fact that elderly patients receive multiple medications at the same time.

In the USA, the health care policy is ever changing to deal with escalating costs of care and medicines. ³² Likewise, there is an expectation that everyone has health insurance as per the Affordable Care Act. The insurance is purchased by the individual or through an employer. Additionally, for qualifying individuals, the option of Medicare exists in the USA as a single-tier system, where the government pays all the bills. In Europe, a 2-tier system is emerging, wherein tier one is government-provided health care and tier two is more expensive care that the rich can afford, although the actual health outcomes may be the same with either approach. As the wealth gap increases and the demographics change, health care provision and the way we develop and deliver medicines become complex. The socioeconomic changes pose unique challenges in supporting the vision care of the elderly, which will only get more expensive.

As a person reaches 60 s and 70 s, it is quite likely that a cataract operation and lens replacement is required. With different materials and designs used for artificial lenses, we are unaware of the impact on bystander aspects such as drug removal through or around these new lenses in patients fitted with an artificial lens.

In old age, the vitreous undergoes a loss of hyaluronate, and the collagen fibers collapse and eventually form a thicker fibrous structure, which remains attached at the back of the lens. This was elegantly demonstrated by Sebag in the late 1980's using side-lit illumination of mounted eyes in a Perspex cell. ³³ Following posterior vitreous detachment and extensive syneresis, the aged eye ends up divided into 2 zones, differing widely in their environment. This poses an important question–would it be of benefit in therapy to replace the liquified vitreous with an artificial vitreous humor? It is of consequence, since partial liquification can result in marked differences in drug distribution and clearance following intravitreal administration. ³⁴ In aphakic eyes devoid of lens, the vitreous humor is completely liquified and forward loss for drugs with low retinal permeability is greatly increased. ³⁵ Because the condition is progressive, partial liquification to complete liquification in aged eyes reflects different degrees of syneresis or mobility of vitreal contents, including dosage forms, resulting in poorly controlled differences in drug distribution and clearance. The way in which delivery systems must be adjusted for these aspects of aging have yet to evolve, else we risk having to employ formulations such as solutions, solids, or suspensions with suboptimal performance in the age group that we are trying to treat.

As a further consideration, elderly patients are very diverse with disparate comorbidities. Thus, randomness in responses as well as difficulty in recruiting large cohorts with similar pathological states is anticipated. It is often noted that studies in the elderly are hampered because they are already medicated. Enrolling diverse group of patients for clinical trials is probably essential to cater for this situation, and animal models are of limited assistance. Within the laboratory, the animal models of disease conditions are often monolithic, making clinical outcomes less than predictable. The knockout mouse may not really translate much beyond determining mechanisms of multiple genetic defects and environmental insults, as the morphology of the eye is so different to the human. In silico modeling might have a key role in crossing these boundaries.

The isolated elderly populations and a shortage of ophthalmic surgeons in some countries complicate vision care of the needy patients. Society will need to rethink provision of care as a pragmatic, cost-effective system that improves quality of life. Extension of treat and monitor remotely might be a new paradigm to consider in some populations.

There will be increase in the number of community-dwelling elders in our population. For example, in Japan in 2017, the number of elderly in the population reached a record 35 million, accounting for 27.7% of the population. ³⁶ Many countries have loci of high-density population, and then dispersed populations at other locations. Therefore, new approaches are needed to deal with these people in the community.

In ophthalmic medication management, compliance has always been a problem and the elderly may lack strength of coordination to use simple topical dosers. In northern India, while 50% of the patients may be noncompliant, 35% have improper administrative technique for antiglaucoma eye drops. ³⁷ Offering dosing aids to patients is of variable success, ³⁸ and anecdotally patients may not like washing dosing aids. Of course, if supplied in one-time dosing kits, the health care system will contribute to the scourge of undesirable plastic packaging.

So, to preserve sight, the health care system has to be thought about holistically, with a full cost–benefit analysis. Society and medicine providers need to address the issues implicit in drug delivery as well as medication management for the elderly to develop a better way to sustain their vision and quality of life, while minimizing the socioeconomic burdens.

