Cynthia A. Challener is a contributing editor to Pharmaceutical Technology.

Cynthia A. Challener

Magnifier with pills on blue background | Image Credit: © Pixel-Shot - stock.adobe.com

Pharmacokinetic (PK)/pharmacodynamic (PD) modeling of small-molecule drug candidates provides information essential for selecting and progressing optimal drug candidates. Effective PK/PD modeling, however, requires specialized expertise, time, and resources that are not always readily available.

Use of advanced artificial intelligence (AI) and machine learning (ML) algorithms can ameliorate resource limitations while also enabling the identification of patterns not readily apparent to human analysts, thus streamlining development while providing more robust predictions of 

PK/PD modeling plays important role in drug development

Understanding the absorption, distribution, metabolism, elimination, and toxicity (ADMET) properties of small-molecule drug candidates can inform the evaluation of small-molecule drug candidates through discovery and development efforts. Detailed information about these attributes can help weed out undesirable molecules and move forward candidates with a higher likelihood of success. In vitro and in vivo PK/PD screening studies are time-consuming and costly, however, with ADMET assessments often becoming a bottleneck in drug development programs (1).

In silico approaches for determination of ADMET properties have, therefore, become crucial to accelerating drug discovery and development efforts, allowing rapid evaluation of large numbers of compounds. “PK/PD modeling provides a quantitative framework that is essential for effective, data-driven drug development,” states Pauline Traynard, product manager for MonolixSuite at SimulationsPlus. “By integrating diverse data from laboratory experiments and clinical studies, these models allow us to understand the complex interplay between a drug’s concentration in the body and its therapeutic effect. This mechanistic insight is critical for optimizing dosing regimens, designing more efficient and informative clinical trials, and making crucial go/no-go decisions,” she explains.

Ultimately, Traynard observes, PK/PD modeling helps reduce late-stage attrition rates and supports a more streamlined, cost-effective, and successful path to regulatory approval and clinical use.

Increasing complexity creating challenges

Greater understanding of disease mechanisms and advances in rapid synthesis of compound libraries based on increasingly complex molecules, many which have limited solubility (and therefore bioavailability) under physiological conditions, has created several challenges to achieving effective PK/PD modeling.

“Emerging trends such as highly potent molecules with narrow therapeutic indices and complex drug delivery technologies (e.g., long-acting injectables or lipid nanoparticles) pose significant modeling challenges,” says Traynard. She also notes that many advanced drug candidates exhibit nonlinear kinetics, require specialized absorption models, or interact with biological barriers in non-intuitive ways.

An additional significant challenge, Traynard adds, arises during clinical development. “Understanding and predicting the high degree of variability in how different patients respond to a drug is difficult to decipher from the limited samples per patient typically collected during clinical trials,” she comments.

AI/ML algorithms overcome many modeling challenges

AI/ML algorithms are instrumental in overcoming many of these modern PK/PD challenges, according to Traynard, due to their superior ability to identify complex patterns in high-dimensional data where mechanistic understanding is still incomplete.

The most impactful applications for AI/ML-based PK/PD modeling are, Traynard observes, in early-stage discovery and the enhancement of mechanistic models. “In discovery, AI/ML-driven prediction of ADME and toxicity is invaluable for rapidly screening vast compound libraries to de-risk projects before significant investment is made,” she explains.

For instance, Traynard notes that AI/ML models can rapidly and accurately predict a full suite of ADME properties directly from a chemical structure, allowing scientists to screen and optimize thousands of candidates virtually and focus resources on the most promising ones. In clinical development, meanwhile, ML-based models can in early development be employed to analyze sparse patient data for efficient identification of factors contributing to variability in drug response, leading to more robust population PK models and better-informed dosing strategies, according to Traynard.

AI/ML techniques are particularly valuable for classes of APIs where traditional modeling methods face limitations due to underlying biological or chemical complexity. Traynard highlights compounds with poor solubility and permeability (Biopharmaceutics Classification System Class II/IV), where AI/ML can effectively model the intricate relationship between a drug’s formulation and its in vivo absorption to predict bioavailability. “It is also used for APIs with complex safety profiles or non-linear pharmacokinetics, as AI/ML is adept at integrating diverse biological data to identify safety signals, such as the potential for organ injury, that are difficult to predict with simpler models,” she says.

Significant benefits can also be obtained when using ML models in conjunction with in vitro-in vivo extrapolation and physiologically based pharmacokinetic (PBPK) modeling because integration of these approaches provides a more comprehensive picture of a drug’s ADMET profile (2). Augmentation of established PBPK and population PK/PD models with model structure learning and parameter estimation is, says Traynard, a second key application. As an example, such modified models are useful for identifying influential covariates or nonlinear relationships within clinical data.

A third important application of advance algorithms in PK/PD modeling is the automation of model development workflows, which Traynard believes can reduce time-intensive manual steps by implementing ML-guided model selection, fit optimization, and diagnostics.

The biggest challenge to enabling the widespread use of AI/ML algorithms to improve PK/PD modeling is the same issue facing the development of AI/ML models for all other applications in the biopharmaceutical industry: the need for large quantities of high-quality, reliable data. “Limited access to large, high-quality datasets needed to train reliable AI models is a major challenge to broader AI/ML adoption in PK/PD modeling, largely due to restrictions and experimental variability,” observes Traynard.

Another important concern noted by Traynard is the inherent “black box” nature of some algorithms, which may lack the transparency needed for regulatory or internal scientific confidence.

Hybrid approaches combining established mechanistic models like population PK/PD or PBPK with interpretable AI components are therefore gaining traction. “This strategy grounds the AI’s powerful pattern-recognition abilities in the context of known biology, making the results more interpretable, scientifically plausible, and trustworthy for both scientists and regulators,” Traynard comments.

Regulatory acceptance is crucial, of course, and there is growing interest from regulators in AI/ML’s potential for population simulation, dose selection, and decision support, according to Traynard, particularly when integrated into PBPK or exposure-response models used in regulatory filings.

“The core regulatory concerns,” observes Traynard, “are model transparency, validation, and managing bias to ensure that any AI/ML tool is fit-for-purpose.” As a result, she says, there is a strong emphasis on establishing “good machine learning practice” and developing explainable AI that allows regulators to understand the basis of a model’s predictions.

Going forward, Traynard believes the evolution of AI/ML in PK/PD modeling is likely to center on a deeper integration with mechanistic science, creating powerful hybrid systems that are both more predictive and explainable. “As these predictive models become more robust, we anticipate advances in automated clinical trial simulation and design, where AI could be used to forecast trial outcomes under different scenarios and optimize protocol parameters in silico before a single patient is enrolled,” she adds.

Looking further out, Traynard notes that leveraging AI to deliver truly personalized medicine represents an important frontier. “The ambition,” she explains, “is to use sophisticated algorithms to integrate vast, patient-specific datasets to enable more accurate personalized predictions of an individual’s unique response to a drug, paving the way for individual treatment optimization.”

