Full text of cancer-vaccine article from Oct 2016 issue of Nature [I highlighted a paragraph pertaining to ADXS in bold-face text (about halfway down); tables and graphics are omitted.] http://www.nature.com/nbt/journal/v34/n10/abs/nbt.3690.html
›A flurry of recent deals has focused on the development of immunotherapies that seek to exploit the full repertoire of a tumor's antigens, revealed through immune profiling at the individual patient level. An agreement by Amgen (Thousand Oaks, CA, USA) acquiring exclusive, worldwide rights to develop and commercialize Advaxis' (Princeton, NJ, USA) attenuated listeria platform for expressing multiple tumor antigens is just the latest in a slew of commercial investments (Table 1). Mutated proteins bearing neoepitopes, novel bacterial and viral antigen delivery systems and, most of all, combinations of immunogens, are renewing optimism in cancer vaccines. Numerous questions remain, but the therapeutic cancer vaccine field has unquestionably gotten a shot in the arm. Now, can it finally live up to its promise?
[Table 1: Select deals involving cancer vaccines]
The clinical success of immune checkpoint inhibitors has spurred a stampede of commercial interest in immuno-oncology. Not only has it been the basis for several approvals and blockbuster products, but it may also rekindle the dying embers of another endeavor with decades of dismal trial results and dashed hopes—the field of cancer vaccines. Whereas prophylactic cancer vaccines for HPV-induced cervical cancer and HepB-induced liver cancer have been in clinical use for years, producing a vaccine against an existing tumor has been more challenging, given the hostile tumor microenvironment and barriers of tolerance.
The hope is that checkpoint inhibitors can remove the most powerful escape mechanisms employed by tumors to elude cellular immunity that have thwarted cancer vaccines of the past. According to Hyam Levitsky, who has worked on cancer immunology in academia (Johns Hopkins University, Baltimore) and pharma (Roche, Basel, Switzerland), and now at the biotech Juno Therapeutics (Seattle), therapeutic cancer vaccines probably never had a chance by themselves, given the complexity of the cancer immune cycle (Fig. 1). “When you look at all the things that need to happen for a successful anti-tumor response, why should you expect that a vaccine would take you from zero to 60 all by itself?” he asks. “A successful vaccine primes or expands the functional T-cell response, which may be naturally lacking in settings where checkpoint inhibitors don't work. But primed T cells aren't enough; they still must find their way into the tumor, and overcome local suppressive mechanisms. “That's why combination therapy is key,” he adds.
[Figure 1: Stimulatory and inhibitory factors in the cancer-immunity cycle.]
Checkpoint inhibitors and adoptive T-cell therapy, including tumor-infiltrating lymphocytes (TILs) and chimeric antigen receptor–modified T cells (CAR-T cells), have provided proof of the cancer immunotherapy concept. Several checkpoint inhibitors have now received US Food and Drug Administration (FDA) approval, such as the fully human IgG1 monoclonal antibody (mAb) Yervoy (ipilimumab, Bristol-Myers Squibb (BMS), New York) targeting CTLA-4 and the humanized IgG4 mAb Keytruda (pembrolizumab, Merck, Kenilworth, NJ, USA), which targets programmed cell death 1 (PD-1) receptor, the first two to receive approval.
Melanomas have long been considered a good target for immunotherapy. It has been known for some time that they are 'hot' tumors—tumors that show considerable immune activity, as demonstrated by the presence of TILs within lesions. In addition, melanomas have the largest mutational load of any cancer, which drives the response to checkpoint inhibitors (Fig. 2).
[Figure 2: Somatic mutation burden by cancer.]
These features make melanoma an obvious target for immunotherapy. But surprisingly, now, “We see responses in other cancer types,” according to Ira Mellman, vice president of Cancer Immunology at Genentech (S. San Francisco, CA, USA). But the low response rate to checkpoint inhibitors is a big drawback. Only a fraction of patients (10–40% depending on the type of cancer) respond to these agents.
Enter cancer vaccines. The logic is compelling: once an immune modulator unleashes the immune system against a tumor, a vaccine can then provide that cellular response with a target and a trigger to sustain the reaction. “Everything suggests the vaccines will work in combination [with checkpoint inhibitors],” says Lélia Delamarre, a cancer immunologist at Genentech.
