Oncology providers understand that each patient's response to therapy may vary depending on several factors, ranging from any preexisting conditions to the stage and grade of cancer. This makes determining optimal treatments incredibly challenging and often leads providers to test their effectiveness without guarantee.
In an attempt to create a new, more personalized approach to cancer treatment, a team of researchers from the University of Geneva (UNIGE) evaluated a theory that organoids having been exposed to various therapies can shed light on colorectal cancer (CRC) patients' treatment response. Referred to as "tumor avatars" these untreated 3D cellular constructs are derived from the patient's tumor and hypothesized to model how the patient would respond to specific therapies.
The study combined ex vivo organoid efficacy testing with the mathematical modeling of results to design personalized CRC patient treatments (J Exp Clin Cancer Res 2023 doi: 10.1186/s13046-023-02650-z). Led by Patrycja Nowak-Sliwinska, PhD, Associate Professor in the School of Pharmaceutical Sciences at the Faculty of Science of the UNIGE, and a member of the Translational Research Centre in Onco-Hematology, the UNIGE team initially began this research based on the direct needs of the CRC patients.
According to Nowak-Sliwinska, her research has always been positioned around the clinical needs, "from patients' bedside to bench and back." These needs were exacerbated in the case of CRC patients, as CRC is a relatively common tumor type with well-characterized sub-populations of patients.
"CRC is a very heterogeneous cancer, meaning that each person with this kind of tumor has different characteristics. The problem is that within the clinics we have a standardized treatment protocol for each stage of the disease," explained George M. Ramzy, PhD, who began his work with Nowak-Sliwinska during his master's project and after completing his PhD in her group, continues working on this research as a postdoctoral fellow. "After a while, this protocol actually fails to cure patients, especially with late-stage disease. Our idea was that for each of these patients, we needed to establish a personalized treatment to make sure that it would actually work."
Demonstrating how the concept of tumor avatars helps make this possible, Ramzy described the process to acquire one of these organoids. When a patient undergoes surgery at the hospital, a small part of the removed tumor is cultivated and given the same conditions as if it were inside the patient's body. The organoids have very similar architecture and characteristics to the tumors themselves. This is what creates an extremely personalized model for the researchers to use to test treatments within the lab.
"The [therapy testing research] process was multi-dimensional because first, we had to understand the problem within the clinics [resulting from the need for] different treatment strategies. In the clinics, the standard of care for CRC is mainly based on a combination of chemotherapeutic drugs. Such therapy is toxic and ineffective after a while due to the emergence of resistance. In our research, we [focused on other] types of drugs, [such as] small molecules targeting specific entities in the cancer cell," noted Nowak-Sliwinska, who began this research line over 10 years ago, but fully developed it to the current stage at the University of Geneva.
She noted that the question became how to find quality drug combinations, as an almost infinite number of candidates exist. This meant having to define the initial set of blocks that could include 11 or more drugs, and then find ways to optimize them to result in treatments that include only three or four drugs.
"For this, we developed a statistically based method that allowed us to model the results obtained in the lab," Nowak-Sliwinska said. "Using mathematical tools, we select drugs that are synergistic with each other and, based on that screen, we can find the optimal drug combinations. It's a hybrid of laboratory testing of selected product combinations and in silico modeling."
Clinical Research Methodologies
The steps to determine treatment effectiveness first included the isolation of patient-derived material in the lab (within 2 hours from the surgery), followed by the establishment of the tumor avatars in the lab. Next was the definition of an initial drug set to be tested, followed by treatment being administered with predefined drug combinations and incubation with tumor avatars for 72 hours. Then, the results readout and data modeling took place and, based on this data, the team selected the optimized drug combination (ODC) and began to look for certain mechanisms and action patterns.
"What's interesting is that, for each patient, these tumor avatars are specific to them, as the optimized drug combination is. So if we compare different avatars from different patients, you will see that architecture-wise, characteristics-wise, and genetic-wise, they're completely different," Ramzy explained.
During the definition of the drug set, the research team's methodology consisted of using the validated phenotypic approach called Therapeutically Guided Multidrug Optimization (TGMO). This approach allowed the researchers to identify four low-dose synergistic ODCs in 3D tumor avatars. The models were identified as either sensitive or resistant to first-line CRC chemotherapy (FOLFOXIRI). Second-order linear regression and adaptive lasso were then utilized to obtain study results.