Myung, Y.; de Sá. A.G.C.; and Ascher, D.B. Deep-PK: Deep Learning for Small Molecule Pharmacokinetic and Toxicity Prediction Open Access. 

 52 (W1), pp. W469–W475. DOI: 10.1093/nar/gkae254

Bassani, D.; Parrott, N. J.; Manevski, N.; and Zhang, J. D. Another String to your Bow: Machine Learning Prediction of the Pharmacokinetic Properties of Small Molecules.

 19(6), 683–698. DOI:10.1080/17460441.2024.2348157

Cynthia A. Challener, PhD, has been a freelance technical writer for more than 25 years and is a contributing editor to 

Articles

Pharmacokinetic/pharmacodynamic (PK/PD) modeling is crucial for drug development, providing insights into drug absorption, distribution, metabolism, elimination, and toxicity (ADMET). However, traditional PK/PD modeling faces challenges due to complex drug properties and patient variability. Advanced AI and ML algorithms offer solutions by identifying patterns in high-dimensional data, enhancing early-stage discovery, and improving mechanistic models. Despite their potential, AI/ML models require large, high-quality datasets and face transparency issues. Hybrid models combining AI with mechanistic approaches are gaining traction, promising more predictive and explainable outcomes.

Artificial intelligence and machine learning can help identify complex patterns.

Overcoming PK/PD Modeling Challenges

medicine and vision concept - woman closeup and eye diagram, future technologies | Image Credit: ©Julia - stock.adobe.com

Aging of the global population and the rise of chronic diseases, particularly diabetes, have contributed to an increase in the occurrence of eye diseases, such as glaucoma, keratitis, age-related macular degeneration (AMD), dry eye disorder (DED), and diabetic retinopathy (DR). Pharmaceutical companies have responded with growing investment in the research and development of novel eye treatments and drug delivery systems designed to address key challenges in ocular therapy. Current treatments range from traditional small-molecule and antibody-based drugs to cell and gene therapies and nanoparticle delivery solutions, along with various types of ocular implants and the use of drug-eluting contact lenses (1,2).

One area of focus has been the development of solutions for the sustained delivery of treatments, particularly to the posterior part of the eye. Many ocular therapies are formulated as eye drops that must be administered frequently to overcome the many physical and mechanistic barriers in the eye and ensure sufficient dosing. Sustained and targeted delivery products are helping reduce the dosing frequency by overcoming ocular barriers and increasing bioavailability.

Delivery of therapeutics to the anterior, or front of the eye, is generally simpler than delivery to the posterior part, because topical eye drops are an option. There are limitations to their effectiveness, however. “Drops do not maintain good bioavailability and are relatively quickly flushed out by nasolacrimal drainage,” explains Jeffrey S. Heier, chief scientific officer with Ocular Therapeutix. “While they are a convenient mode of delivery,” he adds, “they present challenges in terms of not only bioavailability, but limited duration of therapeutic effect, both of which contribute to inconsistent patient compliance and suboptimal outcomes.”

Treatments for glaucoma (high eye pressure), as well as infections and inflammation of the eye, are good examples. Glaucoma eye drops typically need to be administered at least once daily, and in many cases, multiple times a day. For some conditions, Heier observes, drops must be delivered on an hourly basis. “Such a dosing frequency is only sustainable for a short period of time. In fact, it is readily accepted that even with easy-to-administer options like eye drops that must be applied just once daily, adherence remains a major issue due to the frequency of this regimen,” he says.

Another option for treatment of the anterior segment of the eye is straightforward intracameral injections, according to Heier. This method of administration delivers ocular therapies directly into the anterior chamber of the eye, allowing for targeted delivery and resulting in enhanced bioavailability (3).

While treatments for infection, inflammation, high eye pressure, and neovascularization have been shown to be effective, this method of delivery is considered “off-label,” as FDA has not approved products specifically for this use, except two intracameral implants (see below). In addition, many patients are hesitant to receive eye injections.

Targeting the back of the eye presents unique challenges, not only in achieving therapeutic durability but also in navigating complex regulatory expectations, Heier comments. Indeed, delivery to the posterior segment is anatomically more complicated, especially when the durability of treatment is a primary objective.

“Although advances in intravitreal medicines have shown incremental increases in durability compared to prior standard of care, current retinal therapies rely on frequent intravitreal injections to maintain their effect,” observes Peter Jarrett, chief technical officer for Ocular Therapeutix.

For many patients, even therapies that require administration every three to four months can be challenging, as this type of regimen places a significant burden on patients and care partners, according to Jarrett. “To treat the full spectrum of patients with retinal disease or other conditions affecting the posterior segment, there is still a need for more durable treatments, and the development of longer-lasting treatments is such a critical focus,” he says.

New sustained-delivery therapies, however, must be proven to be as safe and effective as existing treatments that may require much more frequent administration. In addition, Jarrett notes that targeted delivery to the retina and posterior segment of the eye necessitates overcoming several blood-ocular barriers, including the blood-aqueous barrier and the blood-retina barrier.

“Historically,” Jarrett explains, “overcoming these physiological defenses has required delivering drugs at high concentrations. However, this approach often results in increased systemic exposure, which in turn can lead to a higher incidence of adverse events and toxicities. Striking the right balance between effective ocular delivery and systemic safety remains a focus in the development of retinal therapies.”

Potential for better treatments with greater convenience

While small steps taken to enhance ocular delivery have had positive impacts, there is still real potential for innovative solutions, particularly those leveraging sustained-release delivery technologies, to have dramatic effects on patient outcomes.

“While incremental improvements can be achieved through sophisticated formulations, such as enhancing solubility, prolonging ocular surface retention, or improving penetration through ocular barriers, these refinements often reach a ceiling in terms of clinical impact,” Heier remarks. “In contrast, transformative advances in treating chronic eye diseases are more directly realized through sustained-release delivery technologies, which bypass many of the pharmacokinetic limitations of topical or injectable drugs by providing continuous, controlled delivery of therapeutics directly to targeted intraocular tissues over extended periods,” he contends.

In addition, Jarrett emphasizes how important it is for developers to realize that results obtained in clinical trials often do not translate to the real world for treatments with dosing regimens patients have difficulty consistently maintaining as part of their daily lives. “A durable therapy that reduces treatment burden could offer not just convenience, but potentially better long-term outcomes for patients who would otherwise fall through the cracks of an intensive treatment schedule. Recognizing this balance between ideal clinical trial conditions and the real-world experience of patients is key to advancing more durable treatment options,” he states.

Nanoscale drug-delivery systems are recognized to achieve enhanced bioavailability and targeted and sustained delivery of many types of therapeutics addressing a wide variety of conditions, including eye diseases (2). Examples of different types of nanoparticles used for ocular drug delivery include nanomicelles, liposomes, polymeric nanospheres, and peptide-based nanoparticles (2).

