Downstream Processing 2023

Navigating lentiviral vector downstream bioprocessing: an engineering perspective

Cell & Gene Therapy Insights 2023; 9(9), 1117–1124

DOI: 10.18609/cgti.2023.148

Published: 5 October 2023
Andrea Rayat

Abi Pinchbeck (Editor, BioInsights) speaks to Andrea Rayat (Associate Professor, Bioseparations and Downstream Processing, Department of Biochemical Engineering, UCL). They discuss the pressing innovation gaps in the viral vector processing field, in addition to how to solve them through scale-down modeling, unit operation engineering, partnerships and collaborations.

Can you tell me about your path to work in the viral vector bioprocessing field?

AR: I started working in viral vector processing in 2018. Before that, I worked on antibody-based products such as monoclonal antibodies (mAbs) and antibody fragments (Fab). I have also been working on recombinant enzymes using E. coli. I work on the recovery and purification of various bioproducts using different expression systems. I see my background as a strength, that has powered my ability to bring in experience from other areas of biotechnology. I find analogies, parallels, or general patterns and connect my knowledge and expertise in these different product-systems to apply, reframe, or repurpose things that I have learned to create a new framework or concept to better understand viral vector processing.

What are you working on right now?

AR: We recently published a paper on viral vector processing, where we have shown, among others, that the impact of shear on viral vectors is not as severe as we thought [1]Perry C, Mujahid N, Takeuchi Y, Rayat ACME. Insights into product and process related challenges of lentiviral vector bioprocessing. Biotechnol. Bioeng. 2023; 120, 1–16., and I am completing a few more similar papers. I have been working on these for three or four years, and it has taken a long time and a lot of effort to reach this point. We gained a real depth of knowledge which enabled us to develop new methodologies. There is a lot of careful consideration in our approach to experimental design and how we conduct studies since they are, in fact, the first of their kind for viral vector applications.

What is your assessment of the current technological state of the art in lentiviral vector downstream bioprocessing—where have the most valuable recent advances been?

AR: In my experience, the study of viral vector processing, specifically lentiviral vectors (LVs), is thin on the ground. I became interested in moving into viral vectors because of my work on membrane processing; I realized that vectors were an up-and-coming class of products without many downstream bioprocess studies reported at that time, specifically on membrane filtration. Moving into the field, I realized that it takes a long time to establish yourself and to develop a track record of robust studies to add to the knowledge pool in this area.

In terms of the state-of-the-art, I found that most of the focus for LVs, and viral vectors in general, is on the first part of the production process itself. There is limited focus on downstream process studies. I am glad to see that we are moving towards working on understanding the process to produce and recover higher titers, including the applications of stable cell lines.

There are only a few stable cell lines out there, mostly inducible ones. Studies on the applications of these stable cell lines, including our work, show that they can be used to produce LVs. I am still seeing very little on the downstream processing side. Recently, I have seen several papers that are working on affinity chromatography for LVs, working to understand the kinetics of adsorption during anion exchange for viral vector production, and looking at producing data with ion exchange chromatography studies. All of these, however, are focused on one unit operation only.

What are the most pressing remaining fit-for-purpose innovation gaps?

AR: The gaps that we see surround creating materials for downstream processing specifically for LVs. However, we understand that this is a big task. Innovation in terms of creating such materials may come a bit later on. New analytical techniques that can help us evaluate the state of the viral vectors at each stage of the process are essential for process and product insights.

The other part that can help, at least for now, is publishing studies to broaden and deepen our understanding of viral vector processing. We need to know exactly what is going on; explain why we get certain results which can then help accelerate the growth in innovation in this space. We are now seeing this more, specifically in chromatography. More studies on membrane processes would also be useful, as well as the interaction between these unit operations, process materials, and equipment. I recently attended the ECI Single-use Technologies conference in Boston, MA. There are many plastic devices and assemblies in the manufacture of viral vectors, and in cell and gene therapy (CGT) applications in general. The pressing need to better understand the impact of single-use technologies in this emerging field was highlighted in that meeting. These include understanding or determining material characterization, supply chain, and regulatory requirements.

What are the impacts on/considerations for downstream processing of the various expression systems and cell lines currently utilized in LV production?

AR: One consideration is finding a more suitable cell line, or making the current cell lines we are using more suitable, for GMP manufacturing. Right now, we see that most people use HEK cells for producing viral vectors. Perhaps there might be other cell lines that are useful to produce LVs. We need to consider the things that are needed to produce such vectors because we know that sometimes their production is detrimental to cells. We are asking a lot from these cells to produce our viral vector products, but we know that the vectors may cause cytotoxicity later on.

The other consideration is seeing how cell engineering, and even engineering biology, can improve production. My background in biochemical and chemical engineering allows me to evaluate each unit operation in the manufacturing process and think about how they could be improved. Sometimes there are ways that we can improve unit operations by looking at the type of cells that we are using, or the type of other materials used in production such as reagents or buffers. Looking at the interactions between cell line expression systems and the engineering environment when producing such materials is important.

I am also an experimentalist. Although I appreciate the benefit of using artificial intelligence (AI) or models, at the moment, in some areas we do not have enough process data for such tools to be effective in guiding our process decisions. I am an advocate for using scale-down experimentation to produce data under different conditions, with different cell lines, and different operating environments. This helps process understanding by generating a lot of data while using mL-scale materials. Our time is now mainly used to focus on the interpretation and analysis of these data, and here we can leverage the power of modelling and even AI.

What are the contributing factors to the loss of LVs during the scale-up of membrane processes for cell and gene therapy manufacturing?

AR: There are many factors. From my understanding, we do not know the compounding effects of the materials used in the production, for example the buffer chosen, the beads used in chromatography, or the membrane filters together with viral vectors. There is also not a lot of information on the impurities that the cells produce together with the vectors or even information on the different variants of viruses that are being produced by the cells.

In a recent study, it is hypothesized that these variants are causing low recovery. They may be becoming bound to the chromatography beads compared to other variants that can be reversibly unbound from the beads. We do not yet have enough understanding of those aspects. Still, we can infer that the interactions between the vectors and the bioprocess material (the beads, the membrane filters, and the buffers) contribute to the losses. Our recent paper [1]Perry C, Mujahid N, Takeuchi Y, Rayat ACME. Insights into product and process related challenges of lentiviral vector bioprocessing. Biotechnol. Bioeng. 2023; 120, 1–16. shows some of these negative factor interactions between process shear and other process conditions or operating parameters during membrane filtration, indicating the need for precise selection of filtration conditions and their operation.

The other thing we do not know yet is the effect of localized processing conditions. As an analogy, let’s say you have a small wound or burn. You might think that as only a small part of your body is affected, it is not a big problem. However, it may cause a lot of issues later if there is an infection, for example. It is easy to see those spots on our skin, but these viral vectors are so tiny, and within a complex fluid, that it is difficult to study any small deformation on them which may cause huge problems in processing, including binding to columns.

An example of these localized effects would be a mismatch in the pH of the buffer and the type of membranes we are using. There are conditions that although good for the vector as a whole (e.g., choice of buffer pH), might cause some localized impact if there are negative effects due to factor-interactions (e.g., buffer pH may affect the fouling profile of the membrane filter as it changes membrane properties). These local factors could be one of the reasons for losses. Much of these is driven by the engineering environment, which has not yet been fully studied in many areas of viral vector processing.

Where do you see opportunities for improvement in the scalability of lentiviral processes during early development and clinical production?

AR: The opportunities for improvement would come with gaining knowledge from experts studying viral vector processing, especially LVs, and publishing and sharing data so we can see which types of conditions work for each specific purpose. Then, there would be the opportunity for AI-driven analysis to see whether there are links within these conditions.

Other opportunities for improvement could lie in scale-down processing. In our group at UCL, I develop ultra scale-down methodologies including devices, but this can only help people if vendor technology companies supporting the CGT manufacturing space also develop commercial products for scale-down methodologies.

