In part 1 our focus was primarily on the PD-1 and CTLA4 pathways, where the biology is well understood and the drug development advanced. See that post here. In part 2 we look at drug development for some newer immune checkpoint targets, and this will drive us a little deeper into the scientific rationale for some of the less known pathways.
I would argue that a good deal of the excitement around some recent deals (Novartis/CoStim and Agenus/4-Antibody) really is driven by the opportunity to get in early with novel targets. While the CoStim portfolio included PD-1 pathway related IP, I think the fact that this deal was so early stage suggests that novel LAG-3 and TIM-3 IP had a lot to do with driving interest. Similarly, emerging details from the BIOCIO conference indicate that Agenus (NASDAQ: AGEN), a somewhat obscure company, acquired novel LAG-3, TIM-3, OX40 and GITR antibodies as well as novel CTLA4 and PD-1 antibodies in its’ 4-Antibody acquisition. It would seem that this company, nominally a cancer vaccine company, is taking a huge leap forward by acquiring assets that could be combined with cancer vaccines. Barron’s labeled this a “genius move”, and I agree. This should make Agenus itself an attractive acquisition candidate. The Smith On Stocks Blog has much more on this (http://bit.ly/1ljmEzx).
So I think it make sense to take on these targets one by one, do a quick update on the therapeutic rationale, and see who is leading the pack. Later we’ll fold this into a landscape analysis to try to understand where the large companies are heading.
We can start with a few targets that are represented by drugs in clinical development. Bristol-Myers Squib (NYSE: BMY), already loaded with anti-CTLA4 and anti-PD-1 programs, is moving their LAG-3 antibody ahead in both monotherapy and combination therapy trials. LAG-3 (lymphocyte activation gene, CD223) is a negative regulator of cell activation. It is expressed on various activated lymphoid cells, including T cells and NK cells that mediate tumor cell killing. The mechanism of action is the binding of LAG-3 to the MHC Class II complex expressed on antigen-presenting cells (B cells, monocytes, macrophages, dendritic cells, and other cell types). The high affinity binding event blocks cell proliferation and effector functions. LAG-3 is also an important mediator of the immune suppressive function of regulatory T cells. Of tremendous interest is the finding that LAG-3 is synergistic with other down-regulatory pathways, specifically PD-1 and TIM-3. As we will see this is driving much of the work on the design of combination therapy testing.
BMS-986016 is an anti-LAG-3 antibody from BMY currently in phase 1 testing in solid tumors and in B cell lymphomas. A very interesting study is NCT01968109: Safety Study of Anti-LAG-3 With and Without Anti-PD-1 in the Treatment of Solid Tumors. This is a Phase 1 dose escalation study of BMS-986016 alone or in combination of one of two defined doses of nivolumab (anti-PD-1). The primary endpoint is safety (AEs, SAEs, fatalities, lab abnormalities). There is also a cardiovascular risk assessment (QTc interval) among the secondary endpoints. Otherwise the secondary endpoints cover PK and exposure, immunogenicity, and RECIST defined tumor responses.
The point is that this is an instructive example of rational combination immunotherapy being investigated at Phase 1.
Other LAG-3 antibodies of potential use in oncology include Immutep’s IMP701, an antagonist antibody. IMP701 ought not to be confused with their depleting anti-LAG-3 antibody IMP731 (partnered with GSK for treatment of autoimmune disease) nor with their activating LAG-3-Fc fusion protein IMP321 (and how this thing works I have no idea). We have already mentioned that CoStim and 4-Antibody had LAG-3 programs and IP, but these would be preclinical. Somewhat better known for its’ anti-CD70 mAb (see below), arGEN-X also lists TIM-3, LAG-3 and VISTA antibodies in its’ preclinical portfolio. No doubt there are other early stage programs, they just are not readily visible yet. I wager that we will see many more of these popping up in the poster aisles at AACR and ASCO this year.
Two proteins related to PD-L1, B7-H3 and B7-H4, are also T cell inhibitory ligands. Both proteins are expressed on tumor cells and expression of B7-H3 or B7-H4 correlates with poor outcome for some tumor types. Both B7-H3 and B7-H4 are normally expressed on myeloid lineage cells including monocytes and dendritic cells. Preclinical tumor model data have supported efficacy with blocking antibodies to these ligands in vivo. The mechanism of action of these ligands is not well understood, as the receptors are not known, or at least cannot be confirmed across different laboratories
The Macrogenics (NASDAQ: MGNX) antibody to B7-H3 has reached clinical development. The phase 1 study in patients with advanced carcinoma, melanoma, or glioblastoma that overexpresses B7-H3. The antibody, MGA271, is licensed to Servier; Macrogenics recently received a milestone payment indicating that the expansion part of the Phase 1 trial had been initiated. Five Prime recently disclosed novel antibodies to B7-H3 and B7-H4 along with TIM-3 and VISTA, as mentioned previously.