Nanomicelles in preclinical testing have been shown to experience good uptake by corneal cells and when administered topically to achieve sustained delivery of drugs to treat DED, neovascular AMD, inflammation, and other eye disorders (1). Biocompatible lipid nanoparticles can deliver drugs with both hydrophobic and hydrophilic properties for extended periods. In one animal model of glaucoma, large unilamellar liposomes achieved sustained delivery of latanoprost, a common drug used to reduce eye pressure, for up to five days (4).

Peptide-based nanoscale delivery systems are particularly attractive because they are not only biocompatible, but can be used to target specific cell receptors, disrupt endosomal membranes, and carry drug payloads to the nucleus (1). Of the two main types of systems (elastin-like polypeptides and silk fibroin polymers), the former have been investigated for sustained release of drug substances from contact lenses due to their ability to remain adhered to their surfaces.

Researchers continue to investigate the potential of nanoscale materials for the formulation of extended-release ocular therapeutics, with efforts focused on increased drug-loading capacity, ocular residence time, and release periods while reducing potential toxicities (1).

Drug-eluting contact lenses and canalicular plugs

Given the limitations and issues associated with the use of eye drops and intracameral injections for treatment of conditions in the anterior segment of the eye, other options for delivery are looking to leverage the accessibility of this portion of the eye, according to Heier. Examples include drug-eluting contact lenses and canalicular plugs.

As indicated previously, sustained, topical delivery of ocular therapies to the anterior of the eye is possible using contact lenses loaded with active drug substances, such as peptide-based nanoparticles carrying drug payloads. It is even possible to improve drug loading by incorporating non-covalent binding sites into the lenses when they are manufactured (1). Delivering drugs using contact lenses is attractive because this approach is non-invasive and can potentially support increased drug uptake compared to topical applications (eye drops, creams, ointments). ACUVUE Theravision contact lenses (Johnson & Johnson Vision Care), which provide sustained delivery of ketotifen for the treatment of itchy eyes, are the first FDA-approved drug-eluting contact lens product (5).

Canalicular plugs belong to a class of treatments known as punctal plugs, which are inserts designed to slow or block the outflow of tears from the eye (6). Punctal plugs can be temporary, semi-permanent (resorbable), or permanent (synthetic, usually silicone), and some are able to provide sustained release of drug substances. They are most widely used to treat DED and have been shown over three decades to be generally safe and effective; although, they can fall out, become partially displaced, or lead to watery eyes.

Recent developments have helped address many of these issues. One example is a canalicular gel device (Lacrifill, Nordic Pharma) introduced in June 2024 (7). The chemically crosslinked particulate hyaluronic acid (HA) hydrogel device is designed to completely fill the canaliculus regardless of its size in different patients, preventing dislodgement and the buildup of stagnant tears and mucus that can result in infection. It is meant to be used for up to six months and can be easily removed by applying a sterile solution.

Significant advances have also been made over the past two decades in ocular implant technology for sustained release of drugs that treat a number of eye diseases. Implants can be biodegradable, although some are not (8). The former can suffer from the rapid release of remaining active drug substance upon degradation, while the latter require surgical removal when the drug substance is fully released.

Bioerodable implants are thus receiving more attention. These products are slowly solubilized upon exposure to biological fluids, leading to slow release of the drug and absorption of the implant material into the eye, avoiding a final “burst” release (8). The rate of dissolution can also be controlled by modifying the chemical composition of the polymer used in the implant. Theoretically, these devices could provide sustained release of therapeutics for months to years.

Jarrett highlights the use of degradable or erodible matrix materials for the sustained delivery of small-molecule drugs. “The choice of drug and implant material are extremely important in designing such a sustained-release system,” he says. “Challenges include drug potency, safety, control of drug release rate, local and systemic drug concentrations, implant biocompatibility and bioresorption, implant size and placement position, and the injection technology utilized,” Jarrett denotes.

As an example, Jarrett points to bioresorbable hydrogel technology as offering a long-term option for delivering small molecules to the posterior segment. Hydrogels, he says, are well-suited to deliver extremely low solubility, lipophilic drugs because the release rate is being controlled by drug solubility. “Drug solubility in both the hydrogel matrix and in the vitreous humor fluid are important rate control factors. These factors may be combined to create a constant release rate in the eye,” he explains.

In addition, the bioresorbable hydrogel vehicle can be designed to break down into water soluble molecules that are neutral pH, non-reactive, and readily cleared from the eye, with timing of resorption tailored to synchronize with drug depletion, allowing repeat injections without crowding the vitreous with empty vehicles or “shells” from earlier injections.

Implants have, in fact, been developed for insertion into both the anterior and posterior segments of the eye. The devices are either implanted surgically or via injection (intracameral for anterior and intravitreal for posterior applications). Injected implants are typically smaller as they must be able to fit into a needle (7).

Two examples of FDA-approved intracameral implants are Durysta (Allergan) and iDoseTR (Glaukos), both of which provide sustained release of drugs (bimatoprost and travoprost, respectively) to treat intraocular pressure in open-angle glaucoma/ocular hypertension.

There are also several intravitreal implants approved by FDA, including Ozurdex (AbbVie) for macular edema and Dexycu (EyePoint Pharmaceuticals) for post-operative inflammation. Encelto (Neurotech Pharmaceuticals), which leverages an encapsulated cell therapy technology to continually deliver therapeutic doses of ciliary neurotrophic factor to the retina to treat macular telangiectasia type 2, received FDA approval in early 2025 (9).

FDA approval was also granted in early 2025 to Susvimo (Roche) for the treatment of DME, which continuously delivers ranibizumab via the Port Delivery Platform, a refillable (every six months) eye implant surgically inserted into the eye (10).

While there remain many challenges to effectively treating diseases of the eye, significant strides have been made with recent advances in technologies supporting sustained, targeted release of effective ocular therapies. With many additional, innovative solutions in preclinical through late clinical development, additional progress is expected.

Whalen, M.; et al. Seeing the Future: A Review of Ocular Therapy. 

Chandel, A. and Kandav, G. Insights into Ocular Therapeutics: A Comprehensive Review of Anatomy, Barriers, Diseases, and Nanoscale Formulations for Targeted Drug Delivery. 

https://doi.org/10.1016/j.jddst.2024.105785

Gautam, M.; et al. Intracameral Drug Delivery: A Review of Agents, Indications, and Outcomes.

 39 (2), 102-116. DOI:10.1089/jop.2022.0144

Natarajan, J.V.; et al. Sustained Release of an Anti-Glaucoma Drug: Demonstration of Efficacy of a Liposomal Formulation in the Rabbit Eye. 

