DR MARK EVERS (Lexington, KY): Dr Fong’s laboratory is a leader in the use of oncolytic viruses as potential therapy for a variety of cancer, and if successful, the strategy in the current study could be generalized to treating pancreatic cancer as well as a variety of solid cancers.
In their in vitro studies, the authors have used 4 human pancreatic cell lines, but for the xenograft study, they only used the AsPC-1 cell line. Was there a reason to choose this single cell line? Also, was there a reason to use only female mice for the experiment? Do the authors see a similar response in male mice?
The authors have chosen to use xenografts in immune deficient mice. Why not test these engineered viruses using well-described genetically engineered and immune-competent mouse models of pancreatic cancer, such as the KPC mouse model, which resembles many of the features of human pancreatic cancer, including the fibrotic response?
Finally, to be effective as a clinical treatment, the viruses cannot be used as direct tumor injections, and most likely would need to be administered systemically. Have you tested these oncolytic viruses expressing CD19 using systemic delivery, and if so, are they safe to be delivered in this fashion, and are they effective?
DR ALLAN TSUNG (Charlottesville, VA): You discovered the effectiveness of intratumoral injection of CD19 oncolytic virus in the chimeric antigen receptor (CAR) T cells and reducing pancreatic cancer tumor growth in the superficial flanks of mice. Although we recognize the success of intratumoral drug delivery for this study and the FDA’s approval of various other malignancies, such as melanoma, it is crucial to examine the distinctive challenges posed by pancreatic cancer.
Pancreatic cancer is located deep in the abdominal cavity, and therapy for pancreatic cancer will often involve addressing widespread metastatic disease.
Given these challenges, exploring alternative delivery methods, particularly intravenous injection, becomes essential. Can you address the safety of intravenous administration of the oncolytic virus?
In addition, do you think intravenous injection can achieve therapeutic effects comparable to intratumoral injection in pancreatic cancer’s unique features, such as dense fibrotic environment, which could inhibit delivery of the virus?
The oncolytic virus expressed in CD19 resulted in almost 100% of CD19 on the cell surface expression in vitro. However, in vivo, with the doses that you use, only 20% of the tumor cells were CD19+.
So, as we consider clinical translation, what factors contribute to variability in the level of expression in vivo among different cancers? Are there ways to optimize CD19 infection so that more tumors can express CD19?
Finally, given that the engineered oncolytic virus alone induces pancreatic cancer cell death, exploring whether the cell death can trigger antigen recognition by unmodified T cells, subsequently activating them to exert a tumor-killing role could be very valuable.
Have you studied whether oncolytic virus treatment alone changes the tumor microenvironment, especially the function of native T cells, so virus treatment alone can be a potential therapeutic option?
DR SELWYN VICKERS (New York, NY): If this is a conditionally replicative virus, what is the actual cellular target and the driver to both promote and prevent replication? Can you tell us what allows it to broadly target cancer in solid tumors?
I agree with Dr Evers’ question in an immunocompetent model—have you looked at it in a genetically driven immunocompetent model, potentially the KRAS model?
Finally, have you (either in vitro or potentially in vivo) looked at it in the tumor microenvironment of pancreatic cancer, which really affects the cold nature of the tumor and its immune response?
DR OMAIDA VELAZQUEZ (Miami, FL): What does the replication curve of the virus in the lung tissue look like? What cell types are being infected or transduced in the lung tissue? Are there associated secondary effects to replication in the lung, such as either thrombogenicity in the adventitia or potential impairment of the oxygenation at the alveoli level?
DR JONATHAN LARYEA (Little Rock, AR): With the expression of CD19, it will presumably be on the surface of the tumor. Have you figured out how this oncolytic virus causes lysis in the middle of the tumor?
DR CRAIG SLINGLUFF (Charlottesville, VA): CD19 is obviously a relevant target for this model, but in humans, the CD19 CAR will deplete circulating B cells. I wonder whether any approaches have been tried using an antigen that will not lead to that toxicity.
DR MARGO SHOUP (Orlando, FL): CAR T-cell therapy has been around for several years and has generated a lot of excitement, although it is very cumbersome to administer, as you mentioned, with leukapheresis with a centralized laboratory to develop the product. It is also incredibly expensive, more than hundreds of thousands of dollars per patient.
Our physicians who specialize in hematologic malignancy are now generating excitement regarding the use of bispecifics instead of CAR T cells in select patients. I am curious whether there is any role down the road for bispecifics for this model.
