Alex Moreau, PhD, from Champions Oncology answers key questions about Drug Development Strategies

1) Do you have patient response correlating to the PDX?
Yes. We demonstrated that PDX models effectively capture responses to oncology therapy in a heterogeneous cohort of patients with solid tumors across multiple indications. This landmark analysis was published in Annals of Oncology in 2017 (Izumchenko et al.). We compared whole-exome sequencing of 237 PDX models to equivalent information in The Cancer Genome Atlas database, revealing that TumorGrafts^® faithfully conserve genetic patterns of the primary tumors. We next screened PDXs established for 92 patients with various solid cancers against the same 129 treatments that were administered clinically and correlated patient outcomes with the responses in corresponding models. Our analysis demonstrates that PDXs accurately replicate patients’ clinical outcomes in over 85% of the cases tested (nearly 100% in some cancer types), even as patients undergo several additional cycles of therapy over time, indicating the capacity of these models to correctly guide an oncologist to treatments that are most likely to be of clinical benefit.

Garima Kaushik, PhD, from Champions Oncology answers key questions from her Virtual Poster Talk - Development of a 3D organoid autologous TIL co-culture platform for IO studies 

1. What is the ratio N of organoids vs N of TILs?
   We show infiltration and resulting tumor death in three different organoids: TIL ratios: 1:2.5, 1:5 and 1:10. Our Keytruda treatment experiments shown here were performed in organoid: TIL ratio of 1:10.
   

Veronica Sacchi, PhD from PhenoVista Biosciences answers key questions from her Virtual Summit Presentation - Novel applications for Ex-vivo PDX models of the Tumor Microenvironment

1. Do you use a Matrigel or other hydrogel in setting up the models?
   The fragmentation process we use maintains the 3D nature without a hydrogel in the majority of models. Having a hydrogel matrix for embedding the tumor fragments would increase costs, reduce throughput (as we’d need to prepare, place, and obtain larger Z-stacks), and complicate assay development (e.g., co-culture with immune cells). Our hydrogel-free setup therefore provides the most flexibility for further development. We have optimized media to support good cell viability in the majority of tumor models tested in the platform. We have also accommodated custom media requests to allow for hypothesis testing (e.g., defined media), have run sequential treatments with washouts, etc.
   
   

Lynn Biderman, PhD, from Stelexis Therapeutics answers key questions from her Virtual Summit Presentation - Models for evaluating cancerous stem cell-targeting therapeutics

1. Is Stelexis TumorOrigin platform focus on any specific cancer indication and how do you confirm maintenance of clinical relevance (as stated in slide #5)? 
   The TumorOrigin platform’s early work focused on hematological malignancies, in particular AML and MDS. In the spirit of demonstrating that the cancer stem cell phenomenon is valid in other indication, the company is currently developing the platform in solid tumor indications, starting with lung and pancreatic cancers. We confirm maintenance of clinical relevance by multiple ways: Characterization of stem and cancer markers before, during and after expansionMutational analysis of expanded TumorOrigin cells comparison to patient cell clinical sequencing dataAnalysis of clonality and heterogeneity as demonstrated by molecular profiling of single cell colonies  
   
   
2. You mentioned that your target is not expressed in mouse – can you talk about the pros and cons of using murine model to study such target?
   
   Cons
   No ability to study the effects of non-tumor-derived secreted Target proteinEven though the receptor is present in the mouse, the pathway may be wired differently since the Target protein is not present in miceAs a result, our ability to fully appreciate the complex cross-talk between the neoplastic cell and the microenvironment is limited 
   
   Pros 
   PDX studies enable modeling disease, especially if the MOA involved relies on the Target protein being secreted by the cancer cellThis enables us to focus on the effect of the secreted Target protein that is derived from the cancer cells, separately from stromal-derived protein, aiding our understanding of MOAThe immunodeficient model allows the engraftment of both tumor and different human immune cell subsets. The dose, location, and time of implant and treatment can be controlled to address different stages of tumor development
   
   
3. How do you explain the elevation in mCD45+ cells despite murine lack of target expression?
   We don’t really know, but we are intrigued by the observation and are following up on it. The way we think of it now is as a “real estate” issue - in the combination therapy not only are the human CD45+ AML cells reduced, but also there is potentially less inflammation in the marrow, allowing for the normal mouse immune cells to re-populate the niche.
   
   
4. What was the sample size of mice used per cohort in these AML PDX studies?
   We had three mice per treatment group, and five mice in the control group.
    
