Spotlight interview with Dr Mark Treherne, CEO of Cellesce
We had the privilege of speaking with Dr Mark Treherne, CEO of Cellesce - a leading Welsh biotech company that has developed patented bioprocessing technology for the growth and expansion of organoids. The company is currently commercialising organoid models as a powerful new technology for enhancing drug discovery and genetics research.
With over 25 years of experience in therapeutic development, we were excited to speak to Mark about his views on the current biopharmaceutical market and what upcoming technologies he feels will help to revolutionise cancer drug development.
How has your scientific career to date led you to Cellesce?
I completed my PhD in Pharmacology at the University of Cambridge and was always interested in early-stage drug discovery. I worked on 3D tissue culture models at the University of Basel before joining Pfizer, where I focused on neurodegenerative diseases. In 1997, I set up Cambridge Drug Discovery with other colleagues from Pfizer, which was subsequently acquired by BioFocus plc (now part of Charles River). Since then I have worked with many early-stage biotechnology companies from foundation through to trade sale. In recent years, I have been involved with Talisman Therapeutics, which focuses on building induced-pluripotent stem cell (iPSC) models for Alzheimer’s disease, so joining Cellesce felt like a hugely exciting natural progression to me.
From my experience in the drug development sector, overall drug success rates have always been low. And I hope by supporting drug discovery programmes in the industry with our bioprocessed 3D organoids, Cellesce can be instrumental by helping pharma and biotechs to be more successful. It will be transformational if we can help our partners get 2 out of 10 drugs into the clinic, instead of just 1 out of 10 as is currently the case on average.
In your opinion, what’s responsible for the low success rate of cancer drugs? What can biopharma do to mitigate this risk?
The reason for this is simple - cancer biology is complex to mimic in vitro and we just do not have an adequate understanding of it. This makes drug development difficult because we do not know what works until we invest considerable time and money to test each candidate in cancer patients. The overall success rate has always been roughly 10%. That 10% may eventually become a new therapeutic product, or it may not, and the other 90% may end up being published in a paper but produces no direct benefit to cancer patients. The question is, how can we ensure that the new drug candidate we are investing in falls within that 10% “success” category? If we can somehow predict which drugs are more likely to progress to the clinic, then it would be possible to discard the failing candidates early, i.e. fail fast and fail cheap, and then re-focus resources on the promising drugs to maximise efficiency. In this case, I think using 3D organoids in preclinical oncology studies, instead of conventional cell lines, could help us improve success rates, as patient-derived tumour organoids can better recapitulate most of the key aspects of the tumour progression observed in vivo.
Can you tell us how 3D organoids can serve as a more predictive preclinical model?
Cell lines are actually also 3D, but they usually have an altered morphology and genetics in order to be viable as a monolayer in plastic culture. This may cause them to respond differently to drugs and ultimately the results may not translate into patients. On the other hand, 3D organoids retain spherical morphology when they self-assemble on extracellular matrix gels, similar to how a tumour starts out in a patient. In terms of genetics, the mutational patterns of organoids also tend to be preserved when you compare the patients’ biopsies and the patient-derived organoids (PDOs). Because PDOs are grown from cancerous adult stem cells rather than embryonic stem cells or iPSCs, the oncogenic mutations typically carry over. In addition, 3D organoids are composed of multiple cell types, they are made up of heterogeneous cells, forming necrotic cores, undergoing metastasis. Ultimately, this enables most biologically relevant cancer properties to be recapitulated inside a dish. The high degree of physiological relevance in 3D organoids could then lead to a more relevant pharmacological response. For example, a team from The Institute of Cancer Research conducted a study with 71 patients enrolled in clinical trials and showed that the PDOs had high similarities in terms of morphological and molecular landscapes with their original tumour biopsies. The study also compared the treatment responses between PDOs and PDO-based mouse tumour models, with the patients in the clinical trials. The result showed that in 100% of cases, if the drugs did not work on the PDOs, the same drugs did not work on the patients either; and in 90% of cases, if the drugs worked on the PDOs, the same also applied to the patients.
How long can organoids be cultured for? What are the benefits of using organoids over in vivo mouse models?
Organoids can in theory be kept in culture indefinitely. While they may form necrotic cores and get bigger, they can be reset back into a more uniform starting state to maintain the consistency of the model. This property also allows resistance studies to be conducted with PDOs, using high concentrations of drug candidates for over 90 days. Of course, the same study could be done in animal models like mice, but generally the use of organoids is cheaper and has a higher success rate when it comes to model generation compared to engrafting human tissues in mice. This brings us to another problem in mice, which is the interaction between human tissues and the host. As we know patient-derived xenograft (PDX) models have to be immunodeficient to prevent an attack on the engrafted human tumour. However, in organoid studies you can include co-cultures with patient-matched immune cells to allow for more comprehensive studies in a more controlled environment. We are not trying to displace the work of animal models, but we hope that with the aid of organoid technology, the use of animals in preclinical testing can be reduced. Nonetheless, one of the reasons we work with Repositive is so pharma and biotech researchers can triangulate the best answer using both PDO and PDX models, as PDOs are also compatible with transplanting into mouse models, allowing the results to be compared side by side.
