home > ebr > spring 2013 > surviving the valley of death
European Biopharmaceutical Review

Surviving the Valley of Death

Developability is increasingly becoming a key area of attention to optimise efficiency and success during biopharmaceutical development. Designing and selecting appropriate lead candidates with the desired quality attributes can have a big impact in reducing attrition and costs associated with the development of new drugs

The pharmaceutical industry is currently crippled by the increasing cost of drug development – today’s figures suggest that developing a new drug costs, on average, many billions of dollars – combined with a reduction in R&D productivity and the pressing need for new treatments to be cost effective (1-3). The high attrition experienced during drug development has prompted the use of the term ‘valley of death’ to refer to those pre-clinical and clinical stages of development where most of the casualties occur (4). In this context, new front-loading, de-risking approaches to drug development are urgently needed in order to reduce failure in later stages of development, where associated costs are considerably larger (5).

Bioprocess development remains a central area of risk, not only because of its impact on the final cost of goods but, more importantly, because of its influence in key aspects of product quality, biological activity and safety. Still, the existing divide between discovery and development functions is posing far too great an obstacle to achieving efficient drug development (6). Furthermore, from a quality perspective at least, traditional bioprocesses are essentially ‘passive’ as their quality management is largely limited to measuring and reporting the output of the process, rather than driving the design or engineering the input to obtain a desired outcome. However, in recent years there has been a progressive shift in the perception of how bioproduction and early drug development should be articulated in order to address all these challenges. Some of the strategies include faster and more agile development, the integration of discovery and early process development inputs into drug design, or better predictability of bioprocesses and their output (7).

Developability in Drug Development

Developability is a concept originated in the early 2000s to define the suitability of a given therapeutic candidate to be developed successfully as a drug. Initially used in small-molecule APIs, it paid attention fundamentally to aspects of manufacturing, formulation, bioavailability and pharmacology, as well as toxicology and, among other things, sparked the development of new computational predictive approaches to address aspects of pharmacology and toxicology (8).

In recent years, the developability concept has started to take off in biopharmaceutical drug development. The main causes behind this are: the complex and largely unpredictable biomanufacturing processes; quality concerns and their impact in drug safety and product recalls; and the a-priori inscrutable outcome of clinical trials.

Developability assessment aims to understand the determinants of product quality, safety and efficacy, and can be grouped into three major categories (see Figure 1):

● Manufacturability: looking at whether the product could be made with an appropriate yield, and whether it would have an adequate quality profile and be able to be formulated for the desired route of administration

● Safety: addressing whether the product could potentially trigger unwanted immunological reactions, promote the production of anti-drug antibodies (ADAs) in patients, or show an adequate specificity profile

● Pharmacology/mechanism of action: assessing whether the product could be delivered via the chosen route of administration at the desired dose and with an adequate half-life. Additionally considering whether it would exert the right biological response, what patients could benefit from the treatment, and at what dosage regime

How to Implement Developability?

There is not a unified framework for the implementation of developability risk assessments in the early stages of development. Several platforms are being put forward, but there are two particular areas that seem to be getting more attention:

● The implementation of novel in silico platforms able to assess large numbers of potential candidates

● The development of surrogate in vitro assays capable of reproducing relevant process or physiological conditions

These two differing approaches show great promise in simplifying the assessment of therapeutic candidates by increasing dramatically the flexibility and throughput behind lead selection and optimisation stages, streamlining the ‘selection of the fittest’ for the given desired outcome.

Aggregation and Stability

Aggregation is perhaps the single most troubling quality attribute impacting biopharmaceutical development. Besides its obvious consequences in process yield, formulability or biological activity, protein aggregates seem to be an important contributing factor to reported cases of immune reactions in patients and other safety concerns (9). For example, aggregates are suspected to be involved in the onset of some of the ADA responses reported in patients. Recently, aggregates have also been attributed to pure red cell aplasia (PRCA) observed in patients treated with recombinant human erythropoietin (10). Other routes of chemical and physical degradation can also have an important impact in the stability of a product, including its ulterior aggregation, and can be linked to safety issues in patients, including immunogenicity.

New in silico predictive tools are being developed in an attempt to assess the relative stability and aggregation risks in biopharmaceuticals. Predictive tools to describe degradation reactions, such as deamidation, oxidation, or undesired post-translational modifications, such as unwanted glycosylation, are nowadays broadly available and can be used as an initial assessment during the product development cycle. More recently, algorithms have been proposed to predict the aggregation potential of polypeptides. Although some of these tools still lack extensive validation in biopharmaceutical systems, their application to re-engineering biotherapeutics with improved manufacturability (including reduced aggregation and increased productivity), has been successfully reported by various groups (7,11).