Conclusions

Noninvasive ocular drug delivery would be ideal for treating back-of-the-eye diseases. However, due to multiple barriers, this is not currently feasible. Some potential approaches to enhance noninvasive topical drug delivery include nanomedicines and physical methods such as iontophoresis or a combination of physical enhancement and better-targeted medicines. For invasively administered back-of-the-eye medications, stimuli-responsive approaches can improve personalization of ophthalmic medicines. Stimuli, either external or internal to the sites of treatment, can allow controlled release in the vicinity of target pathology. An unexplored yet important area of research is drug binding to vitreous humor components. Drug binding in the vitreous, although different compared with plasma protein binding, extends half-life of some small-molecule drugs. Nearly all preclinical research and development efforts are focused on homogenous population of young animals, whereas ophthalmic drug therapies call for the treatment of heterogeneous, aging eyes with a more liquified vitreous. Future efforts are warranted for the design of delivery systems for the aging eyes in conjunction with better medication management for the elderly.

Transcript of Panel Discussion: Drug Delivery Hot Topics

The panel discussion was held in Tel Aviv on April 10th, 2018 at The International Symposium on Ocular Pharmacology and Therapeutics. The panelists included Abraham Domb, Arto Urtti, Ashwath Jayagopal, and Clive G. Wilson. Uday B. Kompella and Diane Tang-Liu moderated and chaired the panel discussion.

Uday B. Kompella: Okay, thank you, all the speakers. Now we have about 15 min or so for questions and answers. What I would like to do is take a couple of questions from the audience and go back to Diane, if she has any questions. And then I have a standard list of questions of my own. Any burning questions from the audience? Yeah? So, Diane, anything that you want to propose to the panel?

Diane Tang Liu: When we talk about smart delivery systems, would they be able to recognize any abnormality induced by aging or physiological changes? Knowing that such changes have quite a lot of variation with the population, how does a smart delivery system address this?

Ashwath Jayagopal: Yes, it remains to be seen how we can do that with a heterogeneous patient population. One thing that could drive better delivery systems will be better biomarkers for disease, which can then be translated into drug delivery systems that are patient tailored. I think we still need to stratify patients better by biomarkers, especially protein biomarkers.

Diane Tang Liu: Any questions from the audience?

Male Speaker: What are some examples of biomarkers? TGF-beta or VEGF? Or what are you thinking about to begin validating these?

Ashwath Jayagopal: Yes, the agents we are really interested in are hypoxia-triggered therapeutics to deliver antiangiogenic agents site specifically to ischemic sites in ischemic retinopathies. That would be a good test bed for this kind of application.

Uday B. Kompella: Okay, Avi, you talked about iontophoresis and ultrasound. What do you think are the safety concerns when you are repeatedly dosing subjects with these physical approaches?

Abraham Domb: Iontophoresis, or the application of current onto the eye surface influences the tissues. We have studied, and many other have studied the effect of current on the eye surface. There is an effect, depending on the influx of current that is applied. But in general, the effect disappears in 24 h.

So, this is not a risk, but it limits the time for applying the current and the amount of current that can be applied. Usually it is up to one milliampere for a few minutes, although there are reports on rat eye, using 20 milliamperes, which I assume might be toxic.

Regarding ultrasound, this is less of an issue as there are many medical devices that use ultrasound, which is considered less damaging. It forms temporary cavities in the tissue to allow drug penetration and usually repairs itself shortly thereafter.

Clive G. Wilson: After my limited experience I am cautious using ultrasound to convectively drive drug penetration. We completed some studies looking at a sonophoresis probe with the eye mounted on a video stage and illuminated from the side and we were able to see particles–probably melanin–being shaken out of the iris in animal models. We saw ejection as little plumes, and that was with quite moderate levels of power applied. This was an ex vivo animal study, conducted within 2 h of collection from the slaughterhouse. It was still fresh, with an intact cornea. So, I think there is a risk that sonophoresis will generate debris and in our case shake out melanin. I wouldn't want it done to me.