A key question is whether cancer vaccines can succeed based on older approaches that use shared antigens, which are not specific for tumors and thus have to overcome immune tolerance mechanisms. Despite the development of dozens of cancer vaccines over the past couple of decades, only two programs have made it over the finish line: Dendreon's (now a totally owned subsidiary of Valeant of Laval, Canada) Provenge (sipuleucel-T; Box 1), a mixture of autologous lymphocyte and myeloid cells engineered with prostatic acid phosphatase (PAP) and an immune signaling molecule (GM-CSF); and Amgen's Imlygic, which is strictly speaking not a vaccine, but an oncolytic herpes simplex virus 1 (HSV-1) with vaccine-like properties.
It is not obvious why so many vaccines failed. “In many cases, an immune response was generated, but it didn't result in a clinical benefit,” says Tibor Keler, founder, executive vice president and CSO at cancer immunotherapy company Celldex Therapeutics (Hampton, New Jersey, USA). That phenomenon raised more questions about what an effective immune response should really look like.
The rise of neoantigens
As an alternative, many in academia and industry are now placing their bets on neoantigens—newly formed mutated proteins that arise as a result of the genomic instability of tumors—which have come to the fore in recent years thanks to advances in sequencing technologies. Neoantigens do not have the tolerance problem seen with shared-antigens, and mutated peptides (neoepitopes) derived from neoantigens, when presented in the context of the major histocompatibility complex (MHC)-1, can elicit robust T-cell responses. What's more, these mutations may play a key role in cancer's development and progression. “The major issue with vaccines is that we didn't know what the drivers were,” says Delamarre. “Now we believe these mutations are the drivers of the anti-tumor immune response, and we can finally identify them through sequencing for a few thousand dollars in a few days.” Several studies showing that patients who respond to checkpoint inhibitors tend to have numerous mutations in their tumors (melanoma and some forms of lung cancer, for example) have stimulated interest in neoantigens, and their associated neoepitopes.
Mellman agrees. “The past 25 to 30 years were probably the greatest impediment to immunotherapy in the history of the field,” he says. “There was an almost polemic view that conserved [or shared] tumor-associated antigens were going to be good immunogens.... Now we recognize that, instead, there is actually a substantial role for all these mutated, or passenger, proteins, which have been ignored for decades.”
An immune therapy pioneer, Steven Rosenberg and his group at the National Cancer Institute (NCI), recently made strides in understanding the role of neoepitopes in immune responses. A 2014 Science paper from Rosenberg's laboratory describes a patient with a cholangiocarcinoma (an epithelium-derived tumor, which is notoriously hard to treat), from whom they extracted TILs that recognized a mutated erbb2 interacting protein. When those erbb2-targeting TILs were expanded ex vivo to be ~25% of cells and returned to the patient, the tumor dramatically regressed and the patient has been cancer free for three years. Furthermore, they showed that nine out of ten patients with gastrointestinal cancer have mutations that can be recognized by their immune system.
This growing body of evidence is providing impetus for a burst of commercial activity in the area of neoantigen vaccine research and development. One of the biggest financings of 2015 was garnered by the neoantigen vaccine startup Gritstone Oncology (Emeryville, CA, USA), which netted a cool $102 million from investors. Another company targeting neoepitopes, Neon Therapeutics (Cambridge, MA, USA), was launched in 2015.
Twelve months ago, Moderna Therapeutics (Cambridge, MA, USA), originally a developer of synthetic mRNA therapeutics, spun out Caperna, a company whose stated goal is to use mRNA vaccine technology to develop “personalized cancer vaccines,” according to a company press release; this January, after only 12 months of collaboration, Merck licensed one of potentially five Caperna vaccine programs as part of a research agreement in which Moderna received $50 million upfront and $50 million in equity.
Another nucleic-acid-based vaccine maker, Inovio (Plymouth Meeting, PA, USA), is actively investigating DNA vaccines as a platform to deliver cancer antigens. “You have to process the antigen in the patient's own cells,” says Joseph Kim, president, CEO and director at Inovio. “That's what the immune system has taught us, because that lets the body process the antigens in the most efficient and natural way.” Ex vivo activation approaches, such as that used for Dendreon's Provenge (Box 1), will not be as effective as direct processing in the body, he says. The company's vaccines are delivered by electroporation, directly into the muscle through the skin. Inovio is targeting both neoantigens and shared antigens, such as INO1400, which targets a telomerase expressed in over 85% of tumors.