Johnson & Johnson Vision. Johnson & Johnson Vision Care Receives FDA Approval for ACUVUE® Theravision™ with Ketotifen - World's First and Only Drug-Eluting Contact Lens. Press Release, March 2, 2022. 

https://www.prnewswire.com/news-releases/johnson--johnson-vision-care-receives-fda-approval-for-acuvue-theravision-with-ketotifen--worlds-first-and-only-drug-eluting-contact-lens-301493964.html

Cleveland Clinic. Punctal Plugs. ClevelandClinic.org (accessed June 20, 2025). 

https://my.clevelandclinic.org/health/treatments/25249-punctal-plugs

Charters, L. New Canalicular Gel Shows Promise for Dry Eye Treatment. 

https://www.ophthalmologytimes.com/view/new-canalicular-gel-shows-promise-for-dry-eye-treatment

Wykoff, C.C.; Kuppermann, B.D.; and Saïm, S. Extended Intraocular Drug-Delivery Platforms for the Treatment of Retinal and Choroidal Diseases. 

Neurotech Pharmaceuticals. Neurotech’s ENCELTOTM (revakinagene taroretcel-lwey) Approved by the FDA for the Treatment of Macular Telangiectasia Type 2 (MacTel). Press Release, March 6, 2025. 

https://www.neurotechpharmaceuticals.com/wp-content/uploads/Neurotech_Press-Release_BLA_Approval_FINAL.pdf

Roche. FDA Approves Roche’s Susvimo as the First and Only Continuous Delivery Treatment for the Leading Cause of Diabetes-related Blindness. Press Release, Feb. 3, 2025.

https://www.roche.com/media/releases/med-cor-2025-02-04#:~:text=Susvimo%C2%AE%20(previously%20called%20Port,edema%20(DME)%20that%20continuously%20delivers

Cynthia A. Challener, PhD, is a contributing editor for 

Hydrogels, drug-eluting contact lenses, and other implant technologies show real promise.

Advanced Solutions for Sustained Delivery of Ocular Therapies

Efficient and effective delivery of complex and challenging drug substances is an ongoing issue facing developers of modern medicines. Delivery systems must overcome poor solubility and bioavailability while achieving targeted delivery to achieve maximum efficacy with lower doses while also minimizing unwanted side effects. Nanoscale technology is proving to be an attractive solution on all accounts, particularly for realizing targeted delivery into specific types of cells and across biological barriers.

Many of the unique properties of nanoscale delivery systems can be attributed to their small size (< 100 nanometers). At this scale, the surface area/volume ratio is high, leading to unique physical, chemical, and biologic properties. The ability to tune the surface properties of nanomaterials through functionalization further increases their utility for drug delivery, particularly with respect to their targeting capabilities. Nanoscale delivery systems can not only increase the bioavailability of poorly water-soluble drugs; they can also protect sensitive APIs from degradation and, when appropriately designed, remain stable in the bloodstream, allowing long-term sustained release and more consistent pharmacokinetics (1). The greater cellular uptake of nanoscale delivery systems also contributes to improved efficacy and reduced side effects.

Another important advantage of nanoscale delivery systems is the ability to deliver two or more APIs simultane

Nanoscale co-delivery systemsoffer improved performance for combination therapies because they deliver the different drug substances in a homogeneous manner (2). Ideal nanoscale carriers are stable, small, but able to support sufficient drug loading of both hydrophobic and hydrophilic APIs; protect the loaded drug from degradation; are suitably functionalized to exhibit a high binding affinity to target cells/ tissues; and have good biocompatibility.

Many different types of nanoscale delivery systems (e.g., lipidic, polymeric, inorganic, biomimetic) provide opportunities for designing solutions with high specificity for the particular API, disease target, and route of administration. In fact, they support the loading of most types of drug substances, including small molecules, proteins, nucleic acids, and diagnostic agents (1). They can also be customized in support of personalized medicines, and some have been developed that are capable of exhibiting both therapeutic and diagnostic activities (theranostics).

As more information is gained about the benefits of nanoscale delivery systems, researchers are exploring the applicability of a growing variety of nanoscale materials for this purpose (1–3). Nanoparticles have attracted particular attention, including those based on lipids, polymers (natural and synthetic), lipid-polymer hybrids, dendrimers, inorganic materials (metals, magnetic compounds, carbon nanotubes, silica-based systems, etc.). Other nanocarriers include nanocrystals; DNA-based systems; extracellular vesicles such as exosomes, microvesicles, apoptotic bodies, and phagosomes; nanoemulsions; and micelles.

Like any drug delivery technology, nanoscale systems are not perfect solutions and, depending on the particular nanotechnology, have different sets of limitations (1). Some of the nanomaterials used for drug delivery can, if the systems are not designed appropriately, pose toxicity and biocompatibility issues. One example is metallic nanoparticles if they accumulate in the body.

Similarly, while some nanoscale delivery systems can exhibit prolonged lifetimes in the bloodstream, others have short circulation times or can be viewed as foreign materials by the immune system and rapidly cleared. Some nanomaterials may also exhibit unexpected/unpredictable interactions with biological systems that lead to detrimental effects. These uncertainties and complexity of nanoscale drug delivery systems often lead to regulatory challenges.

The development of robust, predictable large-scale manufacturing processes can be a significant hurdle for some nanomaterials/nanoscale delivery systems. In addition, stability during storage can be an issue for some nanoscale systems, requiring careful formulation development. All these challenges tend to lead to high costs for therapeutics and vaccines leveraging nanoscale drug delivery systems.

Highly targeted delivery with nanoscale systems can be achieved using passive or active mechanisms. Passive delivery involves accumulation of the nanoscale system in the targeted cell/ tissue through enhanced permeability and retention (EPR), which is dictated by the shape, size, charge, and other surface properties of the unfunctionalized nanomaterial (4). This approach does not, however, completely reduce delivery to healthy cells, which can result in off-target effects.

Active targeting, meanwhile, is realized through functionalization of the surface of the nanocarrier system to both increase cellular uptake and ensure delivery to specific cells/tissues (4). Ligands are attached to the surfaces of the nanocarriers that have high binding affinity to receptors specific to the target cells. These ligands can include a range of compounds, including antibodies, aptamers, peptides, small molecules, and biomimetic materials such as cell membranes, among others. As an example, nanoscale delivery systems can be functionalized with PD-L1 monoclonal antibodies in anticancer therapies.

The targeting ability of nanoscale systems can benefit treatments for all disease classes, but this capability is of even greater importance in certain applications. Targeting of cancer cells is a major goal in oncology, as traditional systemic treatment result in severe side effects. The ability of nanocarriers to selectively attack targeted tumor cells allows for delivery of higher doses to improve efficacy while reducing cytotoxicity (1).

Therapeutic hyperthermia is a unique treatment made possible by use of nanoparticle delivery systems (1). Once the nanoparticles are localized in the tumor, they are exposed to laser radiation or an applied magnetic field, leading to an increase in temperature sufficient to kill the cancer cells. The improved targeting capability of nanoscale delivery systems also leads to improved activation of immune cells in cancer vaccines.