With TIM-3 we have a landscape that is a bit earlier than LAG-3 – the excitement about this pathway is driven by the preclinical tumor model data and the translational medicine data. Like LAG-3, TIM-3 has been identified as co-expressed with PD-1, in particular on tumor infiltrating lymphocytes. Genetic data (knockout, transgenic, etc) clearly indicate that TIM-3 is an important immunoregulatory pathway. This is true as well of CTLA4, PD-1 and LAG-3 – the number of “brakes” on the system is remarkable and hints at how dangerous the immune system can be when it is unregulated, as it is in autoimmune, inflammatory, allergic and similar diseases. One of the interesting observations about TIM-3 is that it is ectopically expressed by some tumors and also by dendritic cells associated with tumors, i.e. within the tumor microenvironment. Therefore by blocking TIM-3 in the tumor setting, multiple responses may contribute to efficacy. A confounding issue in the TIM-3 field is the identification of the relevant ligand for TIM-3, with a number of ligands having been proposed (galectin-9, phosphatidylserine (PS), HMGB1). The binding motif for PS is well defined, while binding to the other proposed ligands is less well understood. In particular, TIM-3 and galectin-9 activities seem distinct, at least as far as we can understand from the published genetic data.
The proteins mentioned so far (CTLA4, CD28, CD80, CD86, PD-1, PD-L1, PD-L2, B7-H3, B7-H4, LAG-3, TIM-3, VISTA and TIGIT) are all members of the immunoglobulin (Ig) superfamily of proteins. Two additional protein families of critical importance in regulating immune responses are the TNF and TNF receptor families. Again the leader in this field, clinically at least, is BMY. The antibody BMS-663513 (urelumab) is an agonist anti-4-1BB antibody that functions by stimulating T cell activation. 4-1BB (CD137) is best known for contributing a signaling moiety to the CAR-T constructs (a discussion for another day). BMS-663513 is now in phase 1/2 testing in lymphoma patients. The antibody had previously completed a phase 2 study in melanoma, but that program was put on clinical hold following dose dependent liver toxicity. The new studies utilize lower doses, as a very low dose appears to be efficacious. An important differentiating feature of anti-4-1BB is the apparent ability to eradicate established tumors, at least in some patients. With this is mind it is encouraging to look forward to combination treatment studies. Pfizer also has an anti-4-1BB antibody in development, PF-05082566. This antibody is in a very interesting phase 1 clinical trial in solid tumors and B cell lymphomas, the latter patients being treated with and without rituximab co-therapy.
4-1BB biology is well understood, and agonist stimulation of this receptor induces CD8+ T cell activation, interferon gamma secretion, secretion of cytolytic compounds and recruitment of helper T cells. Of interest, 4-1BB is only expressed on T cells that have been activated through the T cell receptor and CD28, and so is specifically expressed on those T cells that would potentially have anti-tumor activity.
CD27 expression is also induced upon T cell activation, and the critical role of this receptor in immune responses is shown by patients who lack function CD27, as these patients are grossly immunosuppressed. The role of CD27 is subtly different from 4-1BB in that this receptor seems critical to activated T cell survival. Celldex (NASDAQ: CLDX) has developed an agonist anti-CD27 antibody, CDX-1127. In pre-clinical models, CDX-1127 had anti-tumor effects due to enhanced T cell activation. In addition various cancers, particularly B and T cell lymphomas, can express CD27 at high levels and the antibody may be able to such tumor cells directly and activate immune cell killing. Early data is promising, with no obvious toxicities.
The ligand for CD27 is CD70. Paradoxically (and stretching the limits of our understanding of these systems) CD70 is expressed at very high levels on a variety of tumor types, including solid tumors and hematopoietic cancers. Therefore, antibodies targeting CD70 to effect tumor cell killing have been developed. The most advanced of these are antibody drug conjugates, e.g. SGN-75 (SGEN) and MDX-1203 (BMY); there are other coming e.g. from Ambrx. In January. arGEN-X started a Phase 1b expansion study with ARGX-110, a novel cytotoxic anti-CD70 antibody. There are undoubtedly other antibodies in development.