The challenge for companies developing neoantigen or neoepitope-based vaccines is to identify, from among the thousands of mutated peptides that occur in cancers with large mutational loads like melanomas and lung, only those epitopes that are antigenic in individual patients. As such, they are banking heavily on informatics and proprietary predictive models4. “It's dauntingly personalized,” says Rosenberg. “You have to do a whole exomic sequence of the patient's cancer. You then have to do the bioinformatics, so you know all the mutations that a patient's cancer is expressing,” he adds. Although most vaccines to this point have delivered one tumor-specific antigen, some companies are thinking more broadly, which may also potentiate the approach. For example, the Amgen deal with Advaxis centers on the latter company's product, a live attenuated listerial vaccine, which can deliver plasmids encoding up to 50 epitopes each. “Tumors may have hundreds of mutations, but we don't know which ones can actually successfully prime and activate T cells,” explains Elliot Levy, senior vice president of global development at Amgen. “What we find attractive about Advaxis is that its vector can potentially deliver hundreds of neoepitopes to the patient.” The company still has to analyze the patient's tumor to pull out personalized antigens, but they don't have to work so hard to narrow down how many they will use. “We, as scientists, don't do a good job yet of predicting which neoepitopes will generate a good immune response,” Levy adds. The Advaxis approach gives them a better chance, Amgen believes, of getting at least some that can properly stimulate the patient's immune response.
“Some people think you will just need one neoepitope, as long as it's the right one,” says Robert G. Petit, Advaxis' executive vice president and CSO. That's possible, he says, but the more neoepitopes used, the greater the likelihood of getting a good immune response. “In our animal studies, we find that the more neoepitopes you use, the better the response.” The company's listeria-based delivery system, he says, also induces higher levels of GM-CSF and other inflammatory cytokines and chemokines, which enhance the development of T-cell response against the specific neoepitopes.
Boosting recombinant bacterial vaccines
Even if you identify the right antigen or epitope, there is no guarantee that this alone will trigger a cellular immune response in the tumor milieu that may contain regulatory T cells, myeloid-derived suppressive cells and stromal cells. Thus, investigators are turning to checkpoint inhibitors to further flame the immune response.
What's not clear is whether knowledge of the biological underpinnings of the so-called cancer immune cycle is yet sufficient to start making the call on how to structure these combinations (Fig. 1). “We understand bits of why the immunotherapies are working. We have a decent hypothesis about the checkpoint inhibitors,” says Mellman. “But we still don't really know most of it works.”
Mellman says the area is crying out for more basic research. But many companies are plowing ahead, and the leap into combination trials for oncology has been swift, with dozens of such trials already underway. Some of the most anticipated are combining checkpoint inhibitors, but vaccine developers are trying to make certain that they are involved in hot combination studies as well (Table 2). For any combination, “I think CTLA-4 or PD-1 inhibitors are critical,” says James Merson, CSO of vaccine immunotherapeutics at Pfizer (New York). “But are they sufficient? We may need other agents that can keep the T cells active.”
[Table 2: Select clinical trials combining cancer vaccines and checkpoint inhibitors]
One such program is in development by Aduro Biotech (Berkeley, CA, USA), which, similar to Advaxis, is developing a live attenuated recombinant vaccine using Listeria monocytogenes. Aduro's strain expresses on its surface a fusion of a tumor antigen and a portion of the bacteria's actA peptide, together with a deletion in the internalin B virulence gene. A recent combined trial of the company's GVAX pancreatic cancer vaccine (comprising two irradiated, GM-CSF-secreting allogeneic pancreatic cell lines plus low-dose cyclophosphamide) and CRS-207 (L. monocytogenes expressing a fusion of actA and mesothelin, which is overexpressed in pancreatic tumors, non-small cell lung cancer (NSCLC), ovarian cancers and mesothelioma) failed to meet its endpoints. But the company also has a phase 2 trial ongoing of CRS-207 in combination with BMS's PD-1 checkpoint inhibitor Opdivo (nivolumab). The concept is that the vaccine infects monocytes and releases antigens into the cytosol, where they are processed and presented as epitopes in the context of the MHC, thereby triggering both an innate and adaptive immune response. In 'cold' tumors, such as pancreatic cancer, that have few if any options, researchers are eager to see if such a combination can deliver.
