Read the Full Video Transcript

Andrea Miyahira: Hi, everyone. I'm Andrea Miyahira at the Prostate Cancer Foundation. Thank you all for joining us today. I am joined by Dr. Johann De Bono, a professor at the Institute of Cancer Research in The Royal Marsden NHS Foundation Trust. He'll be discussing his group's recent paper, "Targeting Myeloid Chemotaxis to Reverse Prostate Cancer Therapy Resistance," that was recently published in Nature. Thank you for joining me today, Dr. De Bono.

Johann Sebastian de Bono: Thank you so much, Andrea, and I'm delighted to present our recent manuscript to you today. This was work led by my PhD student, Christina Guo, as well as one of my senior postdocs, who's now a group leader, Adam Sharp, in my group.

So this work really stems from a conviction that inflammation's very important in prostate cancer, with work led by many groups, including the group at Johns Hopkins and Angelo DeMarzo's team. And if I can go back to the A, B, C of all this, when cells are stressed in tissue, they can activate programs that attract inflammatory cells to that site of cellular stress, at a very simplistic level. So, for example, when we stress a normal fibroblast with an oncogene or through treatments like irradiation, that cell can activate a program that involves a change in the cell morphology, and basically, that cell becomes what I call a factory for cytokines that bring in inflammation. That is a normal cell, but the same process can be hijacked by cancer cells, and there is a lot of evidence that inflammation is really important in prostate carcinogenesis and prostate cancer growth, from many laboratories, not least of which are the laboratories of Andrea Alimonti and Angelo DeMarzo and Charles Sawyers. But simply, if tumor cells, when stressed, attract immune cells into them, they can also groom these immune cells to feed them. That's a simplistic argument.

Now, Jesus Gil from Imperial College London published many years ago a work where he stressed these fibroblasts and showed that when you induce, say, KRAS in a fibroblast, you activate these SASP programs. SASP stands for senescence-associated secretory phenotype. And these programs are NF-KB and CEBPB driven and require key cytokines like IL-1a and IL-1b. And these induce the generation by the tumor cell of chemotactic cytokines that attract down the chemotactic gradient, or a chemokine gradient, white blood cells particularly, but also macrophages and other cells, into the tumor. And they can also make these white blood cells secrete factors that feed the tumor.

Now, why did we start studying all this? Well, over a decade ago, one of my laboratory postdocs, David Olmos, who's now a professor at Madrid, a former PCF, a young investigator actually, David, and I were looking for tumor transcripts in blood. And we perhaps foolishly decided, because we're very concerned about RNA stability through the processing of the samples, we decided to actually minimize blood processing time and just took whole blood right into a packaging tube and collected all the RNA from the whole blood straight away, froze the RNA in time exactly at the exit into the vacutainer tube, and then did essentially transcriptome analysis, and analyzed how that transcriptome of the whole blood of a patient associated with survival. And we used Bayesian analysis and, for example, latent process decomposition to find out what signatures associated with the outcome from prostate cancer and what signatures in the whole blood associated with aggressive prostate cancer versus indolent prostate cancer.

And we published this in The Lancet Oncology in 2012, by Olmos et al, back to back with work from William Oh, who at that time was at Dana-Farber. And the long and short of it was we were shocked to find that a signature in the whole blood could completely distinguish pretty much the indolent prostate cancers from the aggressive prostate cancers, but also could tell us which aggressive prostate cancers had the worst prognosis. And working with William Oh's group, we came to the conclusion, the paper published by Wang et al, using our data and data from the Howard Scher Sloan Kettering patients and William's patients at Dana-Farber, that actually the signature that associated with a bad outcome was a myeloid transcript signature. And this was work conducted in the early 2010s.

So, we became very puzzled, why are myeloid cells so important to the outcome from prostate cancer? A bit later, work by two of my fellows, Aurelius Omlin, who now heads the APCCC meeting, and Carmel Pezaro, working with Anthony Joshua and Ian Tannock, looked at myeloid cells in the blood of patients in the full blood count and found, to our surprise, that the neutrophil-to-lymphocyte ratio was the best predictive marker of abiraterone response when I was leading the abiraterone drug development program from phase one to phase three.

And to my even greater surprise, work by another fellow of mine, David Lorente, as well as other work by others, confirmed that not only did the NLR ratio, the neutrophil-lymphocyte ratio in blood, tell us who responded best to abiraterone, but we had very, very similar data for enzalutamide, for taxanes, olaparib, and, most recently, lutetium PSMA. So, something strange was evolving here, that the myeloid cell count in the blood to the lymphocyte ratio told you who could respond to these drugs. So, there's something about the myeloid cells that is impacting sensitivity to therapy.

