Cancer vaccines: déjà vu all over again

News of two cancer vaccine clinical trial failures brought to mind the famous Yogi Berra quote “déjà vu all over again”.

"For the Win"

Argos Therapeutics was forced to halt a Phase III study of rocapuldencel-T in renal cancer patients when an interim data analysis indicated that continuing the study was futile. Agenus, somewhat shyly, announced the halting of a Phase II study of its Prophage G-200 vaccine in newly-diagnosed glioma patients for the same reason.

Both rocapuldencel-T and Prophage® are “personalized” vaccines which present cancer antigens derived from the patient’s own tumour.  In theory, these so –called “neoantigens” should be able to get around the problem of tumour-induced host tolerance and elicit cancer-fighting T cell responses.

Despite their sophistication, the Argos and Agenus candidates were unable to induce effective immune responses. A variety of other personalized cancer vaccines are in development but the complexity of neoantigen design is well-recognized and fine-tuning of epitope-predicting algorithms will require accumulation of a large body of clinical data. 

Although still a long way from validation, the prospect that the bespoke neoantigen approach might achieve the stunning outcomes observed in the minority of checkpoint inhibitor-treated individuals in a far larger percentage of treated patients has prompted several high-ticket licensing deals in the past year.

Cancer vaccine success has proved largely elusive over the past three decades, with only Dendreon’s ill-fated prostate cancer vaccine, Provenge®, managing to secure regulatory approval (call me a purist, but I don’t consider Amgen’s Imlygic® a vaccine. We can agree on “immunotherapeutic”). 

As someone whose professional interest in cancer vaccines goes back to the 90s, I’m hopeful that the broader renaissance in cancer immunotherapy will, directly or indirectly lead to useful cancer vaccines, either through combination with tolerance breaking biologics or small molecules or through a better understanding of the subtle interplay between tumours and the host immune system.

To quote Mr Berra once again, “It ain't over 'til it's over”. 

Image By Bowman Gum (Heritage Auctions), via Wikimedia Commons

Added 1st Match 2017:

A study published in Genomic Medicine underscores the challenges inherent in both "universal" and personalized cancer vaccine design due to tumour genetic diversity (the researchers dubbed the analysis "Project Snowflake"). A hypothetical "semi-universal" vaccine based around 10 neoantigens was predicted to be of relevance to only around 0.3% of the population.

Genomic analysis of 63,220 tumors reveals insights into tumor uniqueness and targeted cancer immunotherapy strategies. Hartmaier, RJ et al.
Genome Medicine 20179:16 doi: 10.1186/s13073-017-0408-2. Online 24th February 2017.Full text available http://tinyurl.com/jus5mwf

An editorial in the February 2017 issue of Nature Biotechnology on the complexity of neoantigen prediction is also worth a read:
The problem with neoantigen prediction.  Nature Biotechnology 35, 97 (2017) doi:10.1038/nbt.3800 Published online 8th February 2017. Free to access at
http://tinyurl.com/zfyejaa