DR QUYEN CHU (Washington, DC): As we all know, 1 of the downsides of using CF33 oncolytic virus is that it is rapidly cleared in the system by neutralizing antibodies. Because patients with pancreatic cancer die of metastatic disease, it is imperative to have effective systemic treatment. What are the authors’ thoughts about overcoming the barriers with CF33?
DR VINOD BALACHANDRAN (New York, NY): One of the challenges for adoptive cell therapy, as well as CAR therapy, is cell durability in vivo. I wondered if you had any thoughts on pairing the antigen delivery along with other signals to allow for enhanced cell durability.
Do you have any comments on antivector immunity that might develop in the host?
DR COURTNEY CHEN (Duarte, CA): Regarding Dr Evers’ comments, for our study, there was not truly a specific rationale as to why it had to only be the AsPC-1 line. However, AsPC-1 has already been proven as a viable xenograft model in our laboratory and in other studies. Additionally, it was neither the most or least responsive to CF33-CD19t or CD19 CAR T cells in vitro, so it is a middle-ground for likely response. We did not have to use only female mice. However, in our experience using previous xenograft models, there would be no expected difference in male mice.
We have previously tested this combination in a triple-negative breast cancer model in an immunocompetent mouse model. In that model, we saw similar effects between the immunodeficient mice and the immunocompetent mice. Hence, in our study, which was to ascertain proof of concept that this combination would work in a different tumor type, we did not pursue an immunocompetent pancreatic cancer tumor mouse model.
Regarding Dr Tsung’s comments, we have previously tested the virus systemically. We have undergone robust preclinical testing for the parental virus CF33 as well as our CD19t version, and we are currently in clinical trials using CF33. In terms of intravenous administration efficacy, we have previously shown in a different tumor type, using a CF33 variant and also when we inject IV oncolytic viruses in the apancreatic tumor model with peritoneal carcinomatosis, that IV oncolytic viruses still allow for primary tumor burden control and can control the peritoneal metastases similar to an intratumoral dose.
In terms of penetration into the dense pancreatic cancer fibrotic environment, as our CF33-CD19t virus continues to lyse, we see further penetration into the tumor and destruction of extracellular matrices. Moreover, CD19-CAR T cells improve our viral replication and capability to penetrate deeper into the pancreatic tumor model. In our triple-negative breast cancer model that was previously published, we have seen in vivo further penetration of both CD19 CAR T cells and CF33-CD19t into the cores of the tumors.
Part of the reason for the disparity in in vitro vs in vivo for our CD19t expression with our virus, we believe, is a dose issue. We have tested our virus in a panel of different cell types, ranging from ovarian, gastric, and breast, to glioblastoma and more. In each, we show that in our virus we can increase the CD19-T expression to 100% in a dose-dependent manner. The dosage we used in our study was not the maximum that the mice could tolerate. The reason we did not use a higher dose was that our CF33-CD19t virus is so efficacious in antitumor activity that if we gave a higher dose, our virus would likely eradicate the tumor before we could show the synergistic effect of the CD19-CAR T cells.
We do see that the oncolytic virus alone does modulate the tumor microenvironment. In our study, we saw that we had increased interleukin (IL)-2 and interferon-gamma when the virus was present. In addition, we demonstrated in vitro that the virus can increase coactivating molecules for T cells, such as CD137. Several other studies by our group in pancreatic cancer models using other CF33 variants likewise demonstrate enhanced T-cell recruiting, proliferation, and expansion. There are various data as well for oncolytic viruses overall that report enhanced immunogenicity of infected tumors.
To Dr Vickers’ questions, the virus is not technically conditionally replicative. The parental virus is a recombinant virus of multiple strains of poxviruses and contains a thymidine kinase deletion, which is shared in CF33-CD19t (the deleted thymidine kinase is replaced by encoded CD19t). Investigation into part of the efficacy mechanisms of CF33 reveals that it has an increased extracellular virion-forming potential. Additionally, investigation into CF33 killing mechanisms demonstrates that it kills cancer through necroptotic immune cell death mechanisms rather than apoptosis. Previous studies in our virus and other thymidine kinase-deleted oncolytic viruses demonstrate attenuation of the virus in normal tissue, whereas tumor cells can complement the deficiency and allow still for viral replication.