A critical pathway found on cells that interact with T cells (dendritic cells, macrophages, B cells) is the CD40 pathway. Although early work is in the monotherapy setting, it is reasonable to speculate that agonists to CD40 would complement other approaches, such as cancer vaccines and modulators of T cell responses. Dacetuzumab, developed by Seattle Genetics (SGEN) was discontinued in phase 2b. The reason was unclear but appeared to involve both toxicity and futility analysis. Toxicities included cytokine release syndrome (common) and thrombosis (< 5% of patients), some liver toxicity and cytopenias. Most of these toxicities could be controlled with prophylactic agents. CP-870,893 (Pfizer) has completed Phase 1 clinical trials in melanoma, pancreatic cancer and other solid tumors. The current development in the US of CP-870,893 seems limited to trials being sponsored by U Penn’s Abramson Cancer Center. Of note, one of these trials is in combination with the anti-CTLA4 mAb, tremelimumab. The antagonist anti-CD40 antibody lucatumumab (NVS) competed a phase 1 trial in refractory follicular lymphoma in May of 2012. Here the hypothesis was that the bound antibody would activate cytotoxic killing of CD40+ tumor cells. This Phase 1 trial was in combination with chemotherapy (bendamustine).
At this point I would characterize the development of CD40 modulators in oncology as stalled, and awaiting a better understanding of the best antibody activity (and associated isotype) to use, the appropriate dose, and the most relevant tumor types.
Two final pathways to mention in this section are the OX40 and GITR pathways, the subject of headlines when Agenus bought out 4-Antibody. Several clinical stage therapeutics have been developed for these targets.
OX-40 (CD134) is another T cell survival pathway, activated downstream of CD28, and essential for the induction of anti-apoptotic proteins that keep activated T cells alive and functional. It may also be required for the establishment of the memory T cell pool. Stimulation of OX40 by the OX40-L or by agonist anti-OX40 antibodies enhances T cell responses.
AZN/Medimmune has developed a murine anti-OX40 agonist antibody designed to stimulate the immune system and block tumor suppression of the immune response. AZN’s OX40 collaborations are complex. AZN/Medimmune has partnered with AgonOx, a tech transfer spinoff from the Providence Cancer Center in Portland, OR. There are several clinical trials of anti-OX40 therapy underway at the Providence Cancer Center. AZN/Medimmune has also partnered with the Cancer Research Institute and Ludwig Institute for Cancer Research specifically to undertake clinical trials evaluating immunotherapy combinations including the MedImmune antibodies to OX40, and PD-L1 (MED14736), together with other agents within the CRI/Ludwig portfolio and the Cancer Vaccine Collaborative network of clinical immunologists and oncologists. There was one clinical trial co-sponsored by AgonOx and the Ludwig Institute, to study anti-OX40 in combination with ipilimumab. However, this trial has been suspended. According to AZN/Medimmune, the partnership trials are designed to complement their in-house clinical development effort.
GlaxoSmithKline (NYSE: GSK) gained rights to an OX40 antibody preclinical program from the MD Anderson Cancer Center’s Institute for Applied Cancer Sciences, as part of a deal focused on immune checkpoint antibodies that can trigger immune responses against cancer.
GITR was the other target grabbing headlines in light of the Agenus/4-Antibody deal. GITR is yet another cell surface receptor that is involved in amplifying T cell responses. It’s mechanism of action is distinct, in that GITR inhibits the suppressive activity of T-regulatory cells, thereby releasing effector T cells from active suppression. Secondarily GITR signaling is a pro-survival pathway for activated T cells.
GITR, Inc., is a biotech company spun out when Tolerx went under. The company is developing TRX518, an anti-GITR agonist antibody designed to enhance the immune response to cancer cells. A Phase 1 clinical trial in melanoma and solid tumors is currently recruiting after being released from clinical hold.
A few thoughts about these newer pathways. One is that some of them are very potent indeed (4-1BB, CD27) and we will have to watch carefully for toxicity issues. A second is that we can begin to outline rational combinations based on the biology of the pathways. For example, the CTLA4 and PD-1 antagonists may pair well with treatments that induce tumor cell death, thereby releasing novel tumor antigens that the newly stimulated immune system can then recognized. Some of the downstream T cell or antigen-presenting cell activators (CD40, OX40 as examples) may be better suited for use with cancer vaccine therapies.
There are two more classes of immune checkpoint modulators to consider. One consists of the IDO inhibitors. The second consists of the innate immune response modulators (TLRs, KIR, NKG2A). There are very exciting companies working in these areas, and these will be the subject of the next update.
as always please leave a comment or email me at firstname.lastname@example.org and follow @PDRennert