	Director, UC Davis Comprehensive Cancer Center

The new generation of therapies targeting specific mutations in tumors, while promising, has delivered underwhelming results. The problem is that while genomic analysis can uncover various “druggable” abnormalities, we have no immediate way to determine which abnormality is actually driving the cancer. This approach has resulted in a meager 12% success rate with a few exceptions.

At UC Davis Comprehensive Cancer Center, we are working to make personalized treatment more precise through a collaborative research effort with the NCI-designated Jackson Laboratory headquartered in Bar Harbor, Maine, with operations in West Sacramento, California.

In August, we published a proof-of-concept paper in 

 that described our research using tumor xenografts in mice to test and identify more precise treatments for bladder cancer patients.

Part of the problem, in bladder cancer treatment, is that while combination chemotherapy can be effective for up to half of all patients with invasive and advanced disease, no other FDA-approved therapy is available when first-line therapies fail. Other therapies are occasionally used, but, in general, less than 20% of patients will benefit. So far, no test can identify potential responders before treatment.

We used the patient-derived xenograft (PDX) platform to test various treatments in mice prior to treating the patient. We obtained bladder tumors directly from individual patients, grafted them into mice at Jackson Laboratory, and identified actionable mutations through next-generation sequencing. We developed as many mice carrying patient-specific xenografts as needed to simultaneously test multiple therapies.

Interestingly, this PDX platform is patient-specific. Cell lines and their derived xenografts have been traditionally used to test drug efficacy. These cells, however, are homogeneous and genetically different from patient cancers. After the cells are cultured, their gene expression profile is dramatically different from those of parental tumor xenografts.

These changes cannot be reversed after re-implantation. Therefore, it is not surprising that prediction models for drug response based on the genetic information of cell lines, such as the Genomics of Drug Sensitivity in Cancer and the Cancer Cell Line Encyclopedia, frequently fail to predict drug efficacy in the clinic. Our PDXs, not only retained the morphology of their parental cancer cells, but also showed remarkable genetic fidelity—between 92%—97%&mdash;of their original patient tumors.

There are many applications of this PDX platform in the era of precision. First, we determined its application in identification of effective chemotherapeutic drugs. The GC (gemcitabine and cisplatin/ carboplatin) regimen is commonly used as a first-line therapy in bladder cancer. We determined the PDX sensitivity to cisplatin, gemcitabine, and the drugs in combination. Of the first six PDXs we tested (10-12 mice per treatment group), five were resistant to cisplatin and two were resistant to gemcitabine. Chemo-resistance to one drug could be overcome by the other drug, leaving four of the six PDXs sensitive to this G/C combination. Our findings suggest that, even though cisplatin and gemcitabine are commonly used in combination, many cancers respond to one drug, while the other has little effect on the cancer, but causes toxicity.

To demonstrate proof of principle, we also used this PDX platform to screen for effective targeted therapy. One of our PDXs has three “druggable” genetic aberrations: 

. Consistent with a previous report showing that only 12% of matched targeted therapies were effective, only the PIK3CA inhibitor efficiently inhibited tumor growth, while the other two inhibitors were not effective at all, but caused toxicity. To study the mechanisms of resistance to targeted therapy, we performed a serial biopsy during treatment of one PDX with an FGFR3 inhibitor, BGJ398. This drug inhibited the growth of one PDX that overexpressed FGFR3 as detected by both RNAseq and immunohistochemical staining. Eventually, that xenograft regrew and became resistant to BGJ398.

Serial biopsies, together with western blots, revealed that the MAPK/ERK and PIK3CA-AKT pathways were activated upon the development of resistance. Inhibition of these two pathways reversed the resistance.

We also showed that this PDX platform can be used for drug repurposing. The EGFR/HER2 dual inhibitor lapatinib is FDA-approved for the treatment of breast cancer with ERBB2 gene amplification or overexpression. In general, this drug is ineffective in bladder cancer. However, in our bladder cancer PDXs, one of two PDXs that had overexpression of ERBB3 (3+ by immunohistochemical staining) responded to lapatinib.

The PDX platform also can be used for drug development. For this study, we determined the efficacy of bladder cancer—specific nanoparticles we developed. These nanoparticles are coated with the cancer-specific targeting ligand PLZ4 (amino acid sequence: cQDGRMGFc) on the surface for specific targeting to bladder cancer cells and loaded with therapeutic and/or diagnostic agents in the core. These PLZ4 nanoparticles loaded with paclitaxel, a salvage therapy, significantly decreased paclitaxel toxicity and increased efficacy.

To begin translation into clinical applications, we compared the response of one patient’s cancer and its derived PDXs with three lines of therapy. Consistent with the fact that her cancer was refractory to the dose-dense MVAC (methotrexate, vinblastine, doxorubicin/Adriamycin and cisplatin), her PDX also was resistant to cisplatin treatment. Subsequently, she was treated with two lines of therapy based on the study results in her PDXs: an anti-EphB4 targeted therapy and the gemcitabine/carboplatin combination. The response of her cancer and PDXs were exactly the same to both lines of therapy. The surprising findings in her cancer and PDXs are that both were resistant to cisplatin/carboplatin and gemcitabine when these drugs were administered as single agents. However, surprisingly, both her cancer and PDXs responded to the G/C combination.

While grafting patient tumors for further study holds promise, our ultimate goal is to be able to biopsy patients, sequence their tumors, and feed that information into a computer that can then determine which treatment is likely to be effective. Even though the PDX platform is promising in the era of precision medicine, a major drawback is the low engraftment rate. In our study, only 41% of patient cancer tissues implanted eventually formed xenografts. Implantation at the subrenal capsule can potentially increase the success rate. Other limitations include a long lag time to establish the first PDX (called P0) from patient cancer tissue (4—5 months) and high cost—$6000 to establish the model and an additional $25,000 to run a four-arm study to test three drugs.

Articles

Researchers at the UC Davis Comprehensive Cancer Center are personalizing treatment using tumor xenografts in mice to test and identify more precise treatments for bladder cancer patients.

Ralph de Vere White, MD

Articles

Patient-Derived Xenografts Could Improve Precision Medicine Outcomes