Angst in the IO Combo field – part 2 (lessons from #AACR17)

I posed this question regarding IO combinations in the last post, leading up to AACR:

“Why the perception of angst then? The sentiment has been summed up as “everything will work a little, so what do we research/fund/advance? How do we choose? How will we differentiate”?

I was mulling over these questions as I prepared remarks for Jefferies Immuno-oncology conference – the slides below are taken from the deck I presented.

Even the comment “everything will work a little” now seems to be an overreach. We could instead say: “most combinations won’t work at all”, meaning they won’t work better than anti-PD-1/PD-L1 monotherapy or anti-CTLA4 monotherapy, or, that they won’t work better than those therapies used in combination with standard of care.

Remember two years ago? We were going to take an anti-PD-1 to “release the brake” and add anti-4-1BB or anti-OX40 to “step on the gas”. While it is still early, this seems to be an empty paradigm. Why? Certainly the 4-1BB and OX40 pathways are intensely potent when used to drive T cells directly (e.g. anti-CD3 + anti-4-1BB in vitro or as used in a CAR-T cell). Is it too early to tell? Have the wrong patients been enrolled in trials? Are the antibodies no good? Is it the Fc? IS THE TUMOR COLD?

So here we go, onto the next paradigm, summed up in the phrase “make cold tumors hot”. What happened to stepping on the gas?

At AACR, Dan Chen (from Genentech, a Roche company) laid out the case for using not 1, not 2, not 3, not 4, not 5 … but up to 11 different therapeutics to successfully treat a given tumor – he exaggerated to make the point that none of the current immune checkpoint inhibitors (ICIs) should be expected to work in synergy with anti-PD-1 therapy, a priori. Why not? Because the ICIs are the really big levers, and the rest are smaller levers, where smaller simply means a pathway or biology that is less fundamental to immune anti-tumor responses than the ICIs. In order to see robust activity with these smaller levers, you need to apply them to carefully selected patients. An example was given by Jennifer Michaelson (from Jounce, and before that, Biogen) who stressed the need for biomarkers to guide the clinical application of an agonist anti-ICOS antibody (another gas pedal). The “cold tumor hot” gang that includes oncolytic virus approaches, onco-vaccine approaches, TLRs, STING and so on have not yet really articulated a strategy to identify patients likely to respond except in those tumor types associated with viral infection.

All of these accessory immuno-modulatory therapies, including the agonist antibodies (anti-4-1BB, anti-ICOS), the myeloid cell modulators (anti-CSF1R), the soluble mediator inhibitors (A2AR, IDO), the innate triggers (STING) we can lump as immune-oncology (IO) drugs, to distinguish these from ICIs.

The apparent strength of some ICI-standard of care combinations, and the apparent weakness of the early ICI-IO combinations has some startling implications. Let’s look at the current landscape:

Screen Shot 2017-04-07 at 7.36.33 AM

and here are the approved ICIs:

Screen Shot 2017-04-07 at 7.58.47 AM

Note the concentration of indications – melanoma, NSCLC, H&N, bladder, Hodgkin lymphoma, with single approvals in Merkel cell and RCC. Certainly the list will expand but if we concentrate on melanoma and NSCLC for a moment, we can outline the key challenges:

1) get the response rates up, and 2) prevent relapses. What do we mean by this? In advanced and/or metastatic melanoma the best overall response rate (ORR) using ICI monotherapy is about 30% in previously treated patients and up to 40% in patients naive to therapy (not previously treated with anything).  Within the responders there are two subsets of interest – durable responders (those that will survive for 3 years or more: about 20% of the responders) and relapses (those who initially responded, but then relapsed on ICI therapy: about 30% of the responders).  So if we just call out the durable responders we have between 6% and 8% of the original patient population in the trial receiving durable benefit. The idea of course is to get this number up.

Before turning to relapses, lets look at NSCLC.

Screen Shot 2017-04-07 at 8.19.53 AM


Again the lessons here are pretty clear – get the ORR up, improve durability of response, move ICI to early line therapy. The median overall survival data (OS) for advanced NSCLC patients treated with ICI looks very modest (12 months v 8-10 months with chemo) but this obscures the fact that the OS curve is pulled to the right by a relatively small number of durable responders: again, a small percentage of patients do very well.

So what next? Pharma is taking diverse approaches to improving ICI responses across many indications, To keep it simple I pulled out just two portfolios – those of Bristol Myers Squibb and Roche/Genentech.  We can note in passing that the Merck (US) portfolio is  relatively similar to the BMS portfolio, and the Astra Zeneca portfolio is  relatively similar to Roche, as is the portfolio from the Merck (Germany)/Pfizer collaboration.

Screen Shot 2017-04-07 at 8.34.43 AM


A few comments on these portfolios from the pharma level and as relates to small company programs and those who invest in those programs/companies:

Screen Shot 2017-04-07 at 8.39.35 AM

The latter point is critically important for biotech and investors to the consider and revisit often: how will my program generate compelling data? How is the indication landscape shifting as I spend 2-3 years moving a program forward? Is there a milestone of sufficient signal to rise above the noise of a thousand other ICI-ICI, ICI-IO, or ICI-SOC trials? I had the uneasy experience of walking through the poster sessions at AACR17 last week, past lovely bits of work that no one was paying any attention to. That’s a lousy feeling for the person presenting the poster and a very lousy place to be if you are a small biotech company.

A final slide:

Screen Shot 2017-04-07 at 8.39.46 AM


It’s an incomplete list of course.

More on resistance and relapses next time.

stay tuned.

Immuno-oncology (IO) combination therapy- why the angst?

Thoughts triggered by discussions over the last month or two, perceived sentiment on social media, reaction to clinical updates, and pre-AACR butterflies.