5. Which CSC markers do you look at while sorting the cells? 
   Markers used in sort include: lineage markers, CD33, CD34, CD38, CD45RA, CD123 and IL1RAP
    
6. How do you envision the therapeutic efficacy of your test agents to be in humanize/immune competent murine models, since you are targeting inflammatory pathways? 
   Our target protein is not expressed in mice and while this may provide some limitations to study inflammatory pathway changes, it can also be an opportunity to modulate and control more aspects of the disease model. The effect of target blockade can be assessed on either the tumor only component and/or the immune cell impact. We envision the humanized models as valuable tools to allow us not only determine efficacy, but also how modulation of the pathway can be further interrogated in order to broaden our mechanistic understanding of the target.
   
   

Natasha A. Karp, PhD, from AstraZeneca answers key questions from her Virtual Summit Presentation - Embracing variation: a multi-batch study delivers robust estimates of efficacy

1. How do you know what variation to introduce to increase reproducibility?
   This is a tricky issue. We know that introducing variation improves reproducibility but we anticipate that each scenario will have its own variables that are interplaying with the biology. Some scientists have published success stories for improving the reliability by introducing variability such as circadian (am and pm) or multi-lab but others have shown little improvement (environmental enrichment variation). This is why I suggest the multi-batch approach as a first step to improving reproducibility for your lab with defined meta environment because this is achievable and ensure the conclusions from your testing space are robust and truly add to the knowledge base.
2. Could this approach work with a tumour volume on a day rather than the rate of growth?
   Yes, you can use tumour volume on a day as the variable of interest. However, tumour volumes on a day are often effected by loss of data in a biased fashion. If you have lost readings due to hitting the welfare limits on size or ulcerations, this isn’t loss in random as it will be the larger tumours.  Consequently, processing the remaining data will lead to bias in your estimates.  If you have a complete dataset then this isn’t an issue.  The other thing to consider is that the growth rate is estimated from multiple measures and is consequently more precise and this can increase sensitivity.
3. Is the growth rate metric tolerant to tumours that have a delay in treatment effect or a regrowth?
   The growth rate metric is based on the idea that tumors grow in a multiplicative fashion.  I.e. you go from 2 cells to 4 and from 4 to 8 cells and so on.  If you log transform the tumour burden measure you should then end up with a linear relationship between the measure and time.  If you then calculate the slope of this relationship for each animal this gives you the average growth rate. A strength of this metric is that it doesn’t have to fit the profile well as it will just return the average growth rate. However, when comparing the growth rates between groups you should be selecting a similar time period and this will give the hypothesis that you are testing boundaries eg  is there a differences in the average growth rate between the control and treated after dosing up until day 40. If you are interested in regrowth only,  you can select the time period when dosing halted and compare the average growth over this period between the groups of interest.  If there is a delay in the treatment having an affect you can select a  time range that only includes data after the mechanism of action kicks in and this will increase sensitivity to see an effect.

Joel Kaye, PhD, from Ayala Pharmaceuticals answers key questions from his Virtual Summit Presentation - Targeting Notch Altered Patient Derived Xenografts (PDX) with AL101. 

1. How does the effect of AL101 in the ACC PDX models compare to other Gamma Secretase Inhibitors?
   Some other GSIs have been tested in the ACCx9 model, and based on publications it seems AL101 is the most potent. However, we have not tested head-to-head in this model.
2. You showed there were 14 TNBC models that are Notch activated, but only data in 9 of them. What about the rest?
   That work is ongoing and we will update at the next scientific meeting.
3. Are there any updates on the clinical studies?
   The latest data from ACCURACY was published at ESMO2020, and the TENACITY study in TNBC was just recently initiated.
4. Is there a good combination strategy planning for AL101 in TNBC? 
   We will use the RNA-Seq data from the PDX models to guide rationale design of combinations, and that too is ongoing work.

Advances in oncology research have led to the development of personalized treatments based on specific knowledge of a patient’s tumor. New therapies have been customized to target signaling pathways that are hyperactivated or block specific variants of cell surface molecules, thus leading to better anti-tumor responses. Next generation sequencing (NGS) technology has been at the forefront of these breakthroughs by enabling researchers to rapidly sequence RNA transcripts (RNA-seq) or exons (whole exome sequencing; WES) within tumor tissue and translate these findings into novel therapies.

Ask the Expert: Mike Ritchie, PhD, MBA, answers key questions about our new bioinformatics platform, LUMIN, and using this incredible tool to build oncology biomarker programs, from his recent webinar.

Ask the Expert: Dr. Chris Koehler, from our partner, PhenoVista Biosciences answers key questions from around ex vivo imaging and his recent webinar.

Ask the Expert: Dr. Pravesh Gupta answers key questions from his ongoing myeloid cell research and recent webinar.