The creation of specific biomaterials is key. We need membrane filters, adsorbents, and beads that are designed specifically for viral vector manufacturing. We need scale-down or miniaturized versions of these materials to be available during early development. There are a lot of larger-scale materials readily available, but for those who are just starting out, given the cost of production and the uncertainty of scale-up, it is risky to go straight to large scale. Scale-down tools and materials being commercially available would enable the early development of clinical production, especially moving toward full-scale manufacturing aligned with the materials used in early development stages. This means developing manufacturing processes from the start rather than reworking processes down the line.

What is your key advice to those looking to achieve high quality, high concentration LV products from robust and scalable processes?

AR: In terms of scalability, three key pieces of advice would be to look at using scale-down methodologies, question the perceptions and preconceptions of vector processing that apply in certain situations but may not apply to your product, and partner with an equipment vendor, or form industry–academic collaborations, to help evaluate and scale up processes. Given that many may not be familiar with the unit operations, vendor partnerships or industry–academia relationships can be valuable when designing early process and product development studies. Do not shy away from looking at the basics of the unit operations.

Testing scale-down versions is important, as is understanding what the scale-down version can offer in terms of process insights. These scale-down versions may be different from the GMP version or may be physically different from the larger scale equipment, such as our ultra scale-down devices. However, knowing what these scale-down versions offer could mean that you gain more insights earlier by appropriately applying these in process studies, early in the development stage. You can go to full scale trial with more confidence in certain aspects, having less guesswork at these crucial late stages of development. In our recent paper, we have shown that many preconceptions of viral vector processing are true in certain aspects but may not necessarily be true for all products or cell lines. Each scale-down process model needs testing. We have shown that shear, which is a factor generally considered to be the main reason for loss of recovery, does not necessarily cause losses. For membrane processing specifically, we need a relatively high shear, to help clean the membrane with the sweeping action of the process fluid parallel to the membrane. This helps the vectors to not adsorb onto the membrane.

In a relatively low-shear environment, which is often used in current processes, we experience losses and very low recovery. While it might be true that for certain vectors or certain cell lines high shear can lead to high losses, shear is not always the reason. Therefore, experimentally testing various membrane processing conditions for your own products and cell lines would be beneficial to optimize the process. Testing combinations of process conditions and materials with scale-down methods allows you to de-risk the cost of experimentation, and save time, as these scale-down experiments are quick and only require small volumes.

Finally, can you sum up your major goals and priorities for your work over the coming 12–24 months?

AR: In the next few years, I will be focusing on scale-down membrane processing for viral and non-viral vector applications, and remain active in other areas of recovery and downstream processing of biotech products. I continue to seek collaborations as this is the only way to unravel complex bioprocesses. Although I am very much an engineer working with the design and performance analysis of unit operations, I collaborate with other experts in virology, engineering biology and cell engineering, among others, to see how changes in the design of vectors, cells, or cell components can affect downstream purification. At UCL, we have a Centre for Doctoral Training (CDT) funded by the Engineering and Physical Sciences Research Council (EPSRC) where we collaborate with industry partners to address bioprocessing challenges. Some of my current projects with the CDT are with the Cytiva and Astra Zeneca Centres of Excellence (CoE) at UCL. We are investigating cell retention devices for perfusion and novel unit operations for process intensification. The insights that we obtain from this area may be applied to viral vectors.

We are completing our Biotechnology and Biological Sciences Research Council (BBSRC) project on LV production through novel engineering biology solutions. In this collaborative project with the UK’s MHRA, we separated the viral particle production from the envelope production and performed the assembly of the envelope and the viral particle outside the cell. I am excited about this project because this can also help us understand whether we can produce vectors in this way at a manufacturing scale, as this has already been performed in the lab.

Since we have separated the production of the envelope from the viral particle, we hope to use these two particles to develop an understanding of how LVs interact with the different process materials I mentioned earlier.