In 2015 Gordon Freeman of the Dana Farber Cancer Institute, one of the discoverers of the PD-1/PD-L1 axis, rang me up and asked if I would help write a review with he and Kathleen Mahoney, an oncologist doing a research rotation in his lab. We ambitiously laid out the argument that PD-1/PD-L1 directed therapeutics would be the backbone of important combination therapies and reviewed the classes of potential combinatorial checkpoints ( We covered new immune checkpoint pathways within the Ig superfamily, T cell stimulatory receptors in the TNF receptor superfamily, stimulatory and inhibitory receptors on NK cells and macrophages, targets in the tumor microenvironment (TME), and so on. Importantly we also stopped to consider combinations with “traditional” cancer treatments, e.g. chemotherapy and radiation therapy, and also with “molecular” therapeutics, those directed to critical proteins that make cells cancerous. Regardless, it’s fair to say that we believed that pairing an anti-PD-1 mAb or an anti-PD-L1 mAb with another immuno-modulatory therapeutic would quickly yield impressive clinical results. A massive segment of the IO ecosystem (investors, oncologists, biopharma) shared this belief, and largely still does. Those stakeholders are betting clinical and R&D resources plus huge amounts of money on the promise of IO combinations. After all, the first IO combination of anti-CTLA4 mAb ipilimumab and anti-PD-1 mAb nivolumab has dramatically improved clinical response in advanced melanoma patients and to a lesser extent in advanced lung cancer patients. The downside is additive toxicity, and so the palpable feeling has been that new IO combinations would give a similar efficacy bump, perhaps even with less toxicity.

It’s now about two and a half years since we began drafting that paper and the inevitable letdown has set in. What happened? Let’s cover a few issues:

- Several marque IO combinations have been disappointing so far. Last year we saw unimpressive results from urelumab (anti-4-1BB) in combination with nivolumab (anti-PD-1) and of epacadostat (an IDO inhibitor) paired with pembrolizumab (anti-PD-1).

- Monotherapy trials of therapeutics directed to hot new targets (OX40, CSF1R, A2AR etc.) did not produce any dramatic results, forcing a reevaluation of the potential for truly transformative clinical synergy in the IO combination setting.

- These first two points also reminded the field of how limited preclinical mouse modeling can be.

- Combinations of standard of care with anti-CTLA4 mAb ipilimumab and with PD-1 pathway inhibitors have begun to show promising results, raising the efficacy bar in a variety of indications. There have been several startling examples: the combination of pembrolizumab plus chemotherapy in first line lung cancer, which doubled response rates over pembrolizumab alone; the combination of cobimetinib (a MEK inhibitor) with atezolizumab (anti-PD-L1 mAb) in colorectal cancer (MSS-type) which produced clinical responses in patient population generally non-responsive to anti-PD-1 pathway inhibition; the combination of atezolizumab plus bevacizumab (anti-VEGF) in renal cell carcinoma, showing promising early results; and so on.

- We can add the realization that relapses are a growing issue in the field, with approximately 30% of anti-CTLA4 or anti-PD-1 pathway treated patients eventually losing the anti-tumor response.

Note here that all of this is happening in a rapidly evolving landscape and is subject to snap-judgment reevaluation as clinical data continue to come in. For example, rumors that IDO inhibition is working well have been spreading in advance of the upcoming AACR conference. Indeed the clinical work on all of the immuno-modulatory pathways and IO combinations has increased, and the race to improve care in diverse indications continues. There will be additional success stories.

Why the perception of angst then? The sentiment has been summed up as “everything will work a little, so what do we research/fund/advance? How do we choose? How will we differentiate”? Such sentiment puts intense pressure on discovery, preclinical and early clinical programs to show robust benefit or, and perhaps this is easier, benefit in particular indications or clinical settings. I started thinking about this recently when a friend of mine walked me through a very pretty early stage program targeting a novel pathway. It was really quite impressive but it was also apparent that the hurdles the program would have to clear were considerable. Indeed it seemed likely that validation of the therapeutic hypothesis (that this particular inhibitor would be useful in IO) would not come from preclinical data in mice (no matter how pretty), nor from a Phase 1 dose escalation safety study, nor from a Phase 1 expansion cohort, but would require Phase 2 data from a combination study with an anti-PD-1 pathway therapeutic. That is, 5+ years from now, assuming all went smoothly. To advance such a therapeutic will take intense focus in order to build a fundable narrative, and will require stringent stage-gates along the way. Even then it will be very hard to pull it off. If this reminds you of the “valley of death” we used to talk about in the biotech realm, well, it should.

What should we look for to shake up this landscape? As mentioned, this is a rapidly evolving space. We have already seen a shift in language (“step on the gas” vs. “make a cold tumor hot” is one good example), but let’s list a few:

- “Cold tumors” have no immune response to stimulate. Making them “hot” is a hot field that includes oncolytic virus therapeutics, vaccines, “danger signals” (TLRs, STING, etc), and, to loop back around, chemotherapy and radiation therapy.

- Relapsed patients – as noted above we are seeing ~30% relapse rate in immunotherapy treated patients. Understanding the basis for relapse is a promising field and one that an emerging therapeutic could (and very likely will) productively target.

- Targeting the TME in cold tumors and in unresponsive tumors (the difference is the unresponsive tumors look like they should respond, in that they contain T cells). This is a vast field that covers tumor cell and stromal cell targets, secreted factors, tumor and T cell metabolism and on and on. One can imagine a setting in which a particular TME is characterized (by IHC, Txp or other means) and the appropriate immuno-modulatory therapeutics are applied. We see this paradigm emerging in some indications already. This would certainly be useful as a personalized medicine approach and could be an excellent way to position an emerging therapeutic.

We could go further to talk about the neoantigen composition of particular tumor types, the role of the underlying mutanome, the plasticity of the TME (it’s a chameleon), metabolic checkpoints, and other, potentially novel, targets.

All of this is under intense and active investigation and important data will emerge in time. Until then, nascent immunotherapy programs need to tell a clear and compelling story in order to attract the interest of investors, biopharma and ultimately, oncology clinical trialists. Those that fail to develop a compelling narrative are likely to struggle.

I’ll just end on a few narratives I really like for IO combinations going forward:

- the role of innate immunity in activating immune responses and expanding existing responses (e.g. immune primers like STING agonists and NK cell activators like lirilumab)

- the role of adenosine in maintaining an immunosuppressed (ie. non-responsive) TME (thus inhibitors of A2AR, CD39, CD73)

- the role of beta-catenin signaling in non-responsive tumors (while carefully selecting the mode of inhibition)

- the role of TGF-beta activity in resistance to PD-1 pathway therapeutics (again, with care in selecting the mode of inhibition)

of course at Aleta we’ve charted a different course, ever mindful of the need to focus where we see clear yet tractable unmet need. so we’ll see, starting with AACR in early April, kicking off an active medical conference season.

stay tuned.