Nanoscale delivery systems are improving treatments for infectious diseases including human immunodeficiency virus (HIV) due to their targeting ability. Patient adherence with traditional medications is often low due to the high incidence of side effects. The high on-target bioavailability of nanoscale drugs thus has a measurable, positive impact.

Finally, nanoscale delivery systems are also improving the effectiveness of treatments for autoimmune diseases due to their improved targeting capability (1). Nanoparticles have been shown to deliver antigens and immunomodulators to antigen-presenting cells and lymphocytes with high specificity, and thus more efficiency, inducing antigen-specific tolerance 

Lipid nanoparticles (LNPs) have received significant attention in recent years as attractive systems for the delivery of nucleic acids for vaccine applications. Their success was clearly demonstrated with the approval of the two messenger RNA vaccines against COVID-19, according to McDavid Stilwell, CFO with CPTx. He notes, however, that some of the properties that make LNPs useful for vaccine applications (inflammatory behavior) also present substantial hurdles for systemic therapeutic delivery applications that require higher doses and for which adverse inflammation is undesirable.

There are also issues with targeting LNPs away from the liver, Stilwell observes. The field is working hard to overcome cell targeting challenges, but solutions are not expected soon. Stilwell notes that LNPs have now been successfully engineered to target immune cells, and CPTx is using this approach to deliver single-stranded DNA for its in vivo chimeric antigen receptor (CAR)-T development program.

In addition to ensuring targeted specificity and minimal off-target effects, key challenges with LNPs— and nanoscale delivery systems in general—include achieving consistent manufacturing scalability, controlling batch-to-batch variability, overcoming biological barriers (immune recognition, clearance, endosomal escape), and demonstrating robust safety profiles, adds CPTx’s CEO Hendrik Dietz. “Achieving reliable, reproducible, and cost-effective large-scale production remains critical, especially for transitioning from preclinical validation to clinical and commercial stages,” he says.

CPTx believes that DNA nanostructures presenting an immune-silent vector are a better alternative to currently available gene delivery technologies. “These nanocarriers are optimal for precision targeting, programmable drug release, and transient genetic expression, making them valuable for controlled gene editing and transient immunotherapies,” Stilwell contends. The goal, he adds, is to present a delivery solution that is not only technically superior, but that can also be manufactured in a cost-effective fashion.

The single-stranded DNA nanocarriers CPTx is developing are intended to enable direct, targeted in-vivo gene delivery. Scalable, high-purity DNA single- strand synthesis platforms, both for the gene encoding components of the nanoparticles and the helper material needed to build the nanocarriers, have been an important initial focus to ensure manufacturing challenges are not an issue, according to Dietz.

These new DNA nanostructures are precision engineered from therapeutic immune-silent DNA single-strands using a technique known as programmable DNA nanofabrication, uniquely combining precise control over composition and 3D shape, as well as delivery functionality. “The result is precise cell targeting without external carriers, bypassing issues associated with existing delivery technologies such as viral vectors and lipid nanoparticles, including manufacturing hurdles, immunogenicity (in the case of viral vectors), and targeting challenges,” Dietz states.

The DNA nanocarriers “integrate therapeutic payloads and delivery functionality into rationally engineered particles designed to realize episomal, non-integrated gene expression in target cells or tissues with minimal side effects,” Dietz explains. Functionalization in a site-specific fashion with molecular targeting motifs promotes highly selective binding to specific cells or tissues and efficient cellular uptake, he says.

The modular approach also allows for simultaneous delivery of multiple therapeutic gene payloads and/or simultaneous targeting of multiple tissues or cell types, enhancing effectiveness, safety, and control compared to traditional approaches, according to Dietz.

CPTx is currently working to improve the delivery efficiency in various prototypes while also validating safety and efficacy in animal models with the goal of licensing this DNA nanocarrier delivery solution to drug developers with established payload genes and target tissues.

1. Tenchov, R. et al. Transforming Medicine: Cutting-Edge Applications of Nanoscale Materials in Drug Delivery. 

. 2025 19 (4). https://pubs.acs.org/ doi/10.1021/acsnano.4c09566

2. Sun, L. et al. Better Together: Nanoscale Co-Delivery Systems of Therapeutic Agents for High-Performance Cancer Therapy. F

ront. Pharmacol., Sec. Pharmacol. Anti-Cancer Drugs

. 2024 Vol. 15. DOI: 10.3389/fphar.2024.1389922.

3. Li, Y.; et al. A Nanoscale Natural Drug Delivery System For Targeted Drug Delivery Against Ovarian Cancer: Action Mechanism, Application Enlightenment And Future Potential. 

Front. Immunol., Sec. Canc. Immun. & Immunother.

, 2024 Vol. 15. https://doi.org/10.3389/\ fimmu.2024.1427573

4. Fan, L.; Wang, A.; and Shi, D. Targeted Nanoscale Drug Delivery Systems For Melanoma Therapy. 

. 2023 Vol. 86, 104724. https://doi. org/10.1016/j.jddst.2023.104724

Unique surface properties and ligand functionalization provide real advantages.

Achieving Targeted Delivery with Nanoscale Systems

Insulin Vial Trapped in Ice for Diabetic Challenges | Image Credit: ©Nisit – stock.adobe.com

While many parenteral products can be stored and shipped in the liquid state at room temperature, some, including most based on biologic drug substances, must be maintained at low temperatures. When refrigeration is not sufficient, two options remain: keeping the products in a frozen state or freeze-drying (lyophilizing) them into a powdered form. There are advantages and disadvantages to both, with many factors to consider when choosing between the two solutions.

The biggest driver for freezing or freeze-drying parenteral products is instability of the drug substance when in solution. This issue does not often arise for small-molecule APIs, but is common for recombinant proteins and antibody-based molecules. “Developing freeze-dried or frozen parenteral formulations is primarily driven by the inherent instability of biological molecules in solution at or below room temperature, limited shelf life in liquid form, high sensitivity to temperature fluctuations and air-liquid interfaces, and excipient-related limitations,” says Maria Fernanda Zuluaga Estrada, principal scientist and formulation development group leader at Lonza.

Protein molecules, explains Jeremy Guo, senior vice president and head of drug-product development and clinical drug-product manufacturing at WuXi Biologics, are physically and/or chemically unstable during long-term storage, suffering from aggregation and visible particle formation, deamidation, oxidation, and other degradation pathways, even in optimized aqueous formulations after formulation screening.

“Freezing or lyophilizing these formulations is the optimal choice, because physical and/or chemical reactions will be inhibited in the frozen glassy state or freeze-dried state with minimal water content,” Guo says. At Lonza, advanced analytical tools are used for diagnosing vulnerabilities potentially leading to instabilities, including mass spectrometry to detect and foresee post-translational modifications, holistic (i.e., stability-indicating) surfactant analytics, subvisible particle characterization to differentiate potential protein-versus-surfactant instability, in-depth root cause analysis in case of surfactant degradation, in-depth host-cell protein, especially lipase, characterization and quantification, and container-closure integrity testing and qualification.

