Antigen Spreading and Immune-Based Therapies: An Interview with the National Cancer Institute’s James L. Gulley

Let’s start with the basic question: how are immunotherapies different from other classes of anti-cancer agents?

James L. Gulley, MD, PhD: To understand why immunotherapeutic agents might produce different outcomes than those expected for other modalities, we need to first lay a foundation for understanding how exactly these agents are different. It’s important to note that therapeutic vaccines, unlike conventional chemotherapy, are not directed specifically against the tumor or the tumor microenvironment; rather, they are directed at the immune system, which then in turn can target and attack the tumor. Furthermore, with conventional therapy, clinicians don’t expect to see much in terms of what we call “memory response,” whereas developing a long-lived memory immune response is a major goal of therapeutic vaccines.


Another difference, based on studies that we’ve seen, is the onset of action. With hormonal therapies and chemotherapies, we expect to see early onset of action in terms of reducing clinical outcomes. With immunotherapy, we only see true clinical benefit after a longer period of time. This is partly because immunotherapy develops an adaptable immune response that can improve and become more clinically relevant over time. This is  an ongoing, iterative process that we refer to as the “antigen cascade” or “antigen spreading.” We can talk about that in depth later.


With conventional therapy, clinicians don’t expect to see much in terms of what we call “memory response,” whereas developing a long-lived or durable memory immune response is a major goal of therapeutic vaccines.

This potentially delayed clinical response following immunotherapy can be active long after the patient has received the immunotherapy. In addition, typically when we see mutations, for instance, to tyrosine receptor kinase inhibitors, that drug will not work nearly as well and may cause resistance to therapy; mutations in tumors for the immune system, however, provides us with another target for the immune system to be able to capitalize on.


Finally, with conventional therapy—and this is perhaps one of the most important differences—we are typically limited in how much we can give by the toxicity associated with chemotherapy or tyrosine kinase inhibitors. Fortunately, this is not the case for therapeutic vaccines; however, they do require adequate immune system function both locally at the site of the tumor and systematically.


This systematic immune response has been a main focus of the discussion of immunotherapy selection. The systemic immune competence is generally intact in patients until late in their disease, but the complex biologic processes in the local tumor microenvironment may be more problematic for immune therapies. We have known this as a community for a while now, but we are learning new ways to help overcome this.


So, when balancing the differences between conventional therapy and immunotherapy, the key points to consider are: the potential benefit of having lower toxicity with immunotherapies and this delayed yet prolonged clinical response following immunotherapy.


How should patients be selected for immunotherapy? Is there any way to predict which patients would benefit best from immunotherapy?

We have been looking for predictive biomarkers for many immune-based studies. In patients treated with anti–PD1 antibody, for instance, there has been a suggestion that if the tumor expresses anti-PD-L1, the patient will be more likely to respond.1 One emerging intuitive hypothesis is that immunotherapy may be more effective when used earlier on in the disease course, when patients have a lower tumor burden and fewer immunosuppressive mechanisms in play at the local tumor microenvironment.


There is good rationale for this, and indeed, there’s also good data from the IMPACT [Immunotherapy for Prostate Adenocarcinoma Treatment] trial, which is a retrospective analysis of patients with metastatic castrate-resistant prostate cancer (mCRPC) who were randomized to receive sipuleucel-T or control.2 In this study, Schellhammer et al. looked at the overall survival of patients who received sipuleucel-T, and then broke that down by quartiles of prostate-specific antigen (PSA).


Patients on sipuleucel-T who were earlier in their disease course—as evidenced by a lower PSA (<22)—actually lived longer than control patients. What was interesting was that patients who received sipuleucel-T lived about 41 months, versus 28 months for patients in the control group. This 13-month improvement over control translates to about a 50% reduction in the risk of death. Even more interesting was the overall survival rates in the higher PSA quartiles: with a PSA of 22-50, the difference in length of survival between sipuleucel-T and control was 7 months; with a PSA of 50-134, the difference was 5 months; with a PSA >134, the difference was only 2.8 months.2


Again, it appears that if this treatment is initiated earlier on in the disease course, we are able to continue to maintain control over cancer growth for a longer period of time. So, simply because the drug was started earlier, we have the ability to slow down the growth rate for a longer period of time. We can think of this as a “compounding interest” situation: the earlier you start saving, the more money you’re going to make from interest down the road.