The conference season grind

With the International Immunotherapy meeting just behind us and #SITC2016 (or #SITC16, we’ll see which wins) looming, we are just at the beginning of the long grind of medical and bio-partnering conferences that characterizes the academic year. I’ll be attending a bunch of these (see below) and telling the Aleta Biotherapeutics story as we develop novel methods of targeting cellular therapeutics for the treatment of cancer. It’s a compelling story we think and we are happy to discuss it, just give a shout to Paul or Roy: and

Here’s a partial conference list:

-  Society for the Immunotherapy of Cancer Annual Meeting (SITC); Bethesda, Nov 10-13

-  American Society for Hematology Annual Meeting (ASH), SanDiego, Dec 3-6

-  JP Morgan Healthcare (JPM), San Francisco, Jan 9-12, 2017

-  World Adoptive T-Cell Therapy Summit, Lisbon : Feb 6-7, 2017 —> Invited talk featuring Aleta cell therapy technology 

-  Sach’s IO Biopartnering, NYC, March 28, 2017 —> IO combinations session chair

-  American Society for Cancer Research Annual Meeting (AACR), Washington: April 1-5, 2017 —-> Aleta cell therapy abstract submitted

As always I’ll be live tweeting every conference @PDRennert.

To discuss  Aleta and other projects of interest just get in touch. We are looking forward to a highly productive conference cycle.  Cheers – Paul

part 2: Second CRI-CIMT-EATI-AACR International Cancer Immunotherapy Conference: Translating Science into Survival

Highlights from Day 2

 The Analysis of Tumor Microenvironments:  gaining depth and granularity.

Dr. Wolf Fridman (Cordeliers Research Centre, Paris), abstract 1A10, presented tools for the analysis of cell populations in the TME of colorectal cancer (CRC) and clear cell renal cell cancers (RCC). The CRC analysis produced 4 distinct subtypes, with wildly variable cell types and pathogenic pathways supported by the dominant cell populations. The subtypes aligned with standard CRC immunohistochemcial analyses, with molecular classifications, and with prognoses. Several subtypes were readily apparent in RCC. The results provide a compelling framework for cancer classification across indications and may allow the more precise pairing of immuno- and other therapeutics in a wide variety of cancer indications.

Dr. Shimon Sakaguchi (Osaka University), abstract 1A12, elucidated distinct T regulatory subsets in tumors and applied a Treg classification scheme to CRC, an indication in which the role of Tregs has been controversial. By carefully delineating suppressive and (paradoxically) inflammatory Treg populations, diverse roles in different CRC subtypes were proposed. Notably, the inflammatory subset appeared in the context of tumor invasion by bacteria that can access the tumor interstitial space as the mucosa is breached. Importantly, anti-CCR4 mAb treatment could selectively deplete the suppressive Treg subset and restore anti-tumor immunity in their models.

Dr David Denardo (Wash U School of Medicine, St Louis), abstract 1A14, introduced the hyper-fibrotic TME that characterizes pancreatic ductal adenocarcinoma (PDAC). The TME is composed of a collagen-I rich desmoplastic stroma that houses large numbers of immunosuppressive cells, creating both physical and biological barriers to T cell entry into the tumor. Fibrosis is induced and sustained by the TGFbeta pathway, leading to hyper-activation of focal adhesion kinase (FAK). A FAK inhibitor had monotherapeutic activity in a PDAC mouse model, leading to collapse of the fibrotic architecture and loss of the immunosuppressive myeloid cell compartment. In combination, FAK inhibition was synergistic with anti-PD-1 and anti-CTLA4 in PDAC mouse model that does not respond to either therapeutic given as monotherapy.

In related TGFbeta therapeutic development, Dr Maureen O’Connor-McCourt (Formation Biologics, Montreal), abstract B058, hosted a poster on a new and novel TGFbeta TRAP protein that is selective for TGFbeta isoforms 1 and 3 (but not 2).  Merck Germany has an anti-PD-L1/TGFbeta TRAP bispecific (presented a few weeks ago in Boston). This remains a very hot area.

Given the positive CAR T news from KITE yesterday, their poster on TCR technology is worth a quick mention. Lorenzo Fanchi, abstract B044, hung a poster detailing the derivation and subsequent creation of patient-specific TCRs targeting the antigens mel526, mel624, and mel888. The TCRs were challenged in vitro and in vivo (in mouse) with PDX-matched tumors and HLA-matched tumor cell lines (the HLA-typing was not disclosed). Although mouse, the TCR therapy (20 x 1oe6 cells) sustained long term survival in 2/6 animals, which was an encouraging result.

Second CRI-CIMT-EATI-AACR International Cancer Immunotherapy Conference: Translating Science into Survival

A Few Day 1 Highlights

Ton Schumacher (Netherlands Cancer Institute), abstract IA04 ,has discovered a novel regulator of PDL1 expression called PD-L1M1. PD-L1M1 associates with PD-L1 and modulates the T cell inhibitory function of PD-L1. The protein is expressed ubiquitously, so unclear if this finding has therapeutic implication.

Michael Peled and Adam Mor (NYU School of Medicine), abstract A059, had a poster on molecules that interact with the cytoplasmic tail of PD-1 using high resolution Mass Spec. Two proteins were highlighted on their poster: EFHD2 and SH2D1A. EFHD2 co-localized with PD-1 and was essential for clustering and signal transduction (thus, ablation of EFHD2 blocks PD-1 mediated inhibitory activity). SH2D1A had the opposite function as evidenced by increased PD-1 inhibitory signaling when SH2D1A was knocked down and reduced PD-1 inhibitory signaling when overexpressed. SH2D1A physically competed with SHP2 for access to the PD-1 cytoplasmic tail.

Dario Vignali (U. Pitt School of Medicine), abstract IA05, focused on several emerging immune checkpoints. The first, IL-35, was investigated using anti-IL-35 antibody in various tumor models, with very nice results (similar to anti-PD-1). I liked the neuropilin story – this is a Sema4a binding protein and was offered up as a central control node for Treg activity. NRP1 controls Treg T cell expression of IFNgamma, acting in cis and in trans (so self-regulation and neighborhood regulation). Of interest he identified subsets of melanoma and H&N cancer patients having high levels of NRP1 in the TME, so this is perhaps an actionable finding.