Beyond instability, time can also be a driver for choosing to freeze or lyophilize a drug candidate. “When the development timeline is very limited for early filing of an investigation new drug (IND) application, the final formulation needs to be locked down without completion of comprehensive formulation screening or generation of sufficient stability data. Freezing or freeze-drying can help ensure long-term stability while still meeting these shorter timelines. Comprehensive screening can then be performed during late-stage development to determine the final, commercial formulation,” Guo observes.

There is one other significant reason for moving forward frozen or freeze-dried formulations during the development of biologics, according to Zuluaga Estrada. Many liquid biologic formulations have limited expected shelf-lives when stored at or below room temperature due to surfactant degradation, which is why Lonza conducts so many surfactant-related analyses. “For cases where extensive lipolytic activity is demonstrated, Lonza’s proprietary lipase assay can support the removal of lipolytic activity during downstream processing,” she notes. The company also has a proprietary approach to lipase inhibition that involves addition of low concentrations of citrate to mitigate surfactant degradation, which can potentially prevent the need for frozen or lyophilized storage, Zuluaga Estrada adds.

Advantages and disadvantages of frozen formulations

The primary advantage of frozen formulations is their ability to preserve unstable formulations due to the immobilization of molecular components, thereby extending the shelf life, observes Zuluaga Estrada. Development of frozen formulations also generally takes less time and is less expensive than the development of lyophilized products, according to Guo. The lack of the need for a reconstitution step, which is required for freeze-dried formulations, is another benefit noted by Guo.

There are several disadvantages to frozen formulations, however. “The complexity of developing frozen formulations is often not considered in full, and the related challenges present a substantial risk in the development of biologics,” Zuluaga Estrada contends. She highlights lack of molecule or excipient stability in certain frozen formulations (e.g., cold denaturation and aggregation of proteins, crystallization of excipients, phase separation) and compromised integrity of the container-closure system at ultralow temperatures as two examples.

Beyond stability issues, the key challenge with frozen products is the need for extensive infrastructure and shipping capabilities to support the cold chain, says Zuluaga Estrada. “Robust freezing parameters, appropriate storage conditions, and adequate supply chain management need to be considered,” she notes.

Temperature excursions during storage and shipment of frozen products, adds Guo, can lead to ice crystal formation, protein denaturation, and a potential decrease in drug efficacy. In addition, the need for temperature control during storage and shipment at manufacturing sites, depots, and clinical centers adds significant cost and inconvenience, he comments.

Other concerns pointed out by Guo include the higher risk of vial breakage and the need for lower filling volumes compared to liquid drug products in the same container to reduce this risk.

To avoid these issues, often redevelopment to a liquid or lyophilized formulation is pursued once a candidate reaches later development stages and looks to be commercialized. However, switching to a liquid or freeze-dried product during late-phase development requires significant additional effort and adds cost and time for programs, according to Zuluaga Estrada.

Lyophilized formulations present opportunities and challenges

Freeze drying also has both positive and negative attributes. In addition to overcoming formulation instability issues, one of the biggest benefits is the ability to store and ship products at refrigerated (2–8 °C) or even room temperatures, Guo says.

Appropriate storage of lyophilized cakes within these temperature conditions (i.e., below their glass-transition temperatures), significantly reduces the logistical and storage constraints compared to frozen products, agrees Zuluaga Estrada. “That translates to greater storage and transport efficiency, particularly for global products and patient-centric delivery,” Guo concludes.

As with frozen formulations, however, it is necessary to use lower filling volumes than those of liquid products in the same container to ensure efficiency of the lyophilization process. In addition, Guo observes that the development of freeze-drying processes requires specialized expertise and capabilities and often adds time. The lyophilization process is also energy-intensive and extends the overall manufacturing time.

The need for reconstitution before administration also poses some concerns. Not only does Guo note the potential risk of product contamination by stopper debris after multiple punctures for reconstitution and sampling, high-concentration formulations can be challenging with respect to the time required and the potential for dosing errors and introduction of contaminants.

Despite the complexity, expense, and extended timelines typically associated with the development of lyophilized products, with the right expertise and robust platform cycles, it is possible to develop freeze-dried formulations in similar timeframes compared to liquid formats, according to Zuluaga Estrada. “We employ specialized characterization techniques including modulated differential scanning calorimetry (mDSC) and freeze-drying microscopy (FDM) to identify key thermal changes and, as required, optimize and develop robust lyophilization cycles in cost-effective timelines and with low material requirements,” she explains.

Given the different advantages and disadvantages of freezing and freeze-drying of parenteral formulations, there are some general factors that often influence the choice of one method over another. First is the ability of the drug substance to withstand the lyophilization process and/or freeze-thaw cycles, according to Zuluaga Estrada. The clinical application and route of administration are also important considerations. For instance, she notes that the acceptance of lyophilized formats is lower for emergency medications that must be administered immediately and medicines given via intravitreal administration. Other aspects that should be taken into account include the development phase, supply chain, timelines, and costs.

In general, says Guo, frozen products are often preferred if the development timeline is severely limited (e.g., six months to IND); the product is unstable and sensitive to the freeze-drying process; and the medicine will only require distribution locally or to clinical centers with sufficient cold-chain infrastructure. In addition, he notes that long-term storage at -20 °C is preferred over storage at ultra-low temperatures due to the greater availability of this capability at clinical centers.

Lyophilization is typically preferred, Guo continues, for unstable products that can withstand the freeze-drying process, there is sufficient time for development of an optimal lyophilization process, a long shelf-life is needed, and/or the product will be distributed to globally, including areas where cold-chain management would pose challenges.

“Overall,” Guo concludes, “when selecting between frozen and freeze-dried formulations, scientific, economic, and practical factors should be considered. Stability of the molecule for long-term storage is the primary concern. Other factors such as timeline, costs, storage availability, and distribution logistics should be evaluated as well.”

The evaluation of lyophilized formats can be a backup option for proteins that are unstable in solution at room temperature or under refrigeration or for cases where considerable surfactant degradation (potentially leading to particle formation) is observed, Zuluaga Estrada says.

When lyophilization does not improve protein stability or a lyophilized format is not suitable, such as when large fill volumes or long reconstitution times are involved, frozen storage of the liquid at sub-zero temperatures should be assessed.

Cynthia A. Challener, PhD, is a contributing editor to 

When referring to this article, please cite it as Challener, C.A. Parenteral Formulation: Deciding When to Go Frozen or Freeze-Dried.

Scientific, economic, and practical factors should be considered when choosing between the frozen state and lyophilization.

Parenteral Formulation: Deciding When to Go Frozen or Freeze-Dried

Drug developers face increasing pressure to accelerate process development and reduce costs while ensuring safety and efficacy of investigational candidates and marketed products. They must achieve these goals while minimizing the impact that development and manufacturing operations have on the environment. Given shorter development times and the challenges associated with making process changes in later development stages, the best approach for improving sustainability is to incorporate sustainable practices into process designs as early as possible.

