The possibilities to predict oncological tumor response for individual patients with GI malignancies in order to tailor a personalized oncological treatment, are currently limited. This means that most of the information used to decide upon oncological treatment pertains to a group of patients rather than individuals. Similarly, most of the chemotherapy regimens used is not individually tailored although it has been shown to be beneficial for certain cohorts of patients. In colorectal cancer, KRAS (Kirsten rat sarcoma) and BRAF mutation status predict the response to drugs targeting various steps in the EGFR/KRAS/BRAF signaling pathway, whereas MSI/MMR status may predict the outcome of immune checkpoint inhibitors . As chemotherapy, as well as other cancer treatments, may have severe side effects the treatment may not only be ineffective but may also cause severe adverse events even shortening the patient's life and inhibit the chance of curative tumor resection. For this reason, pre-treatment response evaluation in an avatar model of zebrafish, would be most valuable.
From a clinical point of view the easiest situation is in the adjuvant setting. At this point the whole tumor is available for analysis and larger pieces of tissues may be used in the avatar model. 5-fluorouracil or capecitabine (a 5-FU prodrug) either alone or in combination with oxaliplatin is the current standard adjuvant treatment in stage III colon cancer and stage II colon cancers with 2+ risk factors. Other types of combination regimens including e.g. irinotecan have so far failed to show benefit in the adjuvant situation.
In the neoadjuvant setting, i.e. preoperative treatment to patients with resectable CRLM, combinations of 5-FU (or capecitabine) and either oxaliplatin or irinotecan may improve the outcome of surgery and long term prognosis. It is currently not known which regimen is most efficient for the individual patient.
In the most severe scenario, where the patient presents with unresectable CRLM, either palliative chemotherapy or downsizing/conversion chemotherapy is indicated. For the latter group of patients, it is of uttermost importance to provide as potent treatment as possible in the first line, as the 10
The therapeutic window is narrow and further clinical and/or radiological progression on the first line therapy may preclude further oncological and surgical interventions. Fit patients may benefit from a triplet of 5-FU, oxaliplatin, and irinotecan (FOLFOXIRI). However, this is a toxic regimen, and older or less fit patients are more commonly offered a doublet of 5-FU and either oxaliplatin or irinotecan. There may be an additional value when the chemotherapy is combined with antibodies targeting VEGF or, for those tumors expressing wild type KRAS, the EGF receptor.
But maybe the most valuable setting for the patient with upfront non-resectable tumor burden, would be before any treatment has started.
Up to 50% of patients with colorectal cancer will develop liver metastases either synchronous or metachronous in relation to the detection of the primary tumor. The mainstay of treatment today is a combination of resection +/-ablation of metastases, with or without preoperative liver volume augmentation and chemotherapy sometimes in combination with antibodies.
A number of negative prognostic factors have been identified both regarding the tumor growth but also in the pattern of oncogens in the tumors. Still, one of the most important factors is the response of the metastases to chemotherapy both radiologically but also in reduction of tumor markers (CEA, CA19/9)[5, 6].
Still we have very little knowledge if the metastases are going to respond to the therapy given or not. Tumor with KRAS mutation will not respond to Erbitux treatment, and mucinous tumors are usually more resistant to any chemotherapy.
Different models have been explored to evaluate tumor specific response to treatment, but most are slow or have shown mixed results. Today, mouse patient-derived xenograft (PDX) model is the most used and validated to predict response to therapy, but evaluation of oncological therapy takes months . Therefore, the mouse PDX model cannot be used for clinical decision-making. Organoid cultures using patient-derived cancers is a well-used in vitro screening tool, with promising results for different tumors. Organoids maintain the genetic characteristics of the original tumor tissue. Still, these models are slow when it comes to evaluating patients' tissue for treatment decisions and lacks the ability to evaluate metastatic or angiogenic potentials.
Recently zebrafish embryos have been used as avatars for human cancer, first in hematogenous cancers but lately also in solid cancers like PDAC, breast cancer and colorectal cancer[11-14].
The model has several advantages, the strongest is response evaluation only 3-5 days after the tumor tissue is implanted in the embryos. The embryos are transparent so tumor growth and spread can be visualized in detail and quantified in a semi-automatic manner. As the zebra fish embryos own immune system does not respond until day 8 days , why the immune response against the tumor comes from the patient's own immune cells, this is also the reason why xenograft engraftment works in this model. On the other hand. So far, biopsy tissue is not enough, but at least a cubic centimetre of tumor tissue is needed to generate these avatars. For this reason, needle biopsies are not providing enough material but resected liver tumor tissues are needed in the current protocols.
In zebrafish embryos, colorectal cancer cells have been shown to grow and response evaluation of different chemotherapy combinations have been successful with very good correlation to the response in humans. Clinical studies on colorectal liver metastases have so far not been performed.
Fish are housed at an average temperature of 28 °C in a recirculating system with a 14:10 h light to dark cycle. Zebrafish fertilized eggs are obtained by natural mating of wild-type AB strain at our facilities and the developing embryos are staged in incubator at 28 °C according to Kimmel et al. Before any procedure, embryos will be anesthetized in 0.02% tricaine. The tumor tissue taken from the surgical specimen by the histopathologist is cryopreserved by submerging the tissue in cryotubes containing cryopreservation medium, adding the tubes to a cryobox and placing the box in a -80 degrees freezer for at least 24 hours. The tube with tissue is then transported on dry ice to the zebrafish laboratory where the tissue will be thawed, washed and incubated in a protease mix within a mechanical disaggregation device (GentleMACS), to produce a single cell suspension from the tissue. The suspension is filtered and incubated for 30 min in 40 μg/mL CM-Dil in phosphate buffered saline (PBS), followed by centrifugation, and washing with PBS. The labeled cells are then resuspended in implantation medium and injected using thin glass capillary needles into the subcutaneous perivitelline compartment of 2-days old zebrafish embryos. Successfully implanted zebrafish embryos are then transferred to embryo medium containing the drugs under investigation (e.g. the drugs that could be considered for treatment of the patient donating the sample) and treated for three days. Images of the tumors are acquired both immediately after transplantation and following the three-day treatment period, and an additional image of the main metastatic site in the zebrafish embryo is acquired after the three-day treatment period to evaluate the metastatic capabilities of the cells. Treatment outcome will be evaluated as change in tumor volume between day 3 and day 0 in the treatment group compared to the vehicle control group (percent), and a significantly stronger reduction of tumor volumes in the treatment group will be considered as a positive treatment outcome. Evidence of a pro-metastatic phenotype will be derived from the number of cells found at the metastatic site in the zebrafish embryos three days after implantation. Cut-off levels for how many metastasized cells are required to predict likely metastatic progression in the patient will be evaluated in the first 10-20 patients included as such information is not currently available from other sources of information.