Personalized neoantigen cancer vaccines: are more targets always better?

In April this year, the spinal malignancy of Carlos Gil became so large that his C7 vertebra fractured. The father of four was forced to sleep in the downstairs guest room, biting a pillow so his children would not hear him scream through the night. Gil said that in those long moments of pain, everything seemed to be over.

But just four months after receiving a personalized cancer vaccine, he said he felt much better. "Before the vaccine, I was on death row," Gil said. "I am proof that this therapy can work."

Doctors at the University of California San Diego (UCSD) said Gil's tumor is 90% smaller than before! At the time, the pain caused by his tumor was so severe that he sometimes fainted, and paralysis was almost certain. This is a miraculous transformation that shows the potential of the cancer vaccine research field.

Improving the hit rate of targeted antigens

During tumor development, mutations and genomic changes that accumulate as the tumor grows can produce neoantigens, which exist uniquely in cancer cells and are immunogenic. The immune system can recognize these mutant proteins as foreign and generate a neoantigen-targeted immune response. A growing number of trials are using recent advances in gene sequencing, bioinformatics analysis and manufacturing to prepare neoantigen vaccines to activate each patient's immune system against the specific genetic fingerprint of their tumor. Other research efforts collect samples of patients' tumors and healthy tissues, sequence their genetic codes, and then compare the results, allowing scientists to discover mutations present in cancer but not in healthy tissue. These differences are used to predict changes in cancer antigens.

But merely predicting which neoantigens may exist is not enough. T cells do not simply encounter a neoantigen and immediately start looking for matching cancer cells. Helper cells are also part of the immune system; they must "present" neoantigens to unresponsive T cells using the major histocompatibility complex (MHC). This complex process involves the presentation of MHC binding to neoantigens. After decades of research, immunologists have learned that MHC can vary from person to person. Therefore, to design an effective cancer vaccine, researchers must not only predict which neoantigens are most likely to appear in each patient's tumor, but also predict which targets are likely to bind and present appropriately. The UCSD team hopes the vaccine will elicit a stronger response in patients; to this end, the approach they are pursuing adds a functional step that removes much of the guesswork from the target selection process. Before incorporating predicted mutations into the vaccine, they first test the predicted targets, manufacturing all the predicted neoantigens and exposing them to the patient's blood samples.

Within two weeks, the method can flag antigens recognized by at least some T cells in the blood, to determine whether they can be recognized by each patient's immune system. To help cope with the low potential hit rate, other neoantigen trials often select as many as 20 different targets for each vaccine. But this means the immune system may amplify many ineffective signals. By selecting fewer targets for the immune system in advance, the effect of the anti-tumor response may be significantly improved. Cohen said in an email, "Our evidence in preclinical models will strongly indicate that selecting the right neoantigens is critical. In fact, if a vaccine is loaded with irrelevant antigens or antigens without strong responses, its effect will be reduced. More is not necessarily better; doing things right is better."

In addition to the vaccine, patients also receive the PD-1 monoclonal antibody Keytruda, which blocks the abnormal interaction between the molecule PD-1 on immune cells and the molecule PD-L1 on cancer cells, allowing immune cells to recognize and attack tumors. Clinical studies have shown that this type of immune checkpoint inhibitor appears to work better in tumors with high mutation burden, and may even show ineffective responses for some tumors with low mutation burden that produce few neoantigens. Combined with the vaccine, PD-1 antibodies are expected to unleash the full potential of vaccine-activated T cells.

So far, the results have been mixed. Of the five patients treated with personalized cancer vaccines over the past year, Gil's response is by far the best. In Gil's case, immunologists specifically selected 5 neoantigen targets. Chief investigator Dr. Ezra Cohen said one patient died after their cancer "shook off" the vaccine, and another withdrew from the trial and returned to chemotherapy, although the vaccine did seem to halt the tumor's growth. Another patient was treated only recently, and early results have not yet been published.