Female Speaker: There are some ultrasound techniques now that are used for glaucoma therapy. Now, do you suggest that at that strength it would shake the melanin out of the iris, as well? Is that something you would be concerned about?

Uday B. Kompella: Do you want to elaborate on what that device is?

Clive G. Wilson: The system was a laboratory sonophoresis probe, used in transdermal studies and generating ∼0.8 W/cm² at 55 kHz. The power would be sufficient to shake up the trabecular meshwork. That might usefully increase the drainage and the applied energy will displace debris but the iris melanin was freed. That's an unwanted complication whilst trying to address the problem of stiffness in an elderly eye.

Of course, to assess risk and cover the safety envelope, I have to use the technique in excess of the energy that I'm contemplating using. I did not anticipate that this damage might occur and if you shake tissue while trying to loosen the matrix, it is not unreasonable to expect that loosely attached particulate matter might be displaced.

Abraham Domb: It depends where energy is applied. If applying on the sclera, this might be less of a problem.

Clive G. Wilson: Yes, absolutely. We applied the probe to the cornea.

Diane Tang Liu: I'd like to ask about the drug-binding targets in the eye. I'd like to make 2 statements first before I go to my question. Number one is for systemic drugs. The plasma protein binding is mainly nonspecific reversible binding for acidic and basic drugs.

But in the eye, a lot of drug binding relates to drugs designed to bind to receptors such as VEGF receptor or PDGF receptor. But the populations of those binding targets are relatively small, when compared to the molecules in a dose. In that sense, the binding does not affect the ocular pharmacokinetics. So, can you give us some idea about the binding targets in the back of eye that could affect the drug pharmacokinetics or therapy?

Arto Urtti: So, to alter the pharmacokinetics, the target should be something that is present stably and at high concentration. For example, collagen, hyaluronic acid and melanin are such components. However, there are difficulties in getting compounds that would bind well to hyaluronic acid. In general, that is not happening with the small molecular drugs. However, a recent paper from Novartis in Nature Communications ³⁷ shows protein pharmaceutical that binds to the vitreous so that retention in the eye is extended.

Several drugs bind to melanin in the pigmented cells, and this is a viable option to modify pharmacokinetics in the eye. Drug needs to be potent enough and high binder.

Diane Tang Liu: Thank you.

Clive G. Wilson: There is another feature of melanin, the pheomelanin and eumelanin are these flat molecules, which probably shield molecules by pi bonding. An interesting question is what the purpose of the presence of melanocytes in the choroid is? It is accepted that there are changes in density past 60 years of age. So, why have you got melanin in your choroid? What function are the melanocytes providing? If these cells serve a key function, could you restore them? And leading from this, could you put a melanin-containing material into this space?

Diane Tang Liu: So, Clive, you talked about melanin binding for many years.

Clive G. Wilson: Yes, sure.

Diane Tang Liu: Okay, so do you think melanin binding is a hindrance to drug efficacy because it tightly bound the drug and the drug is not available to the rest of the eye? Or can it be therapeutically used in an advantageous manner?

Clive G. Wilson: I think that generally, because the IC₅₀ of most molecules that bind to melanin is too high, the targeting of a drug to melanin cannot be used for a practical purpose and in most cases of topical delivery merely attenuates efficacy, at least from intermittent dosing. Binding with continuous exposure is ultimately, saturable. Therefore, I think if you can drive melanin binding to saturation, you might be able to impact efficacy in a positive manner because more free drug would be available. Thinking further about using melanin as a depot, if the drug acted at a short range from a slowly unloading melanocyte then Arto's proposition might be worth exploiting.

Diane Tang Liu: I remember when we did radioactivity studies developing brimonidine. So, when we give the radiolabel in the Cynomolgus monkeys, the radioactivity stayed in the retina and choroid for over 90 days. But that's a different story, because we have got brimonidine given as an eyedrop. So whatever happened in the back of the eye with brimonidine really does not affect its effect in the front of eye. And the point is that melanin binding can be very strong, and in fact it can be very long-lasting.