Lei Zheng's group at Johns Hopkins School of Medicine (Baltimore) are participating in several clinical trials combining GVAX and checkpoint inhibitors against PD-1 in various stages of pancreatic cancer, including metastatic disease. An earlier study combining the checkpoint inhibitor Yervoy with GVAX showed a measurable response, although sometimes delayed, in about 20% of patients (3 out of 15) receiving both treatments5. Although it might seem like a low response rate, Zheng points out that these tumors typically respond to little else. “For pancreatic cancer, we have limited treatment options. You're trying to find anything that works,” he says. Other studies have suggested that GVAX can prime the tumor microenvironment, making it more responsive to checkpoint inhibitor therapy6.
Of course, vaccines that are attenuated versions of intracellular pathogens have their own problems; besides the latest failure to show efficacy in pancreatic cancer, Aduro also has had to cope with safety concerns. Last November, a patient receiving CRS-207 came down with listeriosis. The company claimed a protocol violation was the reason, but it could be a potential problem for which experts will now have to be on the lookout. Notably, earlier in 2015, Advaxis also faced a temporary hold from FDA on its trial of an attenuated vaccine of recombinant L. monocytogenes expressing listeriolysin-O fused to the human papillomavirus 16 E7 antigen (axalimogene filolisbac) when a study patient with cervical cancer died and tested positive for listeria.
Bavarian Nordic (Kvistgaard, Denmark) picked up the mantle of the prostate cancer vaccine PROSTVAC in 2008 after the platform had been in development for over a decade at the NCI and had gone through a phase 2 clinical trial. PROSTVAC is a prime-boost vaccine regimen delivering a 786-bp DNA fragment of human prostate-specific antigen in two different viruses—first, a recombinant vaccinia virus prime, then a recombinant fowlpox boost. Researchers at the NCI had reasoned that using a different virus in the boost—one that is less familiar to the human immune system—would allow repeated boosts without inducing an immune response to the vector. “We had showed preclinically that if you give vaccinia first and then fowlpox vector that had the same gene for the same tumor-associated antigen, you get much better responses, so we incorporated that into clinical trials and then we found that you could—if you gave co-stimulatory molecules—get even better responses,” says James Gulley, director of oncology services at the NCI. The latest generation of 'prime' vaccinia vector and 'boost' fowlpox vector takes advantage of co-expression of three co-stimulatory molecules, LFA-3, ICAM-1 and B7.1 in a product called TRICOM. PROSTVAC has already been through a phase 1 trial with Yervoy for metastatic prostate cancer, which intriguingly showed a trend toward longer overall survival. “For the first time, in small numbers, we can see a tail on a chart where you have true, 80-months-out no-progressers. That doesn't happen in the setting of metastatic disease,” says Seth Lewis, vice president of investor relations, for Bavarian Nordic in Boston.
Larry Fong's group at the University of California, San Francisco (UCSF) will also be taking PROSTVAC plus Yervoy into trials, where he will be looking at TILs before and after vaccination. “That's going to be really exciting as it's in the pre-prostatectomy setting, as so much of what we've done so far is metastatic disease where you don't have a primary or isolated tumor to look for any local treatment effect,” says Lewis. In addition, PROSTVAC continues to be tested in numerous clinical trials, some as part of a Cooperative Research and Development Agreement (known as a CRADA) between the NCI and Bavarian Nordic, including a pivotal, large phase 3 trial in metastatic prostate cancer, which has cleared the hurdles of two interim analyses. The NCI will be testing it in combination with both Yervoy and Opdivo. Bavarian Nordic's program has caught the attention of pharma. The company recently inked a deal with BMS worth about $1 billion with milestones and royalties, including a $60-million upfront payment.