Parallel to the use of such new, predictive computational platforms, there is a need for simpler and faster analytics to report aggregation and other degradation mechanisms. Aggregation is a complex phenomenon that is not easy to describe in its entirety by a single experimental methodology. Furthermore, many of the existing technologies are not simple to implement or can be significantly time consuming. The use of particle imaging technologies or immunological assays are just a few examples of methods that could potentially offer a sufficiently large throughput for an early aggregation assessment or formulation study (12,13).

Formulation and delivery are areas of growing importance in biopharmaceutical development. There is a growing demand for molecules that can be formulated for more patient-friendly uses, and subcutaneous self-administration is a particularly attractive approach, especially for drugs that require chronic administration or multiple cycles of treatment. Many biotherapeutics, particularly monoclonal antibodies, often require administration at relatively large doses. For example, formulations of 150-200mg/mL are frequently required in subcutaneous administration because of volume limitations. Such formulations present a number of challenges in terms of stability, aggregation and viscosity. This is where a suitable formulability assessment could help identify candidates compatible with a required route of administration. In silico platforms able to predict aggregation and stability, combined with novel high-throughput analytical platforms, show great promise (7,14).

Immunogenicity Assessment

As most therapeutic proteins can be degraded to single amino acids, the main safety concerns for biopharmaceuticals are usually immune reactions or exaggerated pharmacology. Immune reactions to biopharmaceutical administration can be diverse in their nature and impact. In benign cases, production of ADA might not substantially affect the efficacy of the treatment, or perhaps only alter the pharmacology of the drug. In other occasions, however, ADA responses can neutralise the drug and render the treatment inefficacious. In more extreme cases, immune responses can take the form of hypersensitivity, anaphylactic reactions, cytotoxicity or autoimmunity, with severe consequences for the patient (15). Examples of adverse immunogenic reactions linked to the administration of biopharmaceuticals include the onset of PRCA in patients treated with recombinant erythropoietin, or anaphylactic reactions in patients treated with cetuximab (10,16).

The human immune system is extremely complex in nature and, most importantly, considerably different from that of animals. Even non-human primate models present significant variations to human subjects. To make matters worse, humans show a tremendous genetic variability in their immune components, particularly in their major histocompatibility complexes (MHC), which ultimately differentiate the ‘self’ from the ‘foreign’.

As a result of this, regulatory bodies state that animal models are not good predictors of clinical immunogenicity and therefore do not recommend their use for pre-clinical safety assessment and put the emphasis in the careful monitoring of ADA and other immune responses during the clinical development phases (17).

Recently, new methodologies have been introduced to assess immunogenicity risks in biopharmaceuticals. In silico tools evaluate the presence of T-cell epitopes as defined by the interaction of protein fragments with MHC Class 2 molecules. This type of methodology, albeit simplifying considerably the antigen presentation process, offers a privileged insight into the differential T-cell epitope content present in various therapeutic candidates, and because of its speed and low cost can be used during lead selection stages to identify molecules with a lower propensity to generate immunogenicity.

Additionally, ex vivo or in vitro cell-based assays using blood samples from human donors can provide very powerful information around comparability, safety of alternative candidates, or even the impact of formulation on product immunogenicity. This type of assay is already commonplace in vaccine development because of its proximity to human subjects compared with animal models.

Interestingly, regulators are starting to encourage the use of in silico and in vitro platforms as pre-clinical predictors of immunogenicity risks (18). By combining these two approaches, biotherapeutics with known immunogenicity problems can be successfully re-engineered and potentially re-introduced in clinical development (7). Furthermore, this type of assay could be extremely relevant in comparability studies and the development of biosimilars and biobetters (19).

Immunomodulation and Immunotoxicology

Many of the biotherapeutic molecules approved for their clinical use in human patients have an immunomodulatory effect. As indicated above, animal models have only limited utility in replicating characteristics of the immune system of human patients. In fact, one of the main obstacles to translational research is the disconnect between pre-clinical animal studies and the clinical reality of patients. This is why even in cases where rodent versions of the therapeutic candidate have been utilised, their mode of action might not necessarily replicate that observed in human subjects. Such reality makes pre-clinical development highly unpredictable and risky both for patients and drug developers. An extreme example of this uncertainty is the infamous clinical trial with the superagonistic anti-CD28 antibody TGN1412, which resulted in four healthy subjects having to be admitted into intensive care suffering systemic organ failure and very serious health complications (20). In recent years, a number of groups have managed to replicate cytokine release syndrome (cytokine storm) responses in the lab using cells from human donors (21). This opens the door to the use of ex vivo cell-based assays to characterise in detail both the mode of action and the side effects of immunomodulatory drugs. The application of these methodologies could be very relevant for patient stratifi cation, identifi cation of relevant biomarkers to be used in clinical trials, or the defi nition of safe clinical dosing for new drugs (22). Furthermore, they also support 3R initiatives in reducing the use of animal experimentation.