Clive G. Wilson: Oh, it “jumps” eyes, as well. If you dose one pigmented eye with a drug that is highly melanin bound (for example in a rabbit) it appears in the other eye, because the dose drains down the nasal cavity. It appears from autoradiographs following dosage of a labeled drug to the cornea, the material can be seen in the sinus. The supplying ciliary artery comes up, and it picks up the drug and then supplies it to the other eye.

Arto Urtti: Yes, maybe I can comment on melanin binding. It is not the melanin binding alone that matters, because also the permeability of the compound in the cell membranes has influence on the accumulation in vivo. There is an equilibrium of the free drug and melanin-bound drug in the cell, and permeability affects the rate at which the drug eliminates through the cell membranes, thereby affecting the free cellular drug concentrations. So, the drug levels can be extended to a great deal if melanin binding is high and permeability is low or medium level.

But melanin binding can extend the retention and drug activity. There was a recent paper on pazopanib and one experimental compound, where they showed on purpose the prolongation of the anti-VEGF effect for several weeks. For atropine, prolongation of effect in anterior chamber was shown decades ago.

Impact on response depends on drug potency, like Clive was commenting. Free concentrations should stay above the IC₅₀. The same logic is also relevant in drug toxicity. Are the levels of free drug below or above the threshold levels for toxicity? Overall, the kinetics is complex in the pigmented cells.

Diane Tang Liu: Thank you.

Clive G. Wilson: There was another thing—is anybody here from Canada? No one from Toronto? If you'd have been in Toronto 20 years ago, and you'd gone downtown, and you'd taken a wrong turning into the rougher areas, you might have bumped into some heavy metal rockers. There was an extreme style statement, they had their eyes (sclera) tattooed with blue acrylic ink. That prompted a thought…. if these people are quite prepared to tattoo their sclera, could you use that as a way of drug delivery? Particularly using a drug binder–I suppose a thought was melanin binding and these guys probably wouldn't be averse to black sclera. Anyway, another possibility for loading up the sclera is to be able to, say, treat a stromal melt.

Diane Tang Liu: But you must have a balance between the strong melanin binding and ability to release the drug over IC₅₀.

Clive G. Wilson: Sure, absolutely.

Diane Tang Liu: So, let me be direct. What's keeping polymeric nanoparticles from being on the market?

Abraham Domb: Nanoparticles are delivery systems, usually made of degradable components that can be delivered into the eye with little effect on vision.

The issue of introducing a new drug delivery system into the clinic/market requires passing clinical trials, almost like a new drug entity, although less intensive and less risky, it is still very expensive. To justify the expense involved in the clinical development of a drug delivery system over an existing drug, the new formulation must be an order or magnitude better than the original drug.

The nanotechnology drug delivery systems should work and improve drug performance, however, the need should be defined or refined to specific applications possessing significant improvement.

Diane Tang Liu: Any questions from the audience?

Female Speaker: Clive, I have one more for you. So Arto showed very nicely the studies looking at different drugs and the vitreous binding and how that might affect the half-life. Now, with the liquefying vitreous, do you have any estimation of how that could affect the half-life of a drug for small-molecule versus antibody drugs?

Clive G. Wilson: That's a good question. There is no doubt that if you look at higher-molecular-weight dextrans injected as probes, they clear much, much faster in a model of liquefaction; for example where you've treated the rabbit with collagenase.

Another thing that changes is the distribution. In the liquefied eye, everything falls out and probably that might be one of the issues around the failure of anti-VEGF therapy in older patients. If the eye is liquefied, retention of the dose not going to happen unless we rethink the whole drug delivery vehicle problem.

Ashwath Jayagopal: Are you looking at it as pools of liquefaction or gross compartmental liquefaction? I believe it is more like pools that progressively grow in size.

Clive G. Wilson: Yes, I think that is a fair description.

Ashwath Jayagopal: And that would complicate the modeling of that outcome.

Clive G. Wilson: Exactly right. In aged vitreous specimens lacunae appear in the vitreous. Then eventually with a posterior vitreous detachment, as the membrane peels away from the back, you then get 2 compartments. Then the vitreous continues to shrink until you've got just the residue of that vitreous humor stuck to the back of the lens. You're fairly old by the time that happens (fortunately), but it is an issue.