Will heterogeneity and patient variation affect the reliability of PDOs as a preclinical model?
Indeed, as I mentioned, PDOs are heterogeneous and may grow at different rates. Similar to a tumour, they could also transform along the way and even develop resistance to drugs. So not all PDOs may end up identical, but our technology at Cellesce allows all PDOs to start off with similar sizes and grow in the same controlled environment. But this heterogeneity, as well as inter-patient variation, may actually make PDOs a better preclinical model.
Firstly, while there might be variation in some of the growth parameters, tumour phenotypes and pharmacology are still preserved. During the growth of PDOs, their cancer phenotypes are driven by strong oncogenic mutations, which allows the cancer cells to dominate in the tumour population and eliminate other less cancerous cells. This means that in a 3D environment, PDOs will undergo a similar evolutionary process and ultimately end up with common oncogenic mutation signatures that are important for cancer progression. In short, PDOs with the same cancer subtype should carry similar mutational patterns for pharmacology studies. Our partners can also take healthy tissues around tumour biopsies and create patient-matched controls for further bioinformatics analysis and validation.
Second, patient variation could actually contribute to a more ethical and efficient clinical trial approach, where PDOs from a clinical trial cohort are tested with the drugs in advance, and only the patients whose PDOs respond well to the drugs get selected to proceed with the testing. This maximises the benefit to patients and provides early clinical readouts, helping to make pivotal decisions sooner.
How is Cellesce’s technology helping to encourage the use of 3D organoids?
Cellesce has patented technology to expand organoids in scale using our automated bioprocessors. Once the bioreactors are fed, we ensure the bioprocessed organoids start with similar sizes and shapes, so that the genetic drift, as well as variation of growth curves, and pharmacological variation is minimised. The current generation of our bioprocessors can reduce variation in these parameters significantly compared with manually generated batches of organoids, while also increasing production by 20-fold for a single scientist, and we are aiming to improve this to 60-fold in the next generation technology. The idea is that with less human intervention and more standardised automation, we can reduce batch-to-batch variation, so you can get similar results from batch #1 in year 1 and batch #30 in year 10, and you do not have to go back and revalidate studies.
Our biobank partner, the Wales Cancer Bank, offers organoids derived from colorectal cancer patients, which are validated with full exome sequencing, mutational analysis and pharmacology data to demonstrate their relevance to tumour pathology. Our technology is a cost-effective option for biopharma to get easy access to large scale, validated and quality-assured organoids for their drug development studies. Of course, it is also fine for organisations to spend their resources if they want to manually expand organoids for technical studies, but our technology allows researchers who are interested in using 3D organoids to use our service and refocus their resources elsewhere.
In the past, 3D organoids were not compatible with the industry standard of 2D high-throughput screening (HTS) frameworks. The challenge of hydrogel pipetting was one of the technical blockers for organoid generation, but now commercial gels are available, which can be pipetted automatically. We also have a partnership with Horizon Discovery to show that using 3D organoids in HTS is possible. Our organoids can be fitted in 96- or 384-well plate standard formats and used to perform HTS with Horizon Discovery’s automated screening robotics and large compound library. Now you can screen hundreds of thousands of compounds with our organoids to uncover more promising leads to pursue.
Can you share any teasers about Cellesce’s upcoming projects?
Currently, we are offering colorectal cancer organoids with a diversity of mutational patterns, but we are working on expanding the offering to include more genetic diversity for colorectal cancer organoids, as well as organoids of other cancer types with our biobank partners. We have a grant from Innovative UK to work with the University of Manchester on scaling and validating breasts cancer organoids. Besides, our licencing deal with Hubrecht Organoid Technology (HUB) allows us to use their technology to stimulate certain pathways in the organoids to form mini guts. This can hopefully aid us in mapping the organoid technology to other epithelial-derived cancers. Another project we are working on with our collaborators is focused on the application of organoids for immuno-oncology. By matching patient immune cells and their corresponding organoids, we’re trying to address the question of whether immune cells can be reprogrammed to attack cancer cells.
We want to broaden the range of applications that our technology can be applied to, tackling more cancer types in parallel, and ultimately cover the full cancer space, but understandably, this will take some time!
What other technologies are you excited to see in the future?
Automated image analysis. For understanding morphological patterns and supporting medicinal chemistry, it is possible to digitalise images to help with downstream analysis. However, tumours have complex shapes, and in hospitals, doctors are trained to look at them by eye. It is important to understand these images to see whether or not the tumour is aggressive or has metastatic potential. We are currently working with the National Physical Laboratory in Teddington to set up international measurement standards. There are certain parameters you can measure in tumours that might give us confidence to deduce which drug will work or which parameters are more important at offering insights into cancer biology. In this case, artificial intelligence and machine learning will help interpret and correlate large amounts of complex data and aid automated analysis. This technology is not unique to organoids, but if the technology is successful, it would make automated drug screening of PDOs much more efficient.
Interested to learn how organoids could be incorporated into your drug development programme? Download our new eBook: Organoids as Cancer Models.
Image credit: © BrAt82 - stock.adobe.com