Susan Kaech (Yale Univ Med School), abstract 1A07, presented data showing that the PEPCK overexpression ups the anti-tumor activity of T cells in the TME, thus showing that T cells – if given the tools – can co-opt the same metabolic pathways (lactate, fatty acids) used by tumor cells in the tumor microenvironment (TME). A consequence of this metabolic checkpoint is the upregulation of PD-1 via fatty acid signaling through the PPARs, delta I think. Of interest is that the metabolic switch is supported by gross upregulation of CD36, a fatty acid active transporter, on T cells in the TME.

Greg Delgoffe (U Pitt Cancer Inst), abstract IA08, picked up this general theme, demonstrating that T cells dividing in the TME rapidly lose mitochondrial (MT) mass, and therefore their ability to metabolize glucose ( a T cells preferred energy source). This is a failure of MT biogenesis, due to the downregulation of PGC1alpha, which is required for the process. In the TME, T cell PGC1alpha expression is regulated by AKT – robust AKT signaling leads to PGC1alpha downregulation. If note, PGC1alpha transgenic T cells retain high proliferative activity, do not lose MT, and are highly active Teffector cells.

KITEs CAR T cells

Screen Shot 2016-03-29 at 4.54.24 PM

Love that cover, if you like the band they live here:

I don’t think we’ll see the very best of the CARs at AACR16; that may have to wait until ASCO, but an abstract jumped out at me today as it addresses one of the issues I discussed in the last post (  As a brief refresh the last post was concerned primarily with 2 issues:

1)  The cell population used to make the CAR T cell prep, with emphasis on a mixture of naive CD4+ T cells plus central memory CD8+ T cells, based on work coming out of labs with close ties to Juno.

2)  The ability of CARs once injected to expand appropriately if antigen were not abundant (or as abundant as CD19 is on leukemia and lymphoma cells). This is a common observation not tied to any particular lab or company.

An interesting paper to be presented at AACR is by Timothy Langer et al. from the lab of Adrian Bot at Kite Pharma, in collaboration with the NCI (Steven Rosenberg, Steven Feldman, James N. Kochenderfer). The abstract is #2305: link.

What Langer et al. describe is the phenotype of Non-Hodgkin Lymphoma patient T cells prior to their transduction/expansion and the relationship of this phenotype to outcome.

In summary they took peripheral blood containing circulating T cells from the patients, then isolated the T cell fraction.  Using textbook CAR T techniques they activated the T cells then transduced them with a retroviral vector that encodes their CAR19 construct. Then they expanded the T cells further in culture with IL-2 until they reached the target “dose” of the T cells (which serve as the “drug”).

Importantly they froze down the starting peripheral blood prep and the CAR T cells then went back, thawed them both out (patient by patient matched samples) and analyzed by flow cytometry. Basically they set out to compare the starting point, the “drug” and then the result (i.e. the outcome for the patient). Now some caveats: sample size is small (n = 14), and it’s not clear to me which trial this is (so what the outcome data are). Regardless, it’s a starting point for a dataset that will surely grow over time.

OK, what did they find?

- all 14 samples yielded useful CAR T cells (i.e. were transduced and expanded successfully)

- all patients samples were different, in particular the ratio of CD4+ T cells to CD8+ T cells varied from patient to patient.

- since the abstract doesn’t show the data we don’t know those ratios, we might guess that there were more CD8s than CD4s in most patients, but we don’t know at this point

- however, the CD4/CD8 ratio in the starting prep for each patient was positively correlated with the CD4/CD8 ratio in the “drug”

- the starting T cell population generally had similar percentage of effector T cells (by definition these include activated T cells, effector memory T cells, and central memory T cells of both the CD4 and CD8 lineage) and naive T cells (not activated at all).

- the CAR T cell preps showed a shift to a greater percentage of naive T cells, that is, either they preferentially expanded (likely) or the effector T cells died out (less likely).

Now it gets interesting:

All of the CAR T cell preps were active in vitro and were delivered to their respective patients. Clinical responses occurred regardless of the CAR T cell prep phenotype including CD4/CD8 ratio or effector T cell/naive T cell ratio. So whatever was determining successful responses in the patients or unsuccessful responses – this wasn’t it. They are now further characterizing the cell phenotypes to look at other parameters, e.g. PD-1 expression. They reference these trials as using essentially this exact protocol: NCT02348216, NCT02601313.

This is a markedly different story than told by the Riddell lab recently, as we discussed last time ( So, as Jules Winnfield (the immortal SLJ) said in Pulp Fiction, “well then allow me to retort…”

I’ll be fascinated to watch how this plays out among different groups, each of which is doing its’ very best to get the manufacturing process highly optimized to produce the best possible CAR Ts for patients.

stay tuned

CAR T updates – tangled tales unwound

Last month we saw a biomedical media campaign go a bit off the rails. A press release from the American Association for the Advancement of Science (AAAS: see for example and the Fred Hutchinson Cancer Center, was picked up by multiple media outlets who quickly spun the story of CAR-T-cell mediated rapid and complete clearance of B cell leukemias and some lymphomas from very ill patients and turned it into the “cancer cured” sort of headlines that serve as great click-bait but don’t do much to really educate the reader.

But what first caught my eye was an odd distortion of the data as presented in the session entitled “Fighting Cancer and Chronic Infections with T Cell Therapy: Promise and Progress” (see Several credible sources were telling very different stories about the progress presented. To take one example, BioWorld Today told the story of the clear benefit of using naive T cells as the recipient for cellular therapy, while FierceBiotech (and many other outlets) focused on the benefit of using memory T cells instead (see Indeed the claim was made that even a single memory T cell could affect a cure – which was not really the point, or an important conclusion of the presented works.

It follows that the pressers were used to talk up CAR T cell company stocks, which have been languishing along with the rest of biotech.

All of this came across as garbled and confusing. I found it all very frustrating.

So now I’ve gone through the abstracts presented at AAAS and some of the primary literature, and I’ve a Cliff Notes version of what data were actually presented and what the data mean and don’t mean. I seems clear that the confusion regarding the results arose from the oversimplified weaving of two talks (by Dirk Busch and by Steven Riddell) into one tangled “story”. Lets untangle the knot and follow the threads.

Riddell’s work is closely followed in the CAR T field – not surprising as Dr. Riddell, from the Fred Hutchinson Cancer Center in Seattle, is a technology leader and a cofounder of Juno Inc. The story presented at the AAAS symposium is interesting but perhaps more controversial than one might have gathered from the press reports. Some of the work was recently published ( They start with the observation that in all reported CAR-19 clinical trials, patients have received back a random assortment of their (now CAR-transduced) T cells, meaning that the cell population is a collection of naive T cells, effector T cells and memory T cells representing both the CD4 and CD8 T cell lineages. This introduces a variable into therapy, as different patients are likely to have different percentages of these various T cell subsets. Indeed there is quite a list of variables that may impact the efficacy of CAR T cell treatment including baseline immune competence, prior treatments and antigen load. With this in mind Riddell and colleagues are trying to control the one variable that they can, which is the composition of the transduced T cells going into the patient. By analyzing CAR cell subsets for tumor cell killing function they arrive at the “most potent” combination of CD4+ T cells and CD8+ T cells and conclude that the findings will be important for the formulation of CAR T cells therapeutics for use in patients.

The data in the paper are derived from normal donor and cancer patient PBMC samples that are tested in vitro using cell culture assays and in vivo using humanized mice (NOD/SCID/yc-deficient mice; NSG) reconstituted with T cells and tumor target cells (Raji) that express CD19. The CAR T construct is a “generation 3″ CAR having CD28, 41BB and CD3 signaling domains downstream of the well-studied FMC63-derived anti-CD19 scFv.

Some results:

- substantial differences were seen in the T cell populations between normal donors and cancer patients, with most patients having a higher percentage of CD8+ than CD4+ T cells.

- patient samples also contained more memory T cells than did normal donor samples. A further refinement to the memory T cell definition allows one to identify effector memory and central memory T cells. The latter are a long-sustained population of antigen-educated T cells that contribute to immunological memory, such as one retains after a vaccination against a virus for example.

- both CD4+ and CD8+ T cells were readily transduced with the CAR-19 construct, and when presented with target cells in vitro both cell types responded. CD8 T cells mediated target cell lysis more effectively than CD4+ T cells, but the latter proliferated more vigorously and produced more pro-inflammatory cytokines such as IFNy and IL-2.

- among the CD4+ subset, naive T cells (those not previously antigen-activated) produced more cytokines than the memory cell subsets. In vivo, naive T cells were more potent in controlling tumor growth than central memory T cells which were in turn more potent than effector memory cells.

- similar analyses of CD8+ cells revealed that, of the three subsets, central memory CD8+ T cells were the most potent in vivo, a result that was most closely associated with the enhanced proliferation and expansion of this subset.

- the activity of CD8+ central memory T cells was further enhanced by the addition of CD4+ T cells, notably those of the naive subset. This effect was seen using cells from normal donors and cells from B cell lymphoma patients (specifically, Non-Hodgkin Lymphoma (NHL) patients). The improved in vivo activity was due to enhanced proliferation and expansion of T cells in the NSG mouse model, specifically an increase in the peak of CD8+ cell expansion, in line with clinical results (see below). I’ll note as a reminder that all of the available clinical results are from CAR T cell populations that had not been sorted into naive and memory subsets. Also, many researchers in the field believe that naive T cells (CD4 and CD8) have the best proliferative capacity and potency.

Regardless, the Riddell work suggests a straightforward improvement in the ability to create more potent CAR T cell preparations for use in the clinical setting. There are some caveats however. In the in vitro and in vivo models used, antigen (CD19) is abundant, even in the NSG mouse, due to robust expression of rapidly dividing CD19+ Raji cells. As noted earlier, antigen availability may be an important limiting feature for some patients, and may be more important than the composition of the T cell subset tested. Fortunately the relative importance of these variables could easily be examined in vivo by using sub-optimal Raji cell numbers, or using transfected cells with different levels of CD19 expression, to vary the antigen load.

The Busch study at the same symposium was notable for dispensing with CD4+ T cells altogether and using just CD8+ central memory T cells to control CMV infection (that can occur following allogeneic hematopoietic stem cell transplantation). Nearly all of this work has been performed in mouse models, with a small number of patients treated under compassionate use protocols (see e.g. In the mouse models very small numbers of antigen-specific memory T cells can expand to control viral infection, and this has been taken as evidence (in the popular press mainly) that similar technology could be applied in the CAR T setting. However, numerous studies have shown conclusively that very large-scale expansion is required to achieve optimal potency, to a degree that would seem beyond the capacity of a small number of cells or a single cell. Further, studies in acute lymphocytic leukemia patients presented by Carl June last fall at the Inaugural International Immunotherapy meeting in NY showed that clonal selection and perhaps competition was a component of successful therapy for some patients, a process that would be eliminated or reduced by using a limited cell number in preparing the CAR T cells. The Busch study makes the further argument that central memory CD8+ T cells themselves possess “stem-ness”, that is, they can give rise to functionally diverse CD8+ T cell lineages and as such should have no limit to their proliferative capabilities. While this was demonstrated convincingly in mouse models it would seem a difficult finding to translate to the CAR T setting, although the work may find utility in the adoptive cell transfer setting (e.g. of selected but not transduced T cells, such as tumor infiltrating T cells).

The “stem-ness” concept reminded me of older literature that aimed to dissect the basis for long-lived CD8+ T cell memory in the context of viral immunity (see here for a recent review: There were at one time two broad classes of thought – first, that such memory required a consistent supply of antigen, for example, a depot that periodically re-stimulated the antigen-specific T cell population. The second school of thought, more reminiscent of the Busch finding, was that memory CD8+ T cells were self-renewing, and therefore did not require life-long antigen stimulus. The “big bang” hypothesis of T cell memory development, a hypothesis that the work of Dr. Busch and colleagues has definitively supported (see: holds that once stem-like T cell memory is created, these cells can be used just like stem cells, i.e. to reconstitute cellular function, in this case, the ability to control viral infection.

Let’s get back to CAR T cells. Recent work has demonstrated clearly that the establishment of persistence in cellular therapy requires a robust response to abundant antigen. Only then can CD8+ T cell memory develop and from that point on be maintained. This observation informs the next set of studies, presented at the Clinical Application of CAR T Cells conference (#CART16 – held at the Memorial Sloan Kettering Cancer Center, the Adoptive T-Cell Therapy Congress held in London ( and the Advanced Cell Therapy Symposium ( held at Guy’s and St Thomas’ NHS Foundation Trust and King’s College, also in London. Much of the work presented highlighted at these meetings addressed attempts to move CAR T cells into solid tumors.

Here I am a little hamstrung, as I’m relying on information presented on slides (as shared on Twitter by @JacobPlieth @VikramKhanna and others). Let’s try to define some themes here regardless.

Jacob has reviewed some #CART16 data: Please see that link for his viewpoints.

First, to stick with CAR19 therapeutics, we have some posted Novartis data on responses in Non-Hodgkin Lymphoma (NHL). NHL is comprised of diverse B cell lymphomas, some of which are highly refractory to treatment. Examples of the refractory class include diffuse large B cell lymphoma (DLBCL) and follicular lymphoma (FL) among others. Here we see some rather impressive results treating these lymphomas:

Screen Shot 2016-03-20 at 9.55.18 AM

The data are hard to read, but let’s pull out some numbers from the table. Note that essentially all patients got the optimal dose of 5 x 10e8 cells (1 exception) and that the peak cellularity is defined as %CD3+/CAR19+ cells in peripheral blood. We can therefore look at expansion, time to peak cellularity and outcome:

Screen Shot 2016-03-20 at 10.00.38 AM

There seems no correlation between day to peak and outcome, unless it is very short – day 1 or 2 – and even then that is likely due to abortive expansion. If we arbitrarily set 10% as an exploratory setting with which to parse the %CD3+/CAR19+ data we quickly see that above 10% (black line), half of the patients responded, which below 10% only a third of patients responded.  So expansion is important, as we already knew. With respect to the earlier discussion, we do not know the critical variable at work here, be it CAR T cell persistence (likely), CAR T cellular composition (per Riddell), patient variability, antigen density, or something else.

The FL data are a little bit more confusing:

Screen Shot 2016-03-20 at 10.05.07 AM

Again we can pull out some of the data:

Screen Shot 2016-03-20 at 10.07.43 AM

And now we are really hard-pressed to see any correlation between outcome and peak cellularity, no matter where we might draw the arbitrary line for analysis. What data is missing? I suspect it is a measure CAR T cell persistence over time, as this is most often associated with positive response. We should note that CD19 is an unusual target antigen in that it is expressed on the cancer cells (B cell leukemia or lymphoma) and on normal B cells that we can deplete without undue harm to the patient.

Other B cell antigen targets are under development as CARs, including CD22 and BCMA. BCMA is expressed on plasma cells (relatively uncommon B cells that secrete antibodies) and on essentially all multiple myeloma cells. Early promising results generated using a BCMA CAR to treat multiple myeloma were presented at ASH (

Screen Shot 2016-03-20 at 9.39.58 AM

Updated results are a little less encouraging as the complete response patient (#10) has since relapsed as was reported at #CART16. It is unclear if advanced multiple myeloma is simply more refractory to CAR treatment, if the lower cell number infused led to poor persistence, if the CAR were different, or if antigen load was too low. Thus we are again faced with multiple variables to assess.

So now we can ask what happens when there is little or no persistence, which is the case with most CARs directed to solid tumors. This is data from Nabil Ahmed and Stephen Gottschalk from Baylor College of Medicine. This group is collaborating with Celgene on cellular therapeutics. Here we see the results of treatment of HER2+ advanced solid tumors with CAR-HER2 T cells. The response is minimal.

Screen Shot 2016-03-20 at 10.12.07 AM

This is very likely due to the very short persistence of the CAR-HER2 T cells, in most cases gone in a week or so.

Screen Shot 2016-03-20 at 10.12.27 AM

Interestingly, analysis of resected or biopsied tumor after treatment revealed that the CAR T cells had migrated preferentially into the tumor, but had not proliferated extensively. Novartis presented nearly identical data on an EGFRvIII targeting CAR T cell study at the Boston PBSS Immuno-Oncology Workshop (, and similar data has been presented on a host of solid tumor targets.

To return briefly to CD19+ tumors, it was reported recently that the response to CAR19 therapy in chronic lymphocytic leukemia was about 25%, but that all of those responders were durable complete responders (i.e. potential cures). Why the seemingly digital nature of response here? Again this is most likely due to CAR T cell persistence which itself is most likely a reflection of antigen load (among other variables). With this in mind I was struck by a slide from Seattles Childrens Hospital (I’m not sure from which meeting):

Screen Shot 2016-03-20 at 10.17.40 AM

In point #3 the presenter is basically suggesting injecting an artificial antigen presenting cell expressing CD19, i.e. increasing the antigen load.

We can conclude by saying that there is a fundamental issue with CAR T cell antigens – those that are tumor specific are either not abundantly expressed and/or have been removed during the course of therapy. This is an issue that may not be solved by adding 4-1BB or IL-12 or anti-PD-1 antibody or whatever other immunological “help” one might envision. This issue impacts the entire field, which is why we now see analysts who once talked of the emergence of dominant CAR T platform companies now wondering who will win the CAR19 race to the finish line. That is still a noble race to run, but the patient numbers cannot justify the number of companies competing for the prize. Yet change will come and progress will be made…

What to do?

stay tuned.

ps. thanks again @JacobPlieth @VikramKhanna and others for kindly sharing slides (and getting great seats at conferences!)

thoughts on attending the Immuno-oncology Biopartnering conference in NYC

Thought bubbles on the way back from the Sach’s IO Biopartnering meeting, a very interesting event:


-  The Sach’s Conferences are small, focused and diverse. Sach’s is much better known in Europe than in the US, and the NYC location brings European firms and companies into the fold. Therefore you meet interesting people you might otherwise never come across.


-  When in provincial Cambridge Massachusetts, i.e. the village of Kendall Square and environs, we visualize investors as the VCs that are in the square, across the river in Boston, or out in Waltham. With more experience one might add investment and hedge funds and the associated bankers of greater Boston. What you find in NY is an entirely new investment ecosystem, one that is very active and a good crowd to know. fyi, you won’t find these guys at Area 4 or Voltage.

-  It follows that similar ecosystems exist on the west coast, in Europe (France stands out), England’s golden triangle (London >>> Cambridge >>> Oxford), and in diverse far flung locales (Tel Aviv, Singapore, Hong Kong to name just a few). While many firms have preferred investment geographies, all of them are aggressively spreading their money further afield. It is well worth getting to know as many investors as possible, worldwide.

Science and deals:

-  The conference cliche was “good science wins”. This sounds trite but contains an essential kernel of truth applicable to the suddenly challenging financial landscape. We should remember that, since 2014, VCs and other investors raised massive funding rounds. This money will need to be invested during each fund’s life cycle (these average about 5 years). Equally large hordes of cash are sitting inside of large biopharma and on the balance sheets of newly IPO’d companies. The upshot of all this is that “good science” has a good chance of attracting financing.

-  A second theme, less of a cliche and more of a hope, was that M&A activity will surge in 2016 due to the suddenly more affordable valuations of small and midsize biotech.  In this scenario the biggest biopharmas will get bigger by buying up a bunch of smaller companies – but do we believe they will all buy “good science”? Investors may harbor a doubt about this, but we’ll see.

Good science:

-  So what is “good science”: I’d argue that in the IO space good science in small biotechs has one or more of the following qualities:

  • Novel IO assets, IO combinations and bispecific antibodies that will synergize with existing checkpoint inhibitors.
  • Next generation vaccines, oncolytics, small molecule inhibitors that have tumor cell cytotoxic activity and/or immune activating activity. This is a broad class with assets of diverse MOA and quality, but winners will emerge here.
  • CAR-T and TCR 5.0 (or 6.0) – whatever that will be.
  • Personalized medicine solutions that guide treatment choices.

What can we observe now? Can we recognize “good science”?

Lets dial back a week or so to the response to three talks given in the cellular therapy session at the recent AAAS meeting:                           (

The three talks were widely (wildly in some cases) misrepresented, misinterpreted or both as documented in detail on social media  - see for example        …

and …

and …

Fortunately two of the three presentations are now represented by publications. The papers are in advanced publication at Leukemia (paper-1) and Nature Biotechnology (paper-2). I’ll review these papers and comment on how these data reflect back on the public perceptions of the AAAS presentations, and how they fit into other recent findings in the field, in the next post.

stay tuned

comments are welcome

Why we should all buy immunotherapy company stocks

Someone recently commented that Bristol Myer Squibb’s (BMS) stock price was getting dangerously high on a forward price/earnings basis (this was before the Dec/Jan selloff). Technically this may have been true, but the science says this company (and Merck; MRK) and a few others are here for the long run, with therapeutics that will fundamentally transform cancer care. So I ask: who cares about PE? My view is to buy the key companies and watch them thrive for a decade or more, regardless of competition and possible price controls.

How can I be so bold (or stupid) you might ask.

I recently gave a talk for a local non-profit, the Boston Pharmaceutical & Bioscience Society (PBSS), which provides mentorship for students and job seekers in the biopharmaceutical industry. I decided to use cancer indications to advance the story of immunotherapy, and as I did I realized that we are sitting right at the bottom of a boon time for cancer treatment – for patients and biopharma of course, and ultimately for the payers and insurers. Why? because cured patients will be relatively cheap to maintain.


A magic word, much thrown about. In some hematologic malignancies and solid tumor indications (breast for example), cures can be anticipated if you catch the right patient at the right time. This has been the exception however, while the rule has been – to be blunt -let’s try to slow down you’re time to death.

So what is the evidence that immunotherapy will produce these extraordinary cures? We can start with advanced, metastatic melanoma. 10 years ago this diagnosis was a death sentence, with more than half of patients dead in a year, nearly all dead in 5 years. What’s changed? We can walk though some clinical trial data to sort this out.

here is where we start sometime pre-2010:

Figure 1

Screen Shot 2016-01-18 at 5.40.18 PM

The key points being that chemo was a lousy solution, with 2/3rds of patients dead the first year, and that the available immunotherapy (high-dose IL-2) was not much better, except for a very few “super-responders” who experienced long term survival. This was a hint that immune-based therapy could work – the key was to get better response with less toxicity.

Most of those following immunotherapy know the “long-tail” story – that when a cohort of patients is treated with an anti-CTLA4 mAb or an anti-PD-1 mAb – a subset (10% or more depending on the indication) do very well and have durable responses and long term survival, creating a long-tail on progression free survival (PFS) and median overall survival (OS) curves, which you can see at the bottom of this figure showing response to ipilimumab (anti-CTLA4; BMS):

Figure 2

Screen Shot 2016-01-18 at 6.24.08 PM

The OS curve on the right shows the long tail of survival out to nearly 5 years, and recent meta-analyses have highlighted patients living 10 years or longer, which is, as far as we can tell, a cure. We’ll note here the dosing regiment used for nearly all of these patients: treated once a month for 4 months. Period. We can conclude from this that when immune therapy works it does so by resetting the immune system response to the cancer – in some cases, permanently. Ipilimumab has a challenging toxicity profile, but in the face of such an improvement in response rate and in survival at 1 year – which jumped 15% over the chemo example in figure 1 – patients and their physicians will try to manage such toxicity.

If we look at results from anti-PD-1 treatment, the picture brightens again, as seen in this data from nivolumab clinical trials:

Figure 3

Screen Shot 2016-01-18 at 6.43.37 PM

Note that a few things have changed here. First, chemo alone (dacarbazine) is doing better here than the data we saw in figure 1: 42% are alive at year 1 although nearly all of them have progressive disease. But the objective response rate (ORR) for treatment with anti-PD-1 (nivolumab) is a whopping 40%, and nearly 2/3rds of patients are alive at 1 year. This is an amazing improvement over chemotherapy.

Now to fast-forward a bit. Ipilimumab and nivolumab are antibody therapeutics developed by BMS. Pembrolizumab is an anti-PD-1 antibody therapeutic developed by Merck (MRK). MRK cleverly designed a advanced metastatic melanoma trial in patients who had tried (and relapsed after) ipilimumab treatment. The results were outstanding:

Figure 4

Screen Shot 2016-01-18 at 6.53.02 PM

Two doses were tested here, but to simplify let’s just focus on the range of survival at one year: 58-72%. And look at chemo, right where we expect it at 30% alive at 1 year. Twice as many pembrolizumab-treated patients are alive at 12 months in this trial.

Rhe obvious thing to do next was to combine antagonists of CTLA4 and PD-1, which BMS got right to work doing. In the resulting trials the ORR jumped to 60%, although toxicity jumped as well, as seen here:

Figure 5

Screen Shot 2016-01-18 at 6.59.59 PM

Grade 3-4 AEs are very nasty and not to be taken lightly. However with further dose modification of ipilimumab, and careful treatment, these adverse events (AEs) can be managed, especially in the context of an otherwise lethal cancer condition. Nonetheless, many patients experienceing AEs dropped out of the trial, i.e. discontinued treatment, and now things get really interesting:

Figure 6

Screen Shot 2016-01-18 at 7.03.26 PM

So, 67% of patients comtinued to respond after treatment discontinuation! Now, consider that these patients only got four doses of ipilimumab – but they are kept on nivolumab for 2 years. Do they have to be on anti-PD-1 (nivolumab) for 2 years? Hell no, and this data proves it. So in a cost-constrained healthcare environment one might very well consider shorter courses of treatment with the potential to go back on to the drug (in this case, nivolumab) as needed. This is of course an abstraction at this point, but I’m certain we will see such trials performed in due course.

So where are we now:

Figure 7

Screen Shot 2016-01-18 at 7.09.19 PM

This approval signals that a fundamental change in the treatment of advance metastatic melanoma occurred in about 7 years, from standard-of-care chemotherapy to immunotherapy. It would be unfair to ignore the contribution of other therapeutic modalities, notably of BRAF and MEK inhibitors, but those do not produce cures, while immunotherapy is clearly able to do so in a small percent of patients (that percentage to be determined).

So just to recap, from where we started to where we ended up:

Figure 8

Screen Shot 2016-01-18 at 7.13.59 PM

The change in numbers from the dacarbazine line to the ipi.nivo combo line is staggering. The larger point in that the immunotherapy companies – BMS, MRK, AZN, Roche and others – will prosecute many other indications in exactly the same manner, with CTLA4, PD-1, PD-L1 and other, novel, therapeutics in a dizzying number of combinations. The value of such a class of medicines cannot be calculated in any meaningful way – each company is paying for hundreds of immunotherapy clinical trials – but it will take only a fraction of these trials to be successful to expand immunotherapy across the universe of oncology indications. The value of that success will be enormous – for patients, companies, shareholders and even, as mentioned, insurers and payers as we move patients from chronic and intensive treatment, to potentially curative treatments.

How does one guess the biopharma winners? Personally I like the bucket approach, as the ultimate winners will more than pay for the losers (and there will be some losers). As we ease into 2016 with certainly bearish sentiment about the biotech sector, I’m buying bluechip immunotherapy companies, and I won’t even look at them again for a year or two.

stay tuned.

Enumeral update – guest post by Cokey Nguyen, VP, R&D

Paul’s introduction:  Enumeral has been sending ’round some interesting updates to several of their programs and I asked for some more detail. Below is a quick primer sent along by Cokey Nguyen. More detail is available in Enumeral’s recent 8K filings, including one that dropped this morning. Also the company will present this and other work at the AACR Tumor Microenvironment Meeting in January ( - see below).

New data from Enumeral, by Cokey Nguyen

PD-1 biology in human lung cancer is an active area of research, as these cancers have shown PD-1 blockade responsiveness in clinical trials.  Enumeral has a drug discovery effort aimed at generating novel anti-PD-1 antibodies to develop into potential therapeutic candidates.  Using a proprietary antibody discovery platform, two classes of PD-1 antagonist antibodies were discovered:  the canonical anti-PD-1 antibody which blocks PD-L1/PD-1 interactions and a second class of antibody which is non-competitive with PD-L1 binding to PD-1.  These antibodies were validated first in a pre-clinical model of NSCLC using NSG mice with a humanized immune system and a patient derived NSCLC xenograft (huNSG/PDX) (Figure 1).  Here either class of antibody demonstrated activity on par with pembrolizumab, confirming that PD-1 blockade can slow tumor growth.

Figure 1

Figure 1

In order to confirm these pre-clinical findings, Enumeral began proof of concept studies with NSCLC samples.  The first question was if resident TILs, as found in tumors, could be reinvigorated (Paukken and Wherry, 2015) or if PD-1 blockade is mainly a phenomenon that affects lymph node-specific T cells that have yet to traffic to the tumor.  In these studies, Enumeral found PD-1 blockade can, in fact, increase effector T cell function, as readout by IFNg, IL-12, TNFa and IL-6.  In addition, in a NSCLC sample that showed PD-1hi/TIM-3lo expression, PD-1 blockade strongly upregulated TIM-3 expression (~5% to ~30%, see Figure 2).

Figure 2

Screen Shot 2015-12-01 at 6.07.24 AM

In these NSCLC-based studies, it was also found that an anti-PD-1 antibody (C8) which does not bind to PD-1 in the same manner as nivolumab or pembrolizumab (PD-L1 binding site) displays differentiated biology:  increased IFNg production and significantly higher levels of IL-12 in these bulk (dissociated) tumor cultures (Figure 3).  As IL-12 is thought to be a myeloid derived cytokine, this mechanism of action is not yet well understood, but has been now observed in multiple NSCLC samples as well as in MLR assays.

Figure 3

Screen Shot 2015-12-01 at 6.08.35 AM

In these NSCLC studies, while a subset of patient samples demonstrates PD-1 blockade responsiveness, the co-expression of TIM-3 on NSCLC TILs suggests this is a validated path forward to increase the response rate in lung cancer.  As with the PD-1 program, armed with a substantial portfolio of diverse anti-TIM-3 binders, Enumeral is actively testing single and dual checkpoint blockade on primary human lung cancer samples.

Look for the companies 2 posters at AACR/TME in January

Screen Shot 2015-12-01 at 6.11.08 AM