In terms of immunotherapy, how important is earlier detection of prostate cancer—especially considering that immunotherapy has that “delayed yet durable” response?

That’s a great question. I think we will get answers when all of these immunotherapies have been tested in large, randomized studies with patients who are later in the disease course. Certainly, the overall survival studies have been in metastatic disease, but there’s no reason why these treatments couldn’t be studied in earlier disease states.


There are ongoing and planned studies looking at these immunotherapies in the newly-diagnosed patient population—even in the active-surveillance patient population, where studies are designed to determine if immunotherapies can prevent the serious adverse events that typically accompany local therapies. I think this is an aspect of treatment that will have to be explored in greater detail in larger trials down the road, but there’s no reason why something that is well tolerated couldn’t be used earlier on in a disease setting. Perhaps one day it would even be possible to identify “high-risk” groups of patients who could be treated in the setting before they are diagnosed with cancer—in other words, immunotherapies could conceivably act as preventive agents.


Earlier you mentioned the antigen cascade or antigen spreading when talking about the potential benefits of immunotherapies over conventional therapies. Could you explain the importance of this process in the development of therapies to treat metastatic cancers?

The antigen cascade is something that I believe has been under-appreciated until recently. Let me give you a little background: in pre-clinical studies, researchers have shown that, in difficult-to-treat tumors, when you can generate an immune response against the tumor that leads reproducibly to eradication of the tumor, you will see vaccine-specific T cells, as well as tumor-specific T cells that are not directed against targets found in the vaccine.


This is important to note because this observed response suggests that administering a vaccine initiates an immune response that can then find the tumor and start to kill it. When that initial immune response comes in, it seems, it may not be attacking the most immunogenic target.


In fact, the initial immune response to vaccine may not lead to the optimal clinical benefit in a particular patient, but if it is possible to kill some of those cells, it may be beneficial for the patient because these cells are being killed in an immunologically relevant manner. They are then taken up by scavenger cells that will present these tumor antigens to the immune system. It’s like a smorgasbord, really. In this cascade, the immunotherapy can select out other targets, as well; so, for example, if there is a novel mutation in a certain type of cancer, the immune system can broaden that immune response.


So, for instance, as a treating clinician, I don’t want to just train the immune system to target prostatic acid phosphatase (PAP) or just PSA—I want it to also focus on other targets that are present in a particular tumor. By doing that, we can broaden the immune response, which may lead to a more clinically relevant outcome or more clinically relevant immune response leading to a better outcome. In a variety of clinical trials, we have started to see evidence of this relationship: patients with a broader immune response to the vaccine and other tumor targets actually fare better.


I also recently published an editorial in Human Vaccines & Immunotherapeutics outlining some recent data showing that broader immune responses to antigens not found in the vaccine translated into improved outcomes.3 This relationship also helps explain the iterative process through which the immune response strengthens over time, and one reason why we shouldn’t expect to see an immediate response with therapeutic vaccines. Yet, you could still see a clinical benefit and disease stabilization or regression later on, even if there is initial progression.


We also performed a trial at the National Cancer Institute that adds to the data produced by the IMPACT trial in which we looked at patients receiving PSA-TRICOM (a novel vector-based vaccine designed to generate a robust immune response against PSA-expressing tumor cells).4 In this study the median predicted survival was 12 months. For patients who had a predicted survival of less than 18 months (i.e., were later in their disease course), they performed about as well as expected: they actually lived 14.6 months, which was 2.3 months longer than the Halabi nomogram prediction. But for patients earlier on in their disease course who had a predicted survival longer than 18 months, the median survival of 21 months had not yet been reached at 37 months. So, ultimately, the survival was 16 months longer than predicted.4


These results strengthened the idea that patients earlier in the disease progression derived more benefit from therapeutic vaccines, just like what was seen with the IMPACT trial. There have also been multiple other retrospective analyses demonstrating this finding. In a review I co-authored a few years ago, we found that results from all of these trials trended in the same direction: the earlier you started the therapy, the fewer the symptoms, the less bulky the disease, and the better off the treatment effect was with immune-based therapy versus the control arm.5


When we’re thinking about applying these findings in clinical practice, are there any special considerations for oncologists when incorporating immune-based therapies into practice?