Now, around this time, a friend and collaborator, Andrea Alimonti, published a paper in Nature that indicated that what this was due to was that when you stress prostate cancer cells with a stressor, like taxanes or abiraterone or other stressors, you make them become senescent. So, they change into this response mode and activate this response to the stress that results in sucking in these myeloid cells in, and these myeloid cells can fuel therapy resistance and tumor growth. So, essentially, inflammation is important to therapy resistance and tumor cell survival.

So, we argued, therefore, that if we could block these myeloid cells from getting into the tumor, that we could block resistance to drugs like enzalutamide. And the best drugs for this are chemokine receptor antagonists, blocking chemokine receptors that are central to attracting myeloid cells into the tumor. And one of the key receptors, although not the only one, is CXCR2. So, we, therefore, got a CXCR2 compound, a drug called AZD5069, that I should say is no longer in development, from AstraZeneca, and we ran an investigator-initiated trial that was a proof-of-mechanism/proof-of-concept trial to ask the question: If we block myeloid cells coming into the tumor, what does that do to enzalutamide resistance? So, the patients have had an AR-targeting drug next-generation inhibitor, and an AR signaling inhibitor, or ARSI, or ARTA, as some call it. They're progressing on that, and we just add the CXCR2 inhibitor and see what happens.

And the main reason we believe this can work is that we've also published papers in Nature, in Cancer Research, and others have published papers in other journals, for example, Angelo DeMarzo with IL-6, but from my lab, it was Veronica Gil et al. with IL-6, Andrea and I published on IL-23 in Nature, that basically these myeloid cells come in and secrete cytokines that in a paracrine fashion, released by the senescent cell, or the stressed cell, drive proliferation in the cells around that stressed cell by bringing myeloid cells into that arena. So, anyway, the long and short of it, we ran a trial. Actually, we're going to go through this step by step.

The first thing is we confirmed in this paper, well, in Nature, first of all, that a high neutrophil-lymphocyte ratio essentially means high myeloid cells in the tumor, specifically high CD15-positive, CD11B-positive cells, that are called by some myeloid-derived suppressor cells, but let's just call them myeloid cells in the cancer. That was the first thing. The second thing is that we also showed that a high neutrophil-lymphocyte ratio was associated with a transcriptome signature in the tumor biopsy at the same time as that blood test, a transcriptome signature of senescence, the key senescence drivers, and the SASP drivers, are highly expressed in tumors where, at the same time, the blood test shows a high NLR, supporting our hypothesis, but not proving it. We then gave this drug to men progressing on these ARSI drugs. They had really progressed on all known therapies for prostate cancer, most of them. And we confirmed this drug could, in many patients, but not all, substantially decrease these myeloid cells, these CD15, CD11B-positive, myeloid-derived suppressor cells, from the tumor biopsies using hyperplex immunohistochemistry.

And I think one of the biggest achievements for me in this trial was that we validated every single antibody with western blots, SI knockdown, IHC before and after the knockdown. This was a highly valued two and a half years of assay development doing four to 50 color IHC in a three-micron section for every immune cell you can imagine. We confirmed that CXCR2 in these biopsies, and this is unselected, pretty much CRPC, apart from the selection for the trial, that actually CXCR2 was only expressed in myeloid cells. Some have reported CXCR2 to be expressed in tumor cells. We have looked very, very hard in unselected biopsies from CRPC patients. We have not yet seen, although some have said otherwise, that we have not yet seen at any point CXCR2 expression on tumor cells, which is contrary to what's been reported by others.

Now, we then were able to show that the chemokine ligands that activate that receptor, that the drug blocks, are highly prognostic of bad rPFS and OS, two treatments, a bad time to castration resistance, a bad time to survival, again proving these chemokines are really important to the outcome from prostate cancer. We then ran a dose escalation phase one trial, where we gave patients higher doses of this drug, and notably confirmed that the drug blocks the myeloid cells getting in by blocking the chemokines, activating the myeloid cells to move down that gradient into the tumor. But we're also able to show in some patients that actually, we could reverse therapy resistance and, without causing any toxicity apart from uncomplicated neutropenia, actually show tumor regression, both radiologically with PSA falls in some patients, but not all, with all the responding patients, to my surprise, although I shouldn't have been surprised, being P53 wild-type. So, their P53 function was normal, as was their RB1 state, and they did not have any evidence of loss of CDK inhibitors, like P21 and P27.