Considering sustainability from the outset

Sustainability-by-design in bioprocess development, according to Adam Goldstein, senior director of R&D Collaborations at Thermo Fisher Scientific, involves integrating sustainable practices and principles from the very beginning of the development process, which includes considering environmental impacts, resource efficiency, and waste reduction during the design phase, to create processes that are not only efficient but also lower the environmental footprint.

Because 80% of a drug’s final environmental impact is determined during the early stages of process design (1), Kami Krista, CEO of Elio Technologies, adds that sustainability needs to be part of the decision-making process during the early R&D and chemistry, manufacturing, and controls (CMC) phases—particularly before changes become difficult due to regulatory constraints (e.g., once process inputs are defined in the regulatory dossier).

In addition, because 42–47% of a pharmaceutical company’s emissions profile can be attributed to purchased goods and services (2), Krista emphasizes the importance of considering both process inputs and conditions. That means employing more environmentally friendly chemicals and processes.

Incorporating sustainability-by-design into process development activities is most effectively achieved by integrating sustainable practices throughout design, construction, and operation of biopharmaceutical production sites, according to Jacob Skou, sustainability manager at FUJIFILM Diosynth Biotechnologies. Key elements that are considered include resource and energy efficiency, waste minimization, and use of less hazardous chemicals and processes. “Processes are designed that optimize the use of raw materials, energy, and water, and alternative renewable energy options such as solar and wind are used where possible to increase energy efficiency. Furthermore, strategies are implemented for reducing, reusing, and recycling waste material generated during bioprocessing,” he says.

Krista concludes that “to successfully integrate sustainability, it is essential that every scientist, process engineer, and other relevant stakeholders can easily assess the sustainability implications of their decisions—alongside technical and cost considerations—and weigh the associated trade-offs.”

Environmental, operational, and cost benefits

Adopting a sustainability-by-design approach in biopharmaceutical manufacturing offers a wide range of benefits in addition to reducing the impact of development and manufacturing operations on the environment. Skou highlights cost efficiency, compliance and risk management, innovation, and increased social responsibility as important benefits resulting from the integration of sustainability into bioprocess development activities.

“Reducing resource and energy consumption and minimizing waste reduction benefits the environment and contributes to cost reductions. Furthermore, as environmental standards become more stringent, adopting sustainable practices helps ensure continued compliance and reduces the risk of legal penalties and reputational damage. Taking a sustainability-by-design approach can also drive innovation that results in market-differentiating processes. Finally, sustainable practices emphasize ethical considerations; promote the health, safety, and well-being of employees and communities; and thus align with broader social responsibility goals,” Skou explains.

Goldstein emphasizes the importance of the innovations that result from adopting a sustainability-by-design approach because they can lead to improved product quality and process efficiency.“Driving higher titer in smaller volumes means yielding higher throughput of product in a smaller footprint. The benefits are lower cost of goods in manufacturing while simultaneously delivering the product in a much more sustainable fashion with regards to emissions reductions,” he comments.

Another important benefit of taking a sustainability-by-design approach for Krista is the fact that by integrating sustainability improvements from the outset, it is not only possible to dramatically reduce the impact of biopharmaceutical operations and the overall environmental footprint of the product and company, the high costs and challenges to making sustainability improvements during later stages are avoided.

There are marketing advantages as well. For contract development and manufacturing organizations (CDMOs), Skou believes having an improved sustainability profile is a competitive advantage.

Krista also notes that in Europe, many hospital procurement systems (i.e., England’s National Health Service and the Joint Nordic Procurement system) are including sustainability performance as a decision criterion when selecting drug products. For public companies and early-stage firms seeking specific investors, having a good sustainability rating (often based on product stewardship, sustainable raw material selection, or similar ecodesign criteria) can attract more investors. Sustainability is also increasingly featured as a decision criterion by Big Pharma companies when making merger and acquisition decisions, according to Krista.

All aspects of biopharmaceutical manufacturing should be considered when taking a sustainability-by-design approach to bioprocess development, starting with the choice of cell line through methods for waste processing, according to Goldstein.

“High-titer cell lines are more productive, while chemically defined media reduce contamination risks and allow for sourcing of materials from sustainability-minded suppliers. Using centrifugation rather than depth filtration can reduce waste and processing times while improving yields, and higher-capacity chromatography resins can reduce buffer usage,” explains Goldstein.

Intensified processing both upstream and downstream can, meanwhile, reduce manufacturing footprints and resource consumption while reducing waste generation and improving facility throughput.

Upstream intensification examples highlighted by Goldstein include the use of high-density cell banking to skip long drawn-out seed expansions (reducing the six to eight days of standard good manufacturing practice [GMP] expansion) saving time, money, and resources (plastic, water, and media) by employing the use of single-use bioreactors (SUBs) that have high turndown ratios, which allows seeding at very low operating values and the growth of cells to high densities as the volume is expanded in the same unit operation SUB step. Downstream opportunities include the use of membrane separations.These technologies replace larger chromatography columns that use large volumes of buffers and take entire shifts to run.

Real-time monitoring of processes via inline sampling, often an important component of process intensification and continuous processing, also helps ensure more robust and high-quality processes and products, according to Goldstein. “Additional efficiencies are gained using ballroom designs vs. traditional costly, broken-up manufacturing suites, allowing for greater use and flexibility of manufacturing space, while funneling of waste into circular waste streams via collaborations with waste-stream recyclers add further benefits,” he notes.

The choice of water purity for certain activities can also have a significant sustainability impact. Highly purified water (like water for injection) is extremely costly and resource laden from a carbon footprint perspective, due to the vast number of steps (steam evaporation, filtration, heating, cooling) that are involved in its creation per liter, Goldstein explains. Specifying a lower quality of purified water (like reverse osmosis quality water Type 2 vs. Type 1) can and should be considered in the makeup of media, cleaning of equipment, and the creation of general early-stage purification buffers (until the process is in the final formulation stage). “Careful consideration and selection of the standard of water used for many manufacturing operations can drive large impactful improvements,” he states.

Collaborating across the entire supply chain is, in fact, an essential component of a sustainability-by-design approach to bioprocess development, according to Ann-Cathrine Rosendahl, global sustainability manager with FUJIFILM Diosynth Biotechnologies. “When we talk about sustainability-by-design in bioprocess development, it’s not just about what happens on the factory floor—it’s about embedding sustainability into every step of the journey, from raw materials to the end-of-life of a product, to build a more resilient and responsible supply chain,” she notes.