There are, and I think the number one consideration is knowing how to manage the expectations of the physician and the patient. This involves letting them know about that “delayed yet durable” response we talked about earlier. Both physician and patient must keep in mind that, for instance in prostate cancer, while PSA declines or changes in symptoms might not be immediately apparent, the immunotherapy can still be an effective therapy.


Secondly, patient selection is important: this is not an ideal treatment in a patient who has significant symptoms because of this delayed response or delayed stabilization. If a patient does have greater symptom burden, then another therapy would likely prove more beneficial.


One also needs to weigh the side effects of different therapies. Personally, when comparing the side-effect profiles of all prostate cancer therapies—hormonal therapies, chemotherapies, and therapeutic vaccines—I find that the therapeutic vaccines certainly have less in terms of side effects. If a therapy can rationally be started early and also has fewer side effects, then, to me, it’s a no brainer to go ahead and initiate that therapy early on—while a patient is asymptomatic or has minimal symptoms, and before it is necessary to “switch gears” and change to a therapy that will be more likely to slow down the disease.


In your opinion, how can physicians elicit the best possible response from immune-based therapeutic vaccines?

There are several factors to take into consideration. I would say that, if a patient has significant symptoms, there may be a role for combining these therapeutic vaccines with other therapies. As I mentioned earlier, in this disease setting, we have focused mainly on what is happening in the peripheral immune response, but not as much on what is happening in the level of the tumor. But it turns out, of course, that if a therapy generates a greater immune response so that the tumor-specific T cells have the ability to kill the tumor as a result, these cells still need to get to where the tumor is to recognize that tumor and to kill it. There are a many processes happening in that tumor microenvironment that can thwart such a response, though, such as negative regulatory cells or negative factors that shut down an immune response, like IDO or TGF-beta or IL10, or PDL1 expression at the level of the tumor. All of these factors could potentially make it more difficult for anti-tumor immune cells to kill tumor cells.


As research has shown, it is possible with certain therapies to essentially change the way tumors look to the immune system to make it easier for the immune system to recognize or kill the tumor. This provides a strong rationale for studying the effect of combination therapies. Currently, there are many ongoing clinical trials comparing sipuleucel-T, for instance, with abiraterone or with enzalutamide,6 as well as the combination of enzalutamide with PSA-TRICOM.7


At the National Cancer Institute, for instance, there are two trials examining enzalutamide with or without PSA-TRICOM in both patients with non-metastatic prostate cancer but elevated PSA8 and in patients with metastatic castration-resistant prostate cancer.9 I think it is trials like this one examining combination therapies that are going to help us move closer to the goal of long-term disease control in patients with advanced disease.


An important note about combination therapy is the new immune checkpoint inhibitors that are currently being studied. There seems to be a lot of excitement about the possibilities of combining therapeutic vaccines with immune checkpoint inhibitors. The immune checkpoint inhibitors might be a little “spicier” with a few more side effects, but they also lead, often times, to decreases in tumor size—this is something we don’t often see with therapeutic vaccines alone. So, the combination of the two may potentiate those effects without causing much additional toxicity.


Along those same lines, we recently published a study in The Lancet Oncology where PSA-TRICOM was combined with ipilimumab, an anti-CTLA-4 monoclonal antibody.10 In the future, though, we would like to design a trial combining therapeutic vaccines with anti-PD1 or anti-PD-L1 antibodies, which may possess even more favorable safety profiles than the anti-CTLA4 antibodies.