I often wonder whether we are going to be dealing with a healthier elderly population with this condition… And of course policy will certainly dictate that we have to keep people functioning for longer.

Uday B. Kompella: Okay, Ash, what kind of nanoparticles do you think industry would be investing in, like the stimuli-responsive or topical nanoparticles, or anything else?

Ashwath Jayagopal: I do not know if it is too early to say, but I could say–industry may be looking for in an ideal formulation, one that could be recalled. A lot of these sustained-release formulations, if the patient has a bad reaction, it's hard to recall the drug. Another concern is drug to material erosion in a bioerodible system. If the drug has been long since released and the material is still remaining in the ocular components, that designates maybe some cleanup, and it may concern ophthalmologists to have biomaterials persisting in the eye for too long. So those are some of the things we're looking at.

And of course, ocular tolerability, looking at the placebo formulation and seeing how well it's tolerated. We're finding many materials very well tolerated intravenously, systemically, and subcutaneously, really perform very poorly in the eye for one reason or the other.

And the last one we're looking for is particle movement. Is the nanoparticle moving, clouding the visual axis? We have seen a lot of nanoparticles or metabolites of nanoparticles disturbing the visual axis and actually moving into the anterior chamber. So, we really want something that stays put.

Uday B. Kompella: So, are you saying that it's probably not a good idea to inject nanoparticles or microparticles in the vitreous, or are you open to that?

Ashwath Jayagopal: I think we have seen a lot of good examples of intravitreal injections that are working very well with the Ozurdex and the upcoming Envisia formulation. So, there are definitely great examples of how this can work. So, I'm optimistic.

Clive G. Wilson: Following injection, forward convective flow can cause the particulate to ride around the edge of the lens, and the risk is that the patient will experience light scatter from sidelighting of the particulate. So, a question that arises in development which we mull over is how many particles present in the anterior eye following posterior injection are a bad idea? I think that is a physiological phenomenon that we have to be concerned about.

Diane Tang Liu: But how do you weigh in on the benefit-versus-risk issue? Because a patient's facing a vision-threatening event,–how do you balance vision restoration versus compromising the function in daily life, such as driving?

Clive G. Wilson: If the problem is flare in a particular situation–for example, driving–would a holistic approach be to avoid side illumination from scattering of particulates at the front of the eye by the arm of spectacles blocking headlight beam? I wouldn't want this risk of scatter to prevent us using particulate-based dosage forms. A problem to be managed from the outset of design.

Diane Tang Liu: So basically, it's not a showstopper in drug development? You just have to manage the use? Okay. I just wanted to get a clarification from you.

Uday B. Kompella: Okay, we have a couple of minutes. Any other questions from the audience? Okay, Avi, do you want to add something?

Abraham Domb: The other method to get a controlled release of a drug into the eye is by making a prodrug. For example, a hydrophilic drug that possesses poor penetration through biomembranes, is bound to a lipid moiety that increases the penetration properties of the drug. The drug is not active as long as the lipid residue is still bound to the drug. A prodrug, if designed properly, may improve drug penetration and effectiveness.

Clive G. Wilson: So you have an implant. It swells, and if hot-melt extruded might curl, because it retains the strain of extrusion. So you have to be careful with that type of implant since it might change form after implantation. Ideally, in a rod design we would want a system that slowly erodes and the matrix not last too long after complete drug release.

Diane Tang Liu: Going back to the stimuli-induced drug delivery, we know that there are circadian rhythms, like the IOP. How much do we know about the daily changes, or even seasonal changes of the biomarkers in disease state so that we can specifically design the right dosing regimen or device?

Ashwath Jayagopal: Well, a collection of aqueous humor and vitreous humor samples isn't yet that sophisticated on a large scale to enable that type of tailored collection schedule. I mean, you get a sample when you can get it and when the permission allows it. I think we're a long way from answering those types of questions. But that's the kind of analysis we need to do. It's a great question.

Uday B. Kompella: Okay, let's thank all the speakers now. [APPLAUSE] Thank you all.