A final example is Amgen's Imlygic, an HSV-1 engineered both to show greater selectivity for growth in transformed cells (through deletions in ICP47 and ICP34.5 and expression of immediate-early gene US11) and to express GM-CSF. Delivered directly into melanoma lesions, the virus has a dual mode of action: first, it selectively replicates in tumor cells, causing them to lyse and, second, it likely induces a systemic immune response by releasing a cascade of potential tumor antigens into the circulation, while also boosting the immune system by means of GM-CSF. Although Amgen has stated that the exact mechanism of action is unknown, they have evidence from both preclinical and clinical studies that the virus induces responses at distant lesions, where no virus can be found.
Last fall, Imlygic was approved as a monotherapy for melanoma based on data from a phase 3 trial of 436 patients with metastatic lesions that could not be surgically removed. That study found that 16% responded to the therapy compared with 2% of the control group, which received GM-CSF alone. The modest size of the responder pool as well as no improvement in overall survival left some people scratching their heads. But both the FDA and the European Medicines Agency were apparently swayed by the durable response rate (patients with a response for longer than six months). And, according to David Reese, Amgen's senior vice president of translational sciences, based on updated phase 2 trial data, they have seen patients surviving for as long as several years after Imlygic treatment.
The next step, according to Reese, is to expand that population of responders. Amgen says they already have preclinical data showing synergy between Imlygic and checkpoint inhibitors, and the company is currently partnering with Merck to do combination trials. In addition, a phase 1 study of the virus with BMS's Yervoy has already delivered encouraging results, albeit in a very small group (19) of patients. “They [the drug combination] doubled the response rate,” says Pfizer's Merson. Each treatment has an objective response rate (defined as reduced tumor volume of prespecified amount) of about 20% on its own, but taken together that rate is ~50%. The durable response rate was 44%—again, about twice what was seen with Imlygic monotherapy. The companies are now doing a phase 2 study of that combination. A phase 3 trial of Imlygic and Keytruda (Merck's pembrolizumab) is also enrolling patients based on promising phase 1 data. The analysis from that study looked at data from 16 patients. The objective response rate was again 57%, whereas the durable response rate was 68%.
In addition, Amgen has an ongoing study of the oncolytic virus in head and neck cancer, and they are doing a phase 1 study in liver metastases. “Once we ascertain a safe dose for hepatic injection we then plan on combination studies with a variety of tumors,” Reese says. One goal is to address those typically cold tumors, including colorectal and triple-negative breast cancer.
Imlygic's early combination data provide the first inkling of what's possible. If even half of Amgen's aims are realized, the vaccine will have vaulted into much wider use. The results from the study with Yervoy are also a step toward confirmation that, indeed, even older style vaccines can be efficacious when combined with other immune therapies.
The jury remains out
With so many combination trials underway, experts are watching closely for the slightest evidence that vaccines can actually augment progress already made with checkpoint inhibitors. Big pharma is, naturally, playing a wide field. “We want to target all steps,” says David Berman, senior vice president and head of oncology innovative medicine at AstraZeneca subsidiary MedImmune (Gaithersburg, MD, USA). “We believe in a broad platform—antibodies, fusion proteins, oncolytic viruses and tumor vaccines.”
However, it is clear that Berman, like Pfizer's Merson, puts checkpoint inhibitors at the top. “I think the checkpoint inhibitors are the key thing,” he says. “We have ipilimumab, and the ability to combine vaccines with checkpoint inhibitors, that's what has ignited the field.” Harlan Robins, Fred Hutchinson Cancer Research Center and co-founder of the immunosequencing company Adaptive Biotechnologies (both in Seattle), agrees: checkpoint inhibitors make it a whole new paradigm. “We can show through some of our tools that the vaccines are truly able to induce an immune response against the epitopes that are expressed in cancer,” he adds. The problem now, as before, is going from there to being able to have clinical efficacy against advanced cancer. A strong enough immune response is needed, as well as epitope spreading, among other factors. “But having this big hammer behind you of anti-CTLA4 or anti-PD-1 might get the response big enough to be functional. We haven't really reaped the benefits but I think we will,” Robins says.
With all the new and emerging possibilities, collaboration is going to be key for cancer vaccine companies to thrive. The concept is gaining wide acceptance. “It's clear that to cure cancer, you need a one-two punch,” explains Kim. Of course, it may actually require even more blows in many cases to actually cure most cancers, but based on what we know now, those key punches should be immunotherapies. “I have come to the view that any successful cancer therapy is working as an immune therapy,” says Mellman.‹