This kind of combination of two different immune-based therapeutic modalities (in this case, vaccine plus immune checkpoint inhibitor) is what we refer to as “immunogenic intensification.” And, with immune checkpoint inhibitors, we are starting to see a unique phenomenon in terms of overall survival: not all patients have a response, but there is a tail on the curve where many patients are experiencing long-term disease control. We now have long-term data—out to 8 to 10 years in the original ipilimumab studies—showing that 15-25% of patients do have this long-term disease control.11


As a community, I think our goal now is to raise that number. If we can have this phenomenal outcome in these patients and we are able to control their disease in this setting that has been considered uniformly fatal, that would be incredible. So, we are asking ourselves, “Can we increase the proportion of patients who respond?” I think one way that perhaps we can accomplish that goal is by adding other immune-based therapies to the armamentarium. While recent studies that looked at the combination of ipilimumab and nivolumab (an anti-PD1 antibody) suggested much higher response rates with these immunotherapies, they also found higher toxicity. So, I am looking forward to the possibilities of achieving those same results, but with lower toxicity, by implementing a combination of therapeutic vaccines with the immune checkpoint inhibitors. I think these are going to be exciting times in the next few years as we try and grapple with some of these issues.



  1. Brahmer JR, Tykodi SS, Chow LQ, et al. Safety and activity of anti-PD-L1 antibody in patients with advanced cancer. N Engl J Med. 2012;366(26):2455-65.
  2. Topalian SL, Hodi FS, Brahmer JR, et al. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N Engl J Med. 2012;366(26):2443-2454.
  3. Schellhammer PF, Chodak G, Whitmore JB, et al. Lower baseline  prostate-specific antigen is associated with a greater overall survival benefit from sipuleucel-T in the Immunotherapy for Prostate Adenocarcinoma Treatment (IMPACT) trial. Urology. 2013;81(6):1297-302.
  4. Gulley JL. Therapeutic vaccines: the ultimate personalized therapy? Hum Vaccin Immunother. 2013;9(1):219-21.
  5. Gulley JL, Arlen PM, Madan RA, et al. Immunologic and prognostic factors associated with overall survival employing a poxviral-based PSA vaccine in metastatic castrate-resistant prostate cancer. Cancer Immunol Immunother. 2010;59(5):663-74.
  6. Gulley JL, Madan RA, Schlom J. Impact of tumour volume on the potential efficacy of therapeutic vaccines. Curr Oncol. 2011;18(3):e150-7
  7. Vogelzang NJ, Vacirca J, Kantoff P, et al. “Abstract no.: 185. “Effects of prior abiraterone (ABI) or enzalutamide (ENZ) on sipuleucel-T (sip-T) manufacture in PROCEED patients (pts).” Abstract presented at the 2014 Genitourinary Cancers Symposium, San Francisco, California. January 30, 2014.
  8. Singh NK, Kim JW, Heery CR, et al. “Abstract no.: TPS5104. A randomized phase II clinical trial of enzalutamide in combination with the therapeutic cancer vaccine, PSA-TRICOM, in metastatic, castration resistant prostate cancer.” Presented at 2013 ASCO Annual Meeting. Chicago, Illinois. June 1, 2014.
  9. National Cancer Institute; National Institutes of Health Clinical Center. Enzalutamide in combination with PSA-TRICOM in patients with non-metastatic castration sensitive prostate cancer. In: [Internet]. Bethesda (MD): National Library of Medicine (US). 2000- [cited 2014 Jun 11]. Available from: NLM Identifier: NCT01875250.
  10. National Cancer Institute; National Institutes of Health Clinical Center. A randomized phase II trial combining vaccine therapy with PROSTVAC / TRICOM and enzalutamide vs. enzalutamide alone in men with metastatic castration resistant prostate cancer. In: [Internet]. Bethesda (MD): National Library of Medicine (US). 2000- [cited 2014 Jun 11]. Available from: Identifier: NCT01867333.
  11. Madan RA, Mohebtash M, Arlen PM, et al. Ipilimumab and a poxviral vaccine targeting prostate-specific antigen in metastatic castration-resistant prostate cancer: a phase 1 dose-escalation trial. Lancet Oncol. 2012;13(5):501-8.
  12. Hodi SF Jr. “Abstract no.: LBA 24. Pooled analysis of long-term survival data from phase II and phase III trials of ipilimumab in metastatic or locally advanced, unresectable melanoma.” Presented at the 2013 European Cancer Congress, Amsterdam, the Netherlands; September 28, 2013.

James L. Gulley, MD, PhD, is the Chief of the Genitourinary Malignancies Branch at the Center for Cancer Research at the National Cancer Institute in Bethesda, Maryland.