Tamara Strauss is the first participant in the trial; she has high-grade stage IV pancreatic neuroendocrine cancer. Her parents, Matthew and Iris Strauss, provided USD 1 million in funding for the trial. In Strauss's case, the blood culture process described above allowed doctors to determine that 3 of the 27 neoantigens predicted by the computer model were already recognized by her immune system. Two of these 3 neoantigens were used to prepare her vaccine; the third could not be included due to manufacturing difficulties. Dr. Stephen Schoenberger, another immunologist leading the study, said: "The accuracy of the results has improved 10-fold. As far as we know, the hit rate of algorithm models is about 3%, but the hit rate of our functional approach is about 35%, which we believe will greatly improve efficiency." Initially, Strauss's tumor remained stable but did not regress like Gil's. Biopsy showed that her tumor stopped expressing the two mutations targeted by the personalized vaccine. Cohen noted that although Strauss's tumor still exists, there are still positive results; although they have not disappeared, they have not grown. Cohen said: "Her disease has been stable for about 10 months, and I think this is evidence that the vaccine is working, just not to the extent we would have liked."

New strategy

The researchers suspect that an important reason the ideal effect was not achieved is that the best targets in Strauss's tumor were not included in her vaccine, because during manufacturing, the short molecules that need to bind to the targets could not stay together. Therefore, they recently began working with a new vaccine manufacturer and tried again to manufacture those molecules that appear to be the best targets inside the tumor. Cohen said: "We would be happy to see it work a second time, so we will vaccinate her again, but this time we will use a different approach." The original study protocol called for three vaccine doses three weeks apart, but Dr. Schoenberger said recent preclinical studies encouraged the researchers to adopt a new dosing schedule: the new approach will start with three injections per week for the first three weeks, followed by six to nine more doses depending on the participant's immune response. The approach aims to mimic the way infectious pathogens gradually stimulate the immune system, ultimately eliciting a strong response. A new arm of the trial will record patients' responses under the new protocol for comparison with the previous three-dose approach. This is the normal law of scientific research. The initial plan looked good on paper, but as real patients began to achieve results in the fight against cancer, adjustments were needed.

Potential hope

Strauss said she has undergone multiple rounds of chemotherapy and multiple abdominal surgeries, and of course she would have preferred that the first cancer vaccine shrink her tumor. However, she said she does not want to back down now. For many people, she said, fighting cancer is simply brutal, and it is time to find better answers. Kristin Peabody's family holds the same view. Kristin died in February this year because the vaccine could not curb the progression of her parotid gland cancer. Her mother, Jan Heaton, together with her husband Wyatt Peabody, established the UCSD Peabody Fund and began raising funds for this trial. Heaton said in an email: "We will continue to support Dr. Ezra Cohen and UCSD's personalized cancer vaccine program. Like Kristin, we believe this is the future of cancer treatment. This is her legacy." This is a rather delicate vote of trust. Cohen said that this support, the ability to see the big picture amid the pain of losing a loved one, ultimately drives science forward. He said: "It is hard for me to imagine any effort in cancer research succeeding at the very beginning. Even the best treatments we have were not universally successful on the first attempt. I think the key is that we do things methodically and scientifically, learning as we go." Initially, the Strauss family's USD 1 million donation was planned to treat 10 patients, but one year after the first vaccination, only half of the patients had been vaccinated. This was because the contract manufacturer took a long time, two to three months, to prepare the vaccine.

Now, working with the new manufacturer, they can produce the needed vaccines more quickly, allowing the trial to treat 10 patients relatively quickly. They also plan to expand the trial to 20 patients. In addition to funding from Strauss and Peabody, the Padres Pedal the Cause charity raised USD 300,000 in donations for the project, which also helped cover the increased costs. Cohen said the current reserve fund will cover 15 patients, but the university needs to raise additional funds to cover the last 5 patients.

Conclusion

The field of therapeutic cancer vaccines is very promising, but has also experienced too many setbacks; perhaps we should be clearer about our strategy in order to break through the obstacles one by one. Personalized vaccines targeting neoantigens have become a hot field, but as mentioned in the article, target prediction and selection, and manufacturing difficulties, all still need more breakthroughs. [Source: 新浪医药]