NEW YORK – In a preliminary data readout from an ongoing Phase I trial, researchers reported notable anti-tumor responses among heavily treated metastatic castration-resistant prostate cancer patients on Poseida Therapeutics' autologous CAR T-cell therapy P-PSMA-101.
The data, which San Diego-based Poseida presented Tuesday during the CAR-TCR virtual summit, were encouraging for the future prospects of P-PSMA-101, despite the analysis being based on the experience of only nine mCRPC patients and on surrogate efficacy measures using prostate specific antigen levels.
"We believe [these responses] are among the best ever described for a CAR T therapeutic in a solid tumor indication," Poseida's CEO Eric Ostertag said in a presentation of the data. Five patients had measurable declines in PSA levels, three patients had a greater than 50 percent reduction in these levels, and one patient saw his tumor completely disappear and continues to respond to treatment five months after infusion with the autologous product.
The Phase I trial, which is designed to assess the autologous treatment's safety, maximum tolerated dose, and preliminary efficacy in up to 40 mCRPC patients, is still ongoing at multiple sites across the US, and Poseida anticipates that the next readout will occur during the first half of 2022.
The primary outcome of interest in the trial is overall response rate, defined as the percentage of participants who have a partial reduction or see their cancer completely disappear on P-PSMA-101. At this early stage, the company didn't report an overall response rate.
Ostertag pointed out that overall response rate, which requires imaging tumors and measuring their size, is more difficult to gauge in mCRPC than other solid tumor types. Although imaging and RECIST criteria are commonly used to track how a tumor is responding to treatment, these modalities work well when tumors occur in soft tissues, he explained.
"The great majority of patients with mCRPC don't have just soft-tissue lesions," he said. "Many have only bone marrow involvement." For this reason, Poseida focused mainly on PSA decline as a measure of its treatment's preliminary efficacy.
Although PSA, measured via blood testing, remains the most well-known biomarker for clinical staging in prostate cancer, in recent years, researchers have also been turning to prostate specific membrane antigen as a biomarker for tracking uptake and response to radio-ligand therapy using PET imaging. At trial sites where PET imaging was available, investigators collected PSMA data from PET scans, including FDG- and PSMA-PET, but not all sites had that capability.
"We're moving to PET imaging, which I personally think is going to replace or supplement the existing measures in prostate cancer," Ostertag said.
Among four patients who received PSMA-PET imaging, the results of the imaging tests were concordant with the PSA decline, Ostertag noted, acknowledging that historically, PSA has not always been the most foolproof biomarker to use as a surrogate endpoint. "That [concordance] doesn't always happen, but it has in our clinical trial," he said.
Additionally, investigators collected bone marrow biopsies from patients when available and tested them for tumor markers such as pancytokeratin and NKX3.1 using immunohistochemistry. Again, the results were concordant with the PSA declines.
In his presentation of the trial's preliminary data, Ostertag highlighted two patients' responses to the CAR T-cell therapy. P-PSMA-101 completely eliminated one patient's tumor based on rapidly declining PSA levels from a baseline of nearly 40 down to 0.2 nanograms per milliliter at around 40 days post infusion. This patient's response according to PSMA PET and FDG PET imaging were concordant with PSA measurements; PSMA levels declined by more than 70 percent at four weeks post infusion and declined even further at 12 weeks.
Additionally, a bone marrow biopsy taken eight weeks post-CAR T-cell therapy from a site where the patient had previously had a tumor showed no tumor by IHC measurements. Five months later, Ostertag shared, this patient is still responding to treatment.
In a second case, 40 days after CAR T-cell treatment, the patient's PSA level dropped by about 75 percent. The PSA drop occurred at the same time as what Ostertag described as a "robust expansion" of the P-PSMA-101 cells. This patient also showed significant improvement in PSMA PET imaging and a bone scan after four weeks post-infusion. A bone marrow biopsy at the tumor site also showed dramatic reduction in tumor burden and T-cell infiltrate.
The safety profile for P-PSMA-101 observed among the nine patients, according to Ostertag, was also encouraging. Investigators observed three cases of possible cytokine release syndrome, but all three were low-grade and managed with early intervention. There were no cases of neurotoxicity.
The news of the encouraging safety profile was particularly welcome to Poseida after the US Food and Drug Administration ordered the firm last year to pause the Phase I trial to investigate a patient death. The patient, Ostertag explained, had died of macrophage activation syndrome "exacerbated by non-compliance." According to Poseida, the patient had been at home following infusion and did not come back into the hospital when he experienced symptoms until he was already in acute liver failure and it was too late.
After this incident, Poseida amended the clinical trial and began requiring patients to be hospitalized for up to 14 days following infusion to keep a closer watch on adverse events. Ostertag explained that that is the time window within which the cells expand, and patients tend to experience cytokine release syndrome. Previously, the treatment was being performed on an outpatient basis. Following the changes, the FDA lifted the hold, and Poseida resumed the trial in November 2020.
Non-viral manufacturing system
According to Ostertag, one of the reasons that P-PSMA-101 has led to significant tumor responses in solid cancers, whereas other autologous CAR T-cell therapies have not, is that the firm takes a unique, non-viral approach to engineering the autologous cells to target cancer cells.
Instead of using viral vectors to engineer the product, Poseida uses a DNA transposon system, dubbed PiggyBac, in which it delivers a therapeutic transgene.
"Compared with viral-based systems, the PiggyBac system offers a number of advantages including very efficient gene delivery into TSCM cells," he said. These TSCM cells, or stem cell memory T-cells, have been identified in recent years to have improved therapeutic persistence versus other T cell types. "PiggyBac prefers to insert your therapeutic transgene into TSCM. … You can't do that with a virus."
The reason for this, Ostertag continued, is that when you use a viral vector in CAR T-cell manufacturing, you first have to infect the cell to activate it, and then the virus transduces DNA into the genome. The infection step requires having certain cell surface molecules on a T cell in order for the virus to get in. With the PiggyBac technology, on the other hand, there is no need to activate the cells first; the process instead involves DNA and RNA going straight into the cell.
Further development
Currently, Poseida is manufacturing these products through a contract development and manufacturing organization, but the firm has recently built a pilot plant capable of producing good manufacturing practices-grade material for clinical trials. The plant is up and running but is currently being used primarily to produce an allogeneic CAR T-cell therapy that Poseida is studying in multiple different solid tumor indications, dubbed P-MUC1C-ALLO1.
The firm is planning a basket trial to evaluate that therapy and has not yet determined whether it will use a biomarker-selected approach and enroll only patients whose tumors express MUC1C, since in certain tumor types, such as breast cancer and ovarian cancer, the protein is expressed on nearly 90 percent of cases.
As for P-PSMA-101, Poseida is continuing its Phase I clinical trial but beginning to set its sights on the next Phase I/II study, which could be registrational.
"Given the already very good anti-tumor efficacy and the tolerability we're seeing at these low doses, it's possible we've already found a dose that would be appropriate for a Phase I/II, possibly registrational study," he said. "But we're still escalating … because you don't know until you try."