Designing Quality and Safety

Although there is still substantial room for improvement, cost of goods (COG) is no longer a signifi cant contributor to the total drug pricing for most established classes of biopharmaceuticals. In fact, the lion’s share of drug prices comes from the absorption of costs derived from the growing attrition and development failures. A focus on quality is replacing COG as the main paradigm in biopharmaceutical development. Quality affects essential aspects of drug safety, process robustness, and even biological activity that can make a difference between the success and failure of a new product.

So can we move beyond current exemplifi cations of quality by design (QbD) in process understanding and control, as defined in existing guidelines, and take them towards designing a desired quality target product profi le (QTPP) from the start? Developability methodologies can indeed be used to this end. Not only can they screen out candidates that might present difficulties or complications in terms of manufacturability, quality and safety, but they also make it possible to engineer products with desired quality attributes from the outset. Such an approach has been successfully implemented in a number of biotherapeutics with existing stability and safety issues; it points to a new way of designing biotherapeutics where de-risking methodologies are at the forefront of the discovery and development process to maximise success and minimise costs (7).

Developability Workflow

Figure 2 shows how a developability workfl ow could be structured. Early risk assessments using the methodologies described above can be utilised before committing to host development. Candidates classified as ‘low risk’ can then move into process development stages, whereas those classified as ‘high risk’ can be subject to a number of risk-mitigation strategies depending on where they are in development, or the type of risk identified and its impact. In cases where programmes are still in early stages of development, the selection of an alternative candidate – for example, from display or affi nity maturation libraries – can be an ideal solution. Alternatively, molecules can be subjected to a re-engineering programme that can very rapidly produce new, improved candidates with the desired properties. In cases where this is not an option (‘late stage’ programmes) processes need to be developed to address the nature and extent of the risk. It is important to note that, although process design can be powerful in modulating product yields and quality, there is only so much process tweaking can accomplish, particularly when dealing with diffi cult molecules. This work can, in fact, be lengthy and expensive, and might not be able to achieve a desired outcome: an example of this would be a stable high-concentration formulation or a non-immunogenic alternative.


To many, manufacturing development is still primarily about a necessary evil that every candidate needs to go through in order to enter clinical development for safety and efficacy validation. Still, there is far more to bioprocessing that meets the eye. The truth is that, although molecules belonging to the same class obviously share similar behaviour, a single amino acid replacement in a molecule can cause havoc in bioprocessing and in the patient. More importantly, many essential aspects relevant to candidate success are not directly related to their ‘binding’ activity, but to quality, safety, delivery and compliance aspects. Unfortunately, these are often not properly addressed during discovery and early development stages.

In this context, there are two main questions to be asked: “What is the value of quality?” and “Is it then worth investing in risk assessment and management strategies, introducing developability aspects in the design and selecting biotherapeutic candidates?” One could argue that quality actually has a great value in drug development and that developability makes economic sense. This is primarily true because the probability of failure and its costs are so absolutely staggering that any measure aimed at reducing failure and increasing efficiency in drug development will prove effective.

This does not mean that drug developers should ‘indulge’ in examining, testing and controlling every single quality aspect before moving forward into the clinic. On the contrary, it is about taking calculated risks based on knowledge rather than pushing candidates forward and betting on a final outcome (commercialisation) that nine times out of ten will never materialise. With better design and ‘filtering’ of candidates early on, attrition in later stages of development should ease with a very favourable impact in the generation of better and more cost-effective treatments.


1. Herper M, The truly staggering cost of inventing new drugs. Visit: www.forbes. com/sites/matthewherper/2012/02/10/ the-truly-staggering-cost-of-inventingnew- drugs

2. Paul SM, Mytelka DS, Dunwiddie CT et al, How to improve R&D productivity: the pharmaceutical industry's grand challenge, Nat Rev Drug Discov 9: pp203-214, 2010

3. Pammolli F, Magazzini L and Riccaboni M, The productivity crisis in pharmaceutical R&D, Nat Rev Drug Discov 10: pp428-438, 2011

4. Butler D, Translational research: crossing the valley of death, Nature 453: pp840-842, 2008

5. Pritchard JF, Jurima-Romet M, Reimer MLJ, Mortimer E, Rolfe B and Cayen MN, Making better drugs: decision gates in non-clinical drug development, Nature Rev Drug Discov 2: pp542-553, 2003

6. Rees H, Supply chain management in the drug industry: delivering patient value for pharmaceuticals and biologics, John Wiley & Sons, Inc, 2011