Given that a significant part of bioprocess emissions comes from upstream activities (92%), engaging suppliers early is essential, Rosendahl says. “The sustainability of suppliers should be evaluated early, and we are working towards ensuring that partners who prioritize environmentally friendly practices—such as responsible sourcing and transport—are prioritized in our selection process,” she states. FUJIFILM Diosynth Biotechnologies is also committed to reducing environmental impact by working closely with suppliers to improve energy efficiency, adopt cleaner technologies, and explore innovative solutions that lower emissions. “By integrating sustainability into our business reviews and setting clear expectations, we are building stronger partnerships focused on continuous improvement,” Rosendahl remarks.

Sustainability aspects for final product management must also be considered. Where possible, packaging and logistics should also be optimized to reduce waste and emissions, according to Rosendahl. Promoting transparency and accountability is equally important to her, which she observes can be achieved by developing metrics for assessing sustainability performance and reporting progress to stakeholders. That, Rosendahl underscores, requires access to accurate data to ensure a clearer understanding of the Scope 3 footprint for biopharmaceutical manufacturers (indirect greenhouse gas emissions generated outside the company across its value chain).

“Having better data helps companies identify where the biggest impacts are and explore opportunities with suppliers to make processes more sustainable over time,” Rosendahl adds. “By addressing these aspects, companies can move toward an approach to sustainability that considers not just the products themselves but also the surrounding processes, infrastructure, and practices, leading to a holistic and effective sustainability strategy,” she concludes.

As an example, Goldstein highlights a program his company has established to connect customers with recyclers that can recycle single-use bioprocess containers (BPCs) after use. “This program helps biopharma manufacturers divert waste from landfills or carbon-intensive incineration, advancing their sustainability goals through a convenient recycling service that turns BPCs into high-quality plastic lumber. Approximately 400,000 pounds of plastic waste has been diverted from landfills or incineration since initiation of the program,” he observes.

Work is now focused, Goldstein says, on identifying recyclers that can accept mixed waste from biomanufacturers and convert those materials to bio-oil, which can then be recycled into new polymers. These new polymers can be used either in the biomanufacturing environment or in entirely separate recycled products, such as faux wood, fillers for roads, etc.

The key to successful pursuit of a sustainability-by-design approach to bioprocess development is effective assessment of the environmental impacts of each unit operation. The first step, emphasizes Krista, is to equip scientists, process engineers, and all others involved in process development with tools that allow them to understand the sustainability impact of the decisions they are making.

Those tools must then be employed to thoroughly assess process development activities and identify opportunities for incorporating more sustainable practices, adds Goldstein.

Growing adoption, but several hurdles to overcome

The adoption of sustainability-by-design is growing, driven by regulatory pressures, consumer demand for greener products, and corporate environmental sustainability goals, according to Goldstein.

As environmental responsibility continues to become a major focus, the bioprocessing industry is transforming its approach to sustainability, with a major driver being requests from customers and investors to deliver on this agenda, evidence that strong sustainability strategy and actions is good business, adds Peter Skals, director of global sustainability for FUJIFILM Diosynth Biotechnologies.

In fact, based on 2024 data, nine of the top 10 pharma companies by revenue have ecodesign targets and activities embedded into their sustainability strategies, Krista observes. “The breadth of these ecodesign targets and activities vary from specific targets like green chemistry to programs that cover the entire value chain. In addition, while Big Pharma companies are leading the charge on ecodesign, due to pressures from customers, investors, and regulators, biopharma service providers, small and emerging biotechs, and suppliers are all stepping up activities focused on improving sustainability,” he comments.

Reaching aggressive sustainability targets requires overcoming some notable challenges, however. For instance, broader application in the bioprocessing sector is inhibited by inherent limitations of the bioprocess itself, which requires a different approach due to quality and safety restrictions, according to Skals. In addition, Goldstein notes that designing in sustainability for process development must begin in the pre-clinical phase, with the capability of selected raw materials demonstrated via qualification early in the development stage.

Sustainability also needs to be integrated into pretty much every design decision, leading to a very high volume of decisions that it must be seamlessly embedded into, Krista says. In addition, he notes there are inherent trade-offs to be made, with each right decision context-dependent, increasing the complexity of the decision-making process because more parameters must be considered.

Other important challenges noted by Krista include the lack of available “clean” data for performing sustainability assessments and the time- and cost-intensive nature of current assessment methods. “A lot of the data that [are] required to assess the sustainability impact of process design and process input selection [do] not exist or [are] hard to get,” he says. Much is based on responses to supplier surveys, which are not in sufficient quantity to support effective decision-making. Traditional lifecycle assessments, meanwhile, require weeks to months to complete, and can cost in the hundreds of millions for typical processes if you include the assessments for each of the different process-design and process-input variations at the relevant decision points (3).

Further increasing the challenge is the fact that most technical staff in bioprocessing plants are not sustainability experts. “Upskilling people in sustainability to the level that would be necessary to complete sustainability assessments is not feasible, which means straightforward guidance is needed,” Krista says.

Finally, Krista emphasizes the importance of not only quantifying and naming the sustainability pressures driving ecodesign today, but assessing those that will be present 10 years from today when investigational candidates eventually reach the market. “The most sophisticated pharma companies in the context of ecodesign are considering this impact lag and extrapolating in order to assess sustainability pressures beyond the response of a particular political party in power today, but where the longer arc of history is bending towards,” he contends.

Despite all these challenges and uncertainties, Skals believes that industry innovations and the inclusion of sustainability practices in the design phase is where the big wins can be achieved. But the incremental improvements of traditional designed bioprocess are important as well, he adds, and reduce waste without compromising product quality, stability, or safety.

Advances in digital technologies are making those assessments easier to perform and providing more valuable insights. Advanced technology facilitates access to reliable process monitoring data, allowing identification and prioritization of potential areas of improvement and implementation of appropriate process adjustments, according to Skals.

Technologies such as automation, real-time monitoring, and advanced analytics help in achieving precision and efficiency, adds Goldstein. He comments that innovations in materials science also contribute to more sustainable bioprocesses.

Machine learning (ML) can help overcome many of the challenges that exist in applying sustainability-by-design today, Krista contends. “At Elio, we believe ML is critical to ensuring sustainability-by-design can be adopted as widely as possible, as early as possible, and thereby have the biggest impact. Specifically, it can help with the challenge of insufficient, high-quality data,” he says.

ML models, Krista explains, in combination with data engineering, can help fill in data gaps with sophisticated estimations using data about the specific product, supplier, and/or specific process aspects from a wide variety of data sources that are available publicly (freely or through a license). “Using this approach can enable bypassing of the data bottleneck coming from relying on supplier surveys,” he adds.

“Pharma companies are already making decisions based on very crude generalizations, such as based on product category (e.g., all alcohols have the same footprint) or other characteristics (e.g., a chemical with a higher molecular complexity has a higher environmental footprint),” Krista observes. “ML models can simply be trained on a much higher dimensionality of data and thus can make higher accuracy predictions,” he contends. Using ML tools thus also addresses the time and cost issues associated with sustainability assessments, as the calculations can be completed much more quickly and at scale.

Incorporating sustainable practices into process designs as early as possible ensures optimal performance.