Developed by Nowak-Sliwinska's lab technology, TGMO merges minimal experimental testing and mathematical modeling to identify the most optimal treatment strategies, ideally composed of drug combinations. The team performs the tests and can rapidly identify the treatments that are effective and safe (non-toxic). They can also identify those that are ineffective (and toxic), therefore eliminating the possibility of treating the patient with something that would not help.
"[CRC patients have been] mostly treated with chemotherapy. So what is the bottleneck here? The treatment will not be working for long because people get resistant to treatment, stop responding, and have very important side effects," Nowak-Sliwinska stated. "Our drug combinations are specific to the patient so that they can potentially be treated longer with better activity and better safety. That [can cause] major changes in the life quality of the patients and hopefully overall survival in time."
Further, she noted that the drug selection used during this research was already being used in the clinic, meaning that nothing new was induced in the patients. Instead, the research involved finding new ways to optimize existing treatments in a way that is more beneficial to patient outcomes. This also means that the cost of potential clinical use is likely reasonable when compared to de novo drug development.
CRC Tumor Avatar Results
Overall, the study resulted in the identification of patient-specific TGMO-based ODCs that were proven to outperform the efficacy of the standard-of-care drug combination FOLFOXIRI. Other findings showed that all ODC activity was validated on patient-derived primary or metastatic CRC organoids.
Using whole-exome sequencing and RNAseq, the researchers' CRC material was molecularly characterized and exhibited that alternative treatments increased cell viability. Specifically, ODCs consisting of regorafenib [1 [mu]M], vemurafenib [11 [mu]M], palbociclib [1 [mu]M], and lapatinib [0.5 [mu]M] were shown to inhibit cell viability up to 88 percent in patient-derived organoid from patients with Stage IV liver metastases (identified as CMS4/CRIS-A) (J Exp Clin Cancer Res 2023 doi: 10.1186/s13046-023-02650-z).
"The organoids technology is not new and has been developed by many groups testing it over the past 10 years," George M. Ramzy said. "However, combining the short-term organoid platform to our drug screening method is innovative because we are able to optimize specific drug combinations on them within a clinically relevant time frame, meaning it's the time between the diagnostic or the surgery to the treatment decision."
He explained that, within 15 days from the day of surgery, they are then able to discuss tumor avatar results with surgeons and oncologists. According to the data, they can relay which combination of drugs is synergistic. Also important to note, the researchers employ dosages below the clinically used doses to ensure they are safe and establish a patient-centric approach to therapy.
This type of insight is significant in regard to the effectiveness of CRC treatment, as it is known to vary greatly from one patient to another. Chemo and other targeted therapies provide only a limited increase in overall survival for this patient population, with significant side effects, including potential drug resistance at the cellular and tumor levels (World J Gastroenterol 2018: doi: 10.3748/wjg.v24.i34.3834). Other possible side effects from chemotherapy used to treat CRC include the rapid clearance of anticancer drugs, low efficiency, and considerable toxicity of common anticancer agents when prescribed at higher doses (Pharmaceutics 2022; doi: 10.3390/pharmaceutics14061213).
"Our method rapidly identifies the most effective and safe treatment options that can be personalized for each patient," Nowak-Sliwinska said. "Of course, this is just a model. What we want to do is...get as much information as we can before we [begin to test therapies on] humans. Each model has its limitations, so while this model does not fully represent a living tumor, it's the closest in vitro model available involved before beginning to test on animals."
Ultimately, not having to test treatments directly on patients-or through animal experiments-would be a major advancement in the world of medicine. Additionally, if the tumor avatar models developed by researchers at UNIGE prove that, when isolated from a patient's tumors that can be used to test the activity of various treatments, this information could revolutionize how providers go about selecting optimal treatment options for their patients. Currently, the team plans to continue their approach to this research on a bigger scale including other tumor types.
"If we could extend [our research] to other cancer types, I think that will be super beneficial for the patients. This is what we hear from oncologists and patients; this is the future they want to see," Nowak-Sliwinska said. "We are currently doing the same for kidney cancer already, and the next step will be [to assess the avatar's impact on treating] breast cancer."
Lindsey Nolen is a contributing writer.