Patient-derived organoids in hepatobiliary pancreatic cancer research: Their uses and limitations

Abstract

In this letter, we discuss the highlights of the paper by Hu et al, including how patient-derived organoids may be beneficial to hepatobiliary pancreatic research. The article provides a review of how organoids can be used in drug sensitivity testing; looking at ways in which successful organoids are created. The literature included in the review revealed heterogeneity in organoid establishment including some differences between organoids from resected tumours compared with liquid biopsies. Additional research is required in creating organoids from liquid biopsies and optimizing these techniques for widespread clinical practice. The article raises awareness of limitations of organoids with suggestions of how co-culture or microfluid platforms may help to simulate the tumour microenvironment for better model fidelity. The article provides a comprehensive review of how organoids are being used in drug testing and ideas about how to harvest or produce these in future.

Key Words: Patient-derived organoids; Organ-on-a-chip; Sensitivity testing; Co-cultures; Tumour microenvironment

Core Tip: The development of organoid models is becoming an exciting field in in-vitro drug testing for chemotherapy for aggressive cancers, such as hepatobiliary pancreatic. The most reliable source of organoid establishment is from resected tumour; however liquid biopsies are becoming more common. There is a large amount of heterogeneity in the way organoids are cultured and there are drawbacks of this technique, such as the lack of fidelity in creating the tumour microenvironment and other variables that affect tumour growth.

TO THE EDITOR

It was a pleasure to read the article by Hu et al[1], which provides insight into the latest updates and applications regarding patient-derived organoids (PDOs) in hepatobiliary pancreatic (HPB) cancer. The paper provides a comprehensive overview of organoid success rates and then using organoids for chemotherapy selection. It provides exciting possibilities of identifying patient tailored drug regimens rather than simply following empirical guidelines. The review focuses on the limitations of traditional organoid models and how co-cultures or microfluid chip platforms may be better at replicating the tumour microenvironment. Like with all in-vitro research, there are a range of pros and cons relating to the effectiveness of organoids in translational medicine. We would like to comment on a few aspects of Hu et al’s timely review article[1].

Heterogeneity in organoid success rates

One key aspect of the article was to discuss the variability of successful organoid establishment. For hepatocellular carcinoma (HCC) success rates varied in the studies ranging from between 26% to 75.6%, with more than half of the studies having success rates below 50%[1]. Similarly, pancreatic tumours and biliary tumours had varied success rates, often lower than HCC, albeit improving with time. There are many reasons for varied organoid success, including unsuitable isolation of cell type. This has been highlighted as quite common in HCC, although not surprising due to heterogeneity in liver cells[1]. The literature highlights these have improved with better microscopic isolation and detection of specific cell markers, including use of adhesion molecules[2]. Data on the methodology for culturing organoids would also help with improving organoid success. It is important to understand how to optimise the process of culturing organoids by comparing growth mediums, the number of passages and whether antimicrobial therapies were given. This has been documented for other types of cancer, although more specific information on culturing methods would be useful in HPB[3]. It is only by optimizing these techniques that this technique can be introduced into widespread clinical practice.

Solid vs liquid biopsies

Tissue harvesting including site of cells and amount of tissue taken is also important for establishing successful organoid cultures. The literature highlights how resected tissue provides the greatest number of cells and success rate in establishing organoids; however, clinicians are becoming more inclined to research liquid biopsies (LBs) like blood or bile as it is less invasive[1,4,5]. Indeed, the article touches upon the use of ascitic fluid for harvesting malignant cells, although much more research is needed in this area to optimise cell collections[1]. The data table comparing organoid establishment success and the site of biopsy was useful, nonetheless could have been improved by specifying the different types of LBs being performed and for what cancer[1]. Some LBs not involving HPB like urine have been highlighted to be of great use in bladder cancer organoid creation[6]. There is, however, limited data highlighted in this paper on bile for detecting cells in cholangiocarcinoma[1]. Work involving LBs has become increasingly common in cancer research for detecting extracellular vesicles and circulating tumor DNA, although limited studies show successful cell isolation[7,8]. This is not only due to lower cell numbers but also a lack of successful technique and heterogeneity in cell types[8]. Once we have improved these techniques, data comparing the genetic stability between cells from resected tissue and LBs will be an important area of research for determining LB use.

Pharmacogenomics and drug sensitivities

Precision medicine or pharmacotyping is one of the main aims of using PDOs. Using a range of genetic analysis techniques like next generation sequencing and whole exome sequencing, oncologists can assess the suitability of comparing the validity of PDO drug sensitivity testing for primary tumour response[1]. The discussion about how HPB tumour cells confer long term stability is an important one, and the article by Hu et al[1] highlights some of the previous work done for liver and bile PDOs, including their genomic stability compared to primary tumours[1,9]. Some studies have shown that for pancreatic adenocarcinoma, there is high genotypic similarity between PDOs and the primary tumour[10-13]. Indeed, one finding was that around 93% of small nucleotide polymorphisms in pancreatic cancer were retained in the PDO[5], although this is for one study. The literature has suggested that frequent passages or exposure to treatments may compromise the long-term genetic stability of these models. However pancreatic PDOs can maintain genetic stability after several passages[11-13]. The article highlights that for precision medicine to work effectively PDOs would need to retain their genetic stability. More work is needed to identify how to improve genetic stability in PDOs.

Limitations of organoids

The article raises awareness about the limitations of organoids due to failure to represent the tumour microenvironment. As per the method mentioned in the paper[1] normally one type of tumour cell is isolated and grown. Whilst this is useful it does not fully consider the crosstalk between cancer cells, stromal cells, immune cells and vasculature surrounding tumours which are increasingly recognised as being an important factor in tumour biology[14]. The literature addresses how the use of mesenchymal stromal cells, peripheral blood mononuclear cells, cancer associated fibroblasts and PDOs for HCCs provide greater fidelity for testing anti-angiogenic agents and immunotherapies[15]. The limitation of chemotherapies in HCC was documented and the need for treating other targets[15]. Creating a suitable tumour microenvironment will also help with considering how specific cytokines shape cancer development[16]. Some limitations of co-culture, however, include failure to simulate the layout and space between different tissues. Co-culture may also not consider other systems in the body and the effect this can have on both pharmacokinetics and pharmacodynamics. Taking a biopsy, however, may be more favourable for the more aggressive, less predictable cancers where empirical regimens may not always be efficacious. Potentially making the results of these important trials publicly available and testing more PDOs will help with drug predictivity research.

Overall, Hu et al[1] provides a timely and important account of how organoids are currently being used and how co-culture and microfluidics may be more favourable when creating a high-fidelity environment with a focus on HPB cancer. However, the techniques and process will be applicable to other gastrointestinal and other tumour types. PDOs are predicted to be a potential game-changer and these initial results are extremely promising in terms of improving success of organoid establishment and genetic similarity of the harvested cells. More work is still needed with analysing the sensitivity and viability of LBs, as well as improved methods of correlating organoid research with clinical trials.