Background:
The chimeric antigen receptor T-cell (CAR-T) and artificial intelligence (AI) derived new drugs have significantly improved the survival rates of hematological malignancies (HM) patients, but many HM are still drug-resistant. More importantly, choosing the best drugs for individual patient with lower side effects remains a huge challenge. The personalized treatment becomes increasingly important, especially for the rapidly growing elderly patient population.
Patient-derived xenograft (PDX) models are powerful tools for personalized treatment. However, traditional intravenous PDX models for HM often have low success rates (10%-40%), long establishment time (4-6 months), and do not adequately reflect patient's clinical characteristics. As a result, the traditional PDX remains largely a concept and fails to meet the clinical needs of most patients. Previous studies of us have proved that magnetically induced cells (MagIC) can promote the homing of various cells to the bone marrow, thereby facilitating the rapid reconstruction of the tumor microenvironment and the establishment of PDX (MagIC-PDX) models.
Objective:
To develop personalized therapy of drug-resistant HM patients, MagIC-PDX is combined with gene marking of malignant cells, creating an innovative and rapid technology for screening antitumor drugs.
Methods:
1.1 Model Establishment: On day 0, mononuclear cells, including primary malignant cells and microenvironment-supporting cells, were isolated from the bone marrow of drug-resistant patients, then labeled overnight with recombinant lentiviral vectors carrying luciferase and GFP. On day 1, the cells were magnetized at 37 degree Celsius for 30min by the “magnetic nanomotors”, then the magnetized cells were injected into the right femur of NCG mice guided by a precise overnight magnetic field (MagIC-PDX group), a conventional IV-PDX (IV-PDX group) was established as a control by tail vein injection.
1.2 Conventional Drug Testing: Once the stable tumor signals were detected in MagIC-PDX group (within 1 week) , clinical chemotherapy drugs such as doxorubicin, cytarabine, bortezomib, and dexamethasone were tested. Further assessment of their efficacy in killing tumor cells are done by bioluminescence imaging, flow cytometry, and pathological techniques. IV-PDX group was used as control.
1.3 Second Round of AI Assistant Drug Screening: If the PDXs showed drug resistant to the 1stround drug screening, a 2nd round of drug screening will be performed based on the MICM information of the patient and AI pre-selected potential drugs. Bone marrow cells of multi-round-drugs' resistant mice were frozen in cell banks for further mechanism studies.
Results:
2.1 Model Establishment: Most mice in MagIC-PDX group obtained stable and precise bioluminescent signals at the injection site on day 4, (95.3±6.5)% (n=5), and the signals expanded over time. All untreated mice in MagIC-PDX group developed HM significant clinical characters in the end. In comparison, the traditional IV-PDX group mice requires at least 3 weeks until the detection of unevenly distributed signals, and has a much shorter lifespan (28.4±2.7 vs 62.4±4.8)d, leading to a much smaller experimental window (14.4±2.7 vs 59.4±4.8)d than MagIC-PDX group.
2.2 Conventional Drug Testing and Second Round of Drug Screening: Based on the immediately diminished tumor load signals of myeloma in the MagIC group, bortezomib and dexamethasone treatment are chosen. However, those treated mice relapsed in 3 weeks, showing a developed resistance to the above mentioned treatments. Therefore, we implemented the second round of AI-designed CD4/TGF-β bi-specific antibody treatment, which showed good effect on inhibiting the development of bortezomib/dexamethasone-resistant myeloma.
Conclusion:
MagIC-PDX is a powerful new kind of PDX model, which allows fast drug screening and mechanism study. This research presents a pivotal tool for advancing personalized treatment.
No relevant conflicts of interest to declare.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal