AACR 2016

CD4 T-cell immunotherapy targeting MAGE-A3 is safe, and shows early clinical responses

Dr Yong-Chen William Lu - National Cancer Institute, Bethesda, USA

First I would like to thank Dr Winner and Julia Gunner and AACR for organising this media-rich programme. Today I’m here to present our research and clinical team from Steven Rosenberg’s lab surgery branch, National Cancer Institute. So I’m here to present the first in human clinical trial using the gene engineered CD4 T-cell targeting MAGE-A3. As a federal government employee I don’t have any financial interest to disclose.

Because this is a media-rich programme I would like to emphasise more about the basic, so make you more have the understanding of the fundamental question that we are facing for the immunology but also let you know that they have significant clinical application based on our finding.

For the immunology 101 the T-cell uses the important molecules called T-cell receptors to communicate with either the immune cell or tumour cell with a group of molecules called human leukocyte  antigen, HLA. So based on these two molecules, based on the and these two, the T-cell can tell the difference between the normal cell and the tumour cell and in this case it’s MAGE-A3.

For general understanding, there are two dominant types of T-cell - one is CD8 T-cell, the other one is CD4 T-cell. CD8 T-cell people usually call it as the cure T-cell because it has the ability to directly kill tumour cells but the CD4 T-cell don’t have such abilities so most people call them helper T-cells because they provide help such as using cytokines or other means to communicate with CD8 T-cells and to help CD8 T-cells to kill tumour cells. However, we believe, based on some clinical evidence and preclinical evidence, that the CD4 T-cell can sit in the driver seat doing the job to kill T-cells so for that we think you have significant clinical application.

So here I’ll just show you the general procedure that we used for the gene engineered T-cell. For a start we obtained the T-cells from the peripheral blood of the patients and then using either the protocol to introduce a gene into the T-cell. In our case we used a retrovirus. And then we need to redirect the T-cell to have the specificity against tumour cells so in our case we used the and then talking to the MLC molecule it’s the general term for the HLA. So for humans it’s called HLA, for the rest of the non-humans we call it MLC and then present a molecule in this case it’s MAGE-A3. Based on this code the T-cell can identify the tumour cell and then go on and go for it. The other approach is using the CAR, chimeric antigen, and for that you’re talking the protein by T-cell. Then after that we the T-cell and then those patients were treated with chemotherapy to clean out the space for T-cells and then we transferred the T-cells and then we may add to push expansion of the T-cell.

So again, I repeat again, the fundamental question we want to ask is currently all the T-cell therapies use either an unpurified mix of CD4 and CD8 T-cells or the purified CD8 T-cells. We want to ask the question whether CD4 T-cells can irradiate tumour cells the same as CD8 T-cells and it could have a very significant clinical application.

I’ll just mention a few key points, the detail is mentioned either in the abstract or in the press release. So, as mentioned before, HLA is a key molecule, it has to be matched with the TCR. For that, we specifically picked the DPB 041 because it presents in 60% of the patient population, this is the highest one. The second highest one is 50% and the rest are then down to less than 20%. The second important point that we picked the MAGE-A3 because of the frequency in a variety of cancer types, not just in melanoma but also in several other cancer types. I want to re-emphasise this because one of the major challenges for cancer immunotherapy, as mentioned in all the talks today, is that now it’s only limited to several types of cancer such as melanoma, renal cancer and leukaemia, lymphoma. So this might be the first clinical trial that’s going to attack a solid tumour such as urothelial or oesophageal cancer. So the most important thing, the major we’re seeing, is a good target because it’s not expressed in normal tissue.

For eligibility we need to have patients which are DPB 041 positive, the tumour needs to be MAGE-A3 positive and the patient must receive at least one first line therapy and fail with it and then we recruit those patients for our trial.

So here is a list of fourteen patients that have been treated. We started a slow dose escalation and the rule is we treat the first patient with 10 million cells, if we did not observe any dose limiting toxicity then we can increase the dose around threefold. Since we didn’t observe any toxicity so we kept increasing the dose to the maximum dose which is around 100 billion cells. So doing the dose escalation we treated eight patients and in the highest dose we treated an additional six patients. At the beginning we treated the majority of patients with melanoma and later we started to have patients with cervical cancer, oesophageal cancer and urothelial cancer and lastly two breast cancer patients. So far we have observed three partial responses, one is ongoing for fifteen months, the second four months and the third one seven months ongoing. So I’m going to show you the response for those three patients.

The first patient with cervical cancer that she has metastasis to two lymph node lesions, as shown here. After fifteen months they have shrunk to a very tiny, small area, as you can see, that’s almost disappeared and still ongoing. The second patient with oesophageal cancer that he has two lesions, two lymph node lesions, showing here and here. After four months they’ve shrunk to the small area but unfortunately there’s a new lesion occurred so the response was about four months. The last patient, patient number 11, that he’s a patient with oesophageal carcinoma and he has two lymph node lesions and three liver lesions and now it’s been seven months and the tumours have shrunk also to a small area for most of them and the pelvic lymph node lesion has been totally dissolved.

So I’ll skim most of the detail in the afternoon presentation but I’ll just tell you the key points. That is we think this first gene engineered CD4 T-cell therapy targeting MAGE-A3 is safe and it has shown early clinical response and we have started a phase II trial. So the major question we want to ask is if this T-cell therapy could be effective in different types of cancer. Thank you for your attention.

Thank you very much, Dr Lu. So there are several important messages, I think, in this presentation. The first is that you heard previously from Dr Hodi about the use of checkpoint antibodies that would essentially release the brakes of the immune system so the T-cells can do their job of attacking a cancer. This is the opposite approach, this is actually not trying to release brakes but essentially press on the accelerator, if you will, by using an expanded tumour specific population of T-cells that will then be able to hone in and attack the cancer. So you can see immediately how that might be complementary, that down the road you could see combinations of checkpoint antibodies applied along with these cellular therapies.

But there’s more to it than this in that here this is not a typical CAR T-cell and many of you have probably heard about the CAR T-cells where chimeric receptor targeting something that is not necessarily tumour specific is being attacked, such as CD19 in leukaemia and lymphoma. In the solid tumours it’s not possible to use similar kinds of targeting strategies because too many normal organs express those targets and the damage to the normal organs is going to be too significant. Here what’s happening is that these T-cells are being endowed with the capacity to recognise a target and go after it in a tumour specific way. And as long as the tumour target is expressed it doesn’t matter if the immune system previously thought it was going to be a tasty target for destruction, these T-cells have been changed so that they now think it’s a tasty target and that’s all you really need.

So this is a very promising conceptual strategy and obviously there are some early responses and that’s exciting. I have one question, though, which is these T-cells that are being transduced this way each have their native T-cell receptors as well. So these native T-cell receptors, when these cells get activated, might also get into the job of attacking something that they were designed to attack again. So do you see any usual side effects in these patients above and beyond what you might have seen as a result of TIL therapy?

Yes, I would thank Dr Winners for the wonderful comments. As for that, it’s not very simple to explain but, as I mentioned at the beginning, the TCR, T-cell receptor, contains two chains, one is the alpha chain and the other the beta chain and we use that to define the specificity of the T-cell. So they have the one that’s already inside every single T-cell and a new one that’s reintroducing and they could also mismatch between them. So to avoid that we added a safety feature which I didn’t explain here, I will explain that in the afternoon session. We used a mouse cancer region receptor. So because we used the mouse forced the TCR that we were introducing would maintain their specificity and you start to increase the safety. So we haven’t seen any non-specific reactivity against the patient’s normal cells.

Thank you. So we can open it up.

Yes, we have time for a couple of questions.

Alex from Medscape. Just a question – what are the typical side effects with this therapy? We know the side effects of CAR T-cells but the dose limiting toxicities, what are the main ones?

The toxicity issue, the same as others, as we mentioned before, those patients need to have the lymphodepletion regimen which is composed of the chemotherapy compound. Because of that they all have the same side effect as the chemotherapy reagents. Other than that, some of the T-cell therapies have more or less toxicity but I need to emphasise this one is much, much safer than CD19 CARs or others. We haven’t seen much side effect for that. I could only mention one is that most of the patients have higher fever, around 39-40o in twelve out of fourteen patients. This fever is very easy to control, it lasts around one to two weeks but it might be doing a good thing. It’s the same as me, as a human being, I catch cold and have a fever and then after that fight off all the flu virus and become healthy again. So it might be a good thing.

Other than that we have one patient with grade 3 toxicity with acute liver and renal injury but we couldn’t find out the cause because it’s one single patient case and she had some individual issues beforehand. So we don’t think that contributes to this T-cell therapy. And that’s all the toxicity that we’ve found so far.

Caroline Helwick, the ASCO Post. So no cytokine release syndrome?

Yes. So about that, first of all my boss, Dr Rosenberg, doesn’t believe in syndrome in this case because those patients did release cytokines and I will mention that in the afternoon session again, but the key issue is those cytokines are very acute and only last for a few days. The only that could correlate with that is the fever but the fever lasts one to weeks and we don’t find them associated with each other and we don’t think there’s any side effect correlating with the cytokines so we don’t think it’s really a syndrome.

OK, but my first question was going to be about your melanoma patients. So they’re previously treated, did they receive an immune checkpoint agent?

At that time they did not. Most of them, I need to double check with all the medical records. At that time, it’s around two years ago, so some of the patients had received checkpoints but some of them did not.

I was wondering how you would get a melanoma patient to go on this trial when they could have Dr Hodi’s potentially fabulous result with another drug.

After that we started to receive those melanoma patients that failed the PD-1 or CTLA4 checkpoint blockade so they come to us to seek for therapy. We have another arm, as you’ve been aware, that we use the TL therapy for melanoma. So the majority of patients will go for that and we think that’s a more curable option if those patients fail with the PD-1. So far from our study, because there are more patients there, it’s totally independent of the checkpoint blockade therapies, either for those patients that have been previously treated with PD-1 or CTLA4 that the response rate is no different.

So just finally, I wonder if Dr Hodi might pose a hypothesis why the melanoma patients wouldn’t have responded to this? Is there a no responses in your melanoma patients?

[Dr Hodi] There are many possibilities of resistant mechanisms to either checkpoint or CD4 directed therapies and we know from immunology, which you very nicely outlined, the interactions of helper and effector, so helping to kill tumour cells and the cells that actually kill are fairly complex. As he alluded to, to include in such a therapy checkpoint blockade and other combinations to increase efficacy would be an exciting area of investigation.

[Dr Lu] Sorry, I probably misunderstood your question. The reason we cannot address the question here and now is because those melanoma patients that have been treated with very low dose. So we only can count out of the six patients the three on high dose. Thank you.

Hello, Jacob Plieth, EP Vantage. You’re isolating CD4 T-cells, the problem I have is if you don’t do that, if you just expand all the T-cells, then your CD4 cells will still be there in the mixture. So what do you gain by effectively eliminating the CD8 cells from the mix?

That’s a very great question and I’ve pre-prepared the answer because it has been asked many times. The answer is as following: it’s very complicated but first we need to emphasise this is first genetically modified CD4 T-cell therapy and that gives us a hard time at the FDA. So we feel you put a CD4 T-cell into the CTA that would go in as a much more hard time. So the first concern is safety and safety is always our first priority. The second is the science. Because we are asking the question whether CD4 T-cells that are going to have anything to do with the tumour therapy so if we, as you suggested, if we use a mixture then if we see response we don’t know where it comes from, it could be from CD4 or it could be from CD8. But again your idea is great and we may go in to pursue that in the future.

Just a quick follow up on a separate issue. Can you please comment on the reasons why all of the NCI cell therapy studies have been suspended, including this one, as of recently?

I need to clarify first. It’s so we just have the clinical trial on hold it’s because the entire NIH is undergoing improvement for all our manufactured process for not just the cell therapy for but also others for. In the meanwhile we cannot enrol new patients, we cannot accommodate those patients for the cell survey so we put them temporarily on hold and we hope the situation will resolve very soon.

Time for one more question.

Thank you, Annette Breindl from BioWorld. Could the tumour cells afford to develop resistance by losing the MAGE-A3 antigen? And if they could do you see that or do they die if they are no longer expressing that antigen?

I’m sorry, could you repeat your question again?

Sure. Can the tumours become resistant to the treatment by no longer expressing the antigen that the T-cells are going after?

Yes, again that’s a great question. So it’s about the heterogeneity of the expression of MAGE-A3 as well as other different antigens. So one of the great challenges for immunology or cancer immunotherapy is not all the tumour displays the same antigen or the same mutation. So when we use a T-cell to attack the tumour cell what’s going on with all the expression of those antigens. So they have several parallel theories. So far the answer is we don’t know. We hope that they have some bystander effect that those T-cells can kill those non-expressers but we don’t know that so far. With the limit of our tools we hope to answer the question in the future but not in this study, not now. Thank you.