7. Zurdo J, Developability assessment as an early de-risking tool for biopharmaceutical development, Pharmaceutical Bioprocessing 1, 2013

8. Venkatesh S and Lipper RA, Role of the development scientist in compound lead selection and optimization, J Pharm Sci 89: pp145-154, 2000

9. Rosenberg AS, Effects of protein aggregates: an immunologic perspective, AAPS J 8: ppE501-507, 2006

10. Seidl A, Hainzl O, Richter M et al, Tungsten-induced denaturation and aggregation of epoetin alfa during primary packaging as a cause of immunogenicity, Pharmaceut Res 29: pp1,454-1,467, 2012

11. Jameel F and Hershenson S, Formulation and Process Development Strategies for Manufacturing Biopharmaceuticals: pp329- 347, John Wiley & Sons, Inc, 2010

12. Zurdo J, Michael R, Stallwood Y, Hedman K and Aastrup T, Improving the developability of biopharmaceuticals, Inno Pharma Technol 37: pp34-40, 2011

13. Mahler HC and Jiskoot W, Analysis of Aggregates and Particles in Protein Pharmaceuticals, John Wiley & Sons, Inc, 2012

14. Buck PM, Kumar S, Wang X, Agrawal NJ, Trout BL and Singh SK, Computational methods to predict therapeutic protein aggregation, Methods Mol Biol 899: pp425-451, 2012

15. Van De Weert M, Horn Møller E, Immunogenicity of Biopharmaceuticals: pp27-56, 2008

16. Chung CH, Mirakhur B, Chan E et al, Cetuximab-induced anaphylaxis and IgE specific for galactose-alpha-1, 3-galactose, N Engl J Med 358: pp1,109-1,117, 2008

17. ICH S6. Visit:

18. Buttel IC, Chamberlain P, Chowers Y et al, Taking immunogenicity assessment of therapeutic proteins to the next level, Biologicals 39: pp100-109, 2011

19. Barbosa MD, Immunogenicity of biotherapeutics in the context of developing biosimilars and biobetters, Drug Discov Today 16: pp345-353, 2011

20. Suntharalingam G, Perry MR, Ward S et al, Cytokine storm in a phase 1 trial of the anti-CD28 monoclonal antibody TGN1412, N Engl J Med 355: pp1,018- 1,028, 2006

21. Weissmuller S, Semmler LY, Kalinke U, Christians S, Muller-Berghaus J and Waibler Z, ICOS-LICOS interaction is critically involved in TGN1412-mediated T-cell activation, Blood 119: pp6,268- 6,277, 2012

22. Brennan FR, Morton LD, Spindeldreher S et al, Safety and immunotoxicity assessment of immunomodulatory monoclonal antibodies, mAbs 2: pp233-255, 2010

Read full article from PDF >>

Rate this article You must be a member of the site to make a vote.  
Average rating:

There are no comments in regards to this article.

Jesús Zurdo is Head of Innovation, Biopharma Development at Lonza, focusing on bioprocessing innovation and biopharmaceutical development de-risking. He has a PhD in Molecular Biology from Autonomous University of Madrid. Jesús cofounded Zyentia Ltd, which developed technologies to combat protein depositional disorders and aggregation problems in biotherapeutics. While there he held various roles, including CSO and CEO, until the acquisition of the AggreSolve platform by Lonza in 2007. He was also Co-founder and Scientific Advisor of Zapaloid Ltd and held several research positions at the Universities of Oxford and Cambridge in the UK and at CBM-CSIC in Madrid.
Jesús Zurdo
Print this page
Send to a friend
Privacy statement
News and Press Releases

ClinOne Closes Series A Funding Round to Scale its Site and Participant Engagement Platform, Data Driven Insights and Interoperability

BOSTON AND BOULDER, COLO. (PRWEB) 13 APRIL, 2023 – ClinOne, the leading provider of Adaptive Experiences in clinical trials announced today it closed a $3 million Series A funding to accelerate the growth of its platform and broaden its reach to serve customers. This latest round, led by Boulder Ventures, with participation from existing investors CU Healthcare Innovation Fund and First Trust Capital Partners, will continue to advance ClinOne’s mission to offer flexible engagement solutions for all participants to support their clinical trial journey.
More info >>

White Papers

Risk-Based Inspection Readiness: Reducing Headaches with Advanced Preparation

The Avoca Group

So often, inspections are a major source of stress in the clinical trial environment. This is especially true when processes involved in preparing for an inspection are reactive. But, it doesn’t have to be that way. Using industry-leading practices and resources enables a proactive approach to mitigate—or even eliminate—the pain points associated with preparing for inspections.
More info >>




©2000-2011 Samedan Ltd.
Add to favourites

Print this page

Send to a friend
Privacy statement