So, to recap, we have shown that high NLR means lots of these myeloid cells in the tumor and SASP, which SASP stands for senescence-associated secretory phenotype, in the tumor RNA-seq, attracting these myeloid cells into the tumor down the chemokine gradient, which requires CXCR2, but possibly also other chemokine receptors. We've also proven that if you block these chemokine receptors, you can, in many, but not all, patients, block the myeloid cells getting in, and in some patients, you can get durable regressions of the tumor and PSA falls. In our hyperplex imaging, we have shown some changes in the macrophage counts as well as the myeloid cell counts, but, to my surprise and disappointment, we have not seen at the time of that second post-therapy biopsy any impact to date on the T-cells or B cells. I had hoped to see some change in T-cells or B cells, but we have not to date seen that.

So why is this paper important? Well, because there is really extensive evidence that the NLR is really, really key to therapy resistance for many therapies. It's also perhaps one of the first papers that really proves inflammation plays a really key role in fueling prostate cancer. And I think for the first time we now have a proof, in my opinion, that these myeloid cells are playing a role in prostate cancer biology. So, Andrea, I think I'll stop there. I'm happy to take your questions.

Andrea Miyahira: Thank you so much, Dr. De Bono, for presenting this. So I do have a few questions. One is, when you look at the higher peripheral blood neutrophil-to-lymphocyte ratio in mCRPC, do you have an idea of what the mechanism is causing this, and have you also looked earlier in disease states to see if this is predictive of outcomes?

Johann Sebastian de Bono: So, we are looking earlier, and that's quite an important question. And if you look at our review in Nature Reviews Cancer by De Bono et al, on inflammatory storms in prostate cancer, that is where we describe our thinking on the subject. But essentially, tumor cells, when stressed, are releasing cytokines into the blood, paracrine cytokines, that are essentially grooming the hemopoietic system to actually change that myelopoiesis, resulting in a high NLR. So, that is why we see high NLR in these subjects.

Andrea Miyahira: So, it's higher neutrophil numbers overall, not lower lymphocyte-

Johann Sebastian de Bono: And decreased lymphocyte function, correct.

Andrea Miyahira: Okay.

Johann Sebastian de Bono: And that involves factors like IL-8, interleukin eight, CXCL1, CXCL2, IL-6, neuregulin-1. As you know, IL-6 activates JAK/STATs, and neuregulin-1 activates AKT, and phospho-ERK, and also IL-23 made by the myeloid cells that we published on before, which activates JAK/STAT and AR signaling. So, these cells, through the cytokines they are making, these myeloid cells, are activating pathways that have been now shown by others to be involved in outcome and therapy resistance.

Andrea Miyahira: Okay, thank you. And your paper mentions several patients did have responses, so are you evaluating mechanisms of action, and do you think it's more related to a reduction of the tumor-supportive myeloid activities as opposed to reducing immunosuppression?

Johann Sebastian de Bono: So, we have not seen any evidence, sadly, that a CXCR2 blockade alters T-cell function or B-cell function to date. Although, I'll be honest, these are limited by the timing of that second biopsy, and maybe if we had done a biopsy at two months or three months further out, that would've been the case, but we haven't seen that to date. So, we believe, based on the preclinical data from the transgenic mouse models, that the way this is working is that you are preventing the myeloid cells from feeding the tumor through cytokines like IL-23, neuregulin-1, IL-6, and other cytokines that can activate things like JAK/STAT, ERK, AKT, and AR signaling.

Andrea Miyahira: Okay, thank you. And what are your next steps for clinical development? You mentioned that AZD5069 production has been discontinued, so what are you planning to do next?

Johann Sebastian de Bono: Yeah, this drug's patent life was very short. It was developed not for cancer but for lung inflammation. And because of that short patent life, AstraZeneca has not pursued its development. So, we are now looking at other, better compounds, in our opinion, that will inhibit this process better, and that we are pursuing further investigator-initiated trials that will start very soon to prove that we can do this better with those trials and really, as we have done with abiraterone, cabazitaxel, lutetium PSMA, and olaparib, transform the care of prostate cancer patients, but also understand disease biology, which really can be transformative. And I believe we can do that.

But I think the bigger question for us all is: If myeloid cells are important to prostate cancer growth, what else are they important for in the prostate cancer process? And as Angelo DeMarzo and others have told us for many years, more than 20 years, understanding inflammation may be the key to understanding prostate cancer. I truly believe that.

Andrea Miyahira: Okay. Well, thank you so much for this, for presenting, and congratulations again on this study. I look forward to your next results.

Johann Sebastian de Bono: Thank you so much. I pray that we can change and improve patient care and decrease suffering through this work. So, thank you for your support and the PCF's support.