Finally, and most importantly, to accomplish all these, training our future bioprocess engineers and scientists in viral vector production, and in the CGT area, in general, is key. Educating students in our biochemical engineering degree programs and providing leadership, supervision and mentoring to the researchers in my research group are very important aspects of my role. I look forward to helping my final year students establish their publication record and complete their doctoral studies. I am excited to start three new doctoral projects this year, all on LV production and downstream processing. The projects are funded by the EPSRC-CDT with Cytiva CoE, the BBSRC-CTP (Collaborative Training Partnerships) with Oxford Biomedica, and Autolus through an Industrial Fellowship with the Royal Commission for the Exhibition of 1851. Although not without its challenges, seeing the progression of these students and researchers as contributors to the field is very satisfying, personally, and for the whole CGT industry, this is very much needed.


1. Perry C, Mujahid N, Takeuchi Y, Rayat ACME. Insights into product and process related challenges of lentiviral vector bioprocessing. Biotechnol. Bioeng. 2023; 120, 1–16.  Crossref

2. Perry C and Rayat ACME. Lentiviral vector bioprocessing. Viruses 2021; 13(2), 268.  Crossref


Andrea Rayat is an Associate Professor in Bioseparations and DSP at University College London (UCL). She holds Biochemical and Chemical Engineering degrees from UCL (PhD), TU Delft (MS), and UPLB (BS). Her main research aim is to contribute to the understanding of bioseparations of novel, biologically-derived products and to reveal the science and engineering that underpin recovery and purification operation, and their scale-up. Her group achieves this through the design and application of ultra scale-down and millifluidic devices, and other high throughput techniques. Together with scale-down studies, they apply modelling and multi-variate data analysis (and in the future, AI) to accelerate the robust manufacture of these novel biotech products. Her work on ultra scale-down as applied to industrial enzyme production has been recognised by the IChemE Global Awards (2020) and by the InnovateUK KTP awards (2021). In the area of LV downstream processing, her research group is one of the very few groups that study membrane filtration and other membrane processing steps for LV production. Their review paper [2]Perry C and Rayat ACME. Lentiviral vector bioprocessing. Viruses 2021; 13(2), 268., has comprehensively reviewed unit operations employed in LV production, outlining the critical process parameters and conditions for LV recovery. Their latest research paper demystifies the impact of process shear on LVs [1]Perry C, Mujahid N, Takeuchi Y, Rayat ACME. Insights into product and process related challenges of lentiviral vector bioprocessing. Biotechnol. Bioeng. 2023; 120, 1–16.. In the autumn of 2023, she is to start several projects in collaboration with cell and gene therapy companies which demonstrate the rapid expansion of her work in viral vector recovery and purification.


Andrea Rayat
Associate Professor,
University College London (UCL)

Authorship & conflict of interest

Contributions: The named author takes responsibility for the integrity of the work as a whole, and has given her approval for this version to be published.

Acknowledgements: None.

Disclosure and potential conflicts of interest: Rayat A discloses she has grants/contracts from UK BBSRC, UK EPSRC, UK MHRA, Pall (now Cytiva) Centre of Excellence and Astra Zeneca Centre of Excellence at UCL. She is to start new projects with Cytiva (with EPSRC), Oxford Biomedica (with BBSRC) and Autolus (with the Royal Commission for the Exhibition of 1851) from October 2023. She also received support for attending meetings and/or travel from Cambridge Healthtech Institute and Engineering Conferences International (ECI).

Funding declaration: The author received no financial support for the research, authorship and/or publication of this article.

Article & copyright information

Copyright: Published by Cell & Gene Therapy Insights under Creative Commons License Deed CC BY NC ND 4.0 which allows anyone to copy, distribute, and transmit the article provided it is properly attributed in the manner specified below. No commercial use without permission.

Attribution: Copyright © Rayat ACME. Published by Cell & Gene Therapy Insights under Creative Commons License Deed CC BY NC ND 4.0.

Article source: This article is based on an interview with Andrea Rayat carried out on Aug 16, 2023

Interview held: Aug 16, 2023; Revised manuscript received: Sep 18, 2023; Publication date: Oct 5, 2023.