Regarding Dr Velazquez’s questions, the viral replication in the lung tissue was not directly tested in normal cells in vitro but rather as a harvest at 3, 7, and 10 days after intratumoral injection. Our viral replication was tested using a viral plaque assay rather than a nucleic acid-based polymerase chain reaction detection, as we wanted to measure active, functional virus. A gross curve can be estimated by undetectable virus at day 3, borderline minimal functional virus at day 7, and approximately 1 log-fold higher by day 10. We did not perform a specific analysis of lung cell types infected, as we homogenized the lung tissue to create our cell lysate to harvest our virus for functional assay.
There do not seem to be clinically relevant secondary effects in the lung tissue regarding damage, thrombogenicity, or impairment at the alveoli level. Clinically, our mice did not appear to have any respiratory distress or impairment of oxygenation, nor did there grossly appear to be any significant edema, effusions, necrosis, or hemorrhage during lung tissue harvesting or euthanization dissection.
In reply to Dr Laryea, the oncolytic virus is self-propagating, with exponential growth per lysis cycle. As it continues to infect and lyse, the virus will continue to penetrate deeper into the tumor. Moreover, CD19 CAR T cells enhance viral replication (which we have shown in vitro), and penetration. As mentioned earlier, a previous study in triple-negative breast cancer demonstrated viral and CAR T-cell penetration deep into the tumor.
Addressing Dr Slingluff’s questions, CD19 was primarily chosen as a target because CD19 CAR T cells have already been approved by the FDA for therapeutic use in B-cell leukemia or lymphomas and therefore would likely have a quicker path to testing and hopefully approval or use using CF33-CD19t. Previous studies examining the effect of B-cell depletion after CD19 CAR T-cell treatment have also demonstrated that it is relatively well tolerated with minimal impact on quality of life. Hence, CD19 was also chosen given its favorable side-effect profile. Certainly, there are several other targetable epitopes, such as claudin, HER2, CEA, and more. Some, however, are shared on normal cell tissue, such as HER2 on cardiac tissues, and suffer on-target off-tumor effects, limiting their application. Still, CAR T cells are being examined for a variety of tumor-specific and novel epitopes and are still being modified to limit off-tumor effects. However, we have not yet fully tested our combination therapy with a delivery of a different tumor antigen and paired CAR T cells but are currently investigating potential avenues.
In response to Dr Shoup, we currently have FDA approval for a trial testing our CF33-CD19t virus in combination with Blincyto, which is a bispecific T-cell engager. We are hoping to pursue our additional bispecifics as well, with viruses encoding other antigens.
Regarding Dr Chu’s points, we have previously tested a different CF33 derivative in a pancreatic tumor model with peritoneal carcinomatosis and have found that it is able to reduce both the primary tumor and the peritoneal tumor burden, even via IV injection. We have also tested another CF33 derivative in a gastric cancer peritoneal carcinomatosis model with intraperitoneal injection with significant reduction in metastatic burden. From that aspect, we would expect our CF33-CD19t virus to likewise exhibit peritoneal carcinomatosis control. When we tested a peritoneal metastases model in our triple-negative breast cancer model with our CF33-CD19t and CD19 CAR T model, we were able to demonstrate the combination reduced tumor burden and improved overall survival when treated with an intraperitoneal injection. Given that IV and IP injection demonstrate efficacy in metastatic models with both CF33-CD19t plus CAR T cells and other tumor models with other CF33 derivatives, we think that the barrier to metastatic treatment is surmountable, based on our current data.
Finally, to Dr Balachandran’s comments, we have noted that the presence of CF33-CD19t with CD19 CAR T cells increases T-cell activation, shown by an increased CD137 expression. Additionally, we have performed other studies where we specifically monitor T-cell anergy, propagation, and durability, and show that combining our oncolytic virus with the CAR T cells enhanced cell durability. We have also investigated other viruses carrying other proinflammatory molecules, such as IL-2 and IL-15. We agree that enhancing the immune response with additional signals may be beneficial, although some degree of caution must be taken, as there may be additional toxicity when combining such a potent signal molecule with our already proinflammatory virus.
Regarding preexisting immunity, there have been some studies on whether preexisting immunity affects intravenous delivery of oncolytic viruses. So far, it seems that, rather than hindering the delivery or the antitumoral effects, the existing antiviral response can boost the immune proliferation of the cells and increase our antitumoral effects. There are some clinical studies that support the theory that preexisting immunity increases oncolytic virus efficacy, or at least does not hinder it. For example, talimogene laherparepvec is still efficacious, despite most of the population having been naturally exposed to herpes simplex.
(20min delay)