Description

Annexon is a development-stage biotech that represents a timely and attractive option-like bet with positive expected value ahead of the immanent topline readout of a Phase 3 trial of its lead compound in Guillain-Barre syndrome (GBS).

While recognizing that biotech ideas will be outside the usual comfort zone of many VIC investors, I think the dramatic drawdown in the XBI from its 2001 peaks has created some interesting option-like opportunities, where adding small positions in a few carefully selected names having clearly identifiable near-term readouts with a reasonable probability of success can add significant asymmetric upside uncorrelated with the market. In many cases, it's now possible to make investments on much more favorable terms than experienced specialist biotech VCs have done in recent years, while also being much closer to an identifiable hard catalyst associated with an impending data readout.

There is of course always going to be significant inherent uncertainty in the results of ongoing clinical trials, and this will make most value investors uncomfortable regardless of how far valuations have come down, resulting in many small-cap biotech names being effectively orphaned in the current environment. By taking an event-driven perspective and calculating evidence-based estimates of expected value based on historical probabilities of success specific to the therapeutic area and stage of development, I'd argue that it's possible to effectively identify some attractive mispriced options, and this is part of what I've been focusing on recently.

Anyway, biotech value rant over. I'll briefly present my assessment of ANNX so you can decide whether or not taking a small position here is your cup of tea.

Pipeline overview - scientific rationale for targeting complement overactivation in autoimmune disease

ANNX is a development stage biotech developing a portfolio of antibody-based treatments targeting the complement cascade, part of the innate immune system which when overactivated can contribute to a variety of serious autoimmune and degenerative disorders in humans.

The biology of the complement system is complex, but a high-level description is that it is part of the innate immune system consisting of a series of complement proteins, which can progressively become activated in a cascade-like fashion when an initial component (the C1 complex) is fixed by antibodies following exposure to a pathogen. Activation of complement components promotes inflammation, opsonization (marking cells to be attacked by white blood cells), and direct lysis (destruction of cells by insertion of a membrane attack complex directly into membranes). Importantly, ANNX’s two lead programs target a component of the C1 complex that is activated at the very start of the classical complement pathway (upstream of previously approved inhibitors targeting downstream components such as C3 and C5). While benefiting from the fact that the precedent of successfully approved complement inhibitors provides strong clinical validation for targeting this pathway, targeting C1 could have the unique advantage of interrupting the cascade from its very start, perhaps providing more comprehensive efficacy in some autoimmune indications.

During early response to a bacterial infection, complement activation is helpful in rapidly mounting an aggressive immune response against the initial pathogen. However, following recovery, some patients eventually develop autoantibodies (antibodies targeting proteins of the original pathogen, which unfortunately also react against proteins present on normal cells in the body). These autoantibodies bind to normal cells and initiate the classical complement cascade, with the first step being fixation of complement complex C1 (composed of C1q, C1r, and C1s subunits). In many autoimmune and degenerative disorders including GBS and geographic atrophy, overactivation of this cascade directly leads to significant autoimmune damage to a variety of cells and tissue types.

GBS is a rare but serious disorder where an autoimmune reaction (often following recovery from a previous infection) results in initial fixation and overactivation of complement complex C1 and progressive overactivation of the classical complement cascade, resulting in rapidly progressive nerve damage that in severe cases can lead to respiratory paralysis requiring ICU admission and mechanical ventilation. ANNX's lead compound, ANX005, is a monoclonal antibody (mAb) designed to potently inhibit C1q activity following a single dose, thereby preventing the cascade of downstream complement components involved in inflammation and nerve damage in GBS.

Geographic atrophy is the advanced stage of dry age-related macular degeneration (AMD), an increasingly common age-associated cause of progressive vision loss. Overactivation of the complement cascade is also strongly implicated in dry AMD pathology, and there is strong clinical validation supporting targeting components of the complement cascade thanks to successful development and approval (and >$5B subsequent valuations) of treatments targeting C3 and C5 as further described below. ANNX’s ophthalmology compound ANX007 is an antigen-binding fragment (Fab) directed against C1q, delivered directly into the eye by intravitreal injection.

To boil this down: considering the well-established involvement of complement overactivation downstream of C1 in the pathogenesis of GBS and geographic atrophy, directly targeting complex C1 with an anti-C1q antibody is a highly rational approach to interrupting overactivation of the classical complement cascade at its source, and could be considered to have a reasonable probability of success in the clinic.

As I'll describe below, this has been borne out by encouraging results in animal models and a Phase 1b randomized clinical trial in GBS, further supporting a reasonable probability of success for the impending readout of ANNX's Phase 3 study in GBS. Using historical data on average Phase 3 success rates to incorporate uncertainty in outcomes (and conservatively assigning zero value to Annexon's other pipeline programs), expected value of an investment at these levels appears attractive and timely heading into the impending Phase 3 GBS readout.

Targeting C1q overactivation in GBS - preclinical and clinical evidence

Consistent with the scientific understanding above that C1 overactivation directly drives nerve damage in GBS, in mouse models of GBS, it has been demonstrated that anti-C1q monoclonal antibody treatment attenuated complement cascade activation and deposition, and reduced immune cell recruitment and axonal injury, along with improvement in respiratory function. Importantly, this has seen encouraging clinical validation in Annexon's previously completed Phase 1b randomized controlled trial in GBS patients as further described below.

The company's GBS R&D day presentation from March of this year presents a detailed overview of the disease background, unmet medical need, and the scientific rationale for targeting autoantibody nerve damage via C1q inhibition with ANX005, as well as the existing clinical data in GBS which I'll summarize below.

ANNX previously completed a randomized, placebo-controlled Phase 1b proof-of-concept study in 50 adults with GBS (12 randomized to placebo and 38 to ANX005, with a conventional Phase 1 ascending dose design used to progressively explore higher doses of ANX005).

The safety profile was broadly positive, with no deaths and no treatment-related serious adverse events (SAEs); the highest dose of 75mg/kg was determined to be well tolerated.

Importantly, the study delivered clear and objective evidence of target engagement and relevant pharmacodynamic effects, with C1q activity rapidly inhibited by a single 75mg/kg dose. Comparison of an assay of overall complement component activity (CH50) following an infusion of IVIG (a nonspecific mixture of immune globulins which is part of current standards of care) versus a single infusion of ANX005 showed that ANX005 rapidly decreased complement activation to a much greater extent.

Pooled data from the 3 highest dose cohorts demonstrated meaningful improvements in muscle strength on the Medical Research Council Sum-Score (MRC-SS) and improvements in function on the GBS Disability Scale (GBS-DS).

Neurofilament light chain (NfL), an objective biomarker that is well established to reflect damage to neurons, was significantly decreased following ANX005 treatment.

Overall, I would say these data are encouraging in terms of the consistency of numeric benefits of ANX005 versus placebo seen across multiple endpoints, and are consistent with the scientific rationale that C1 overactivation is directly associated with nerve damage in GBS. It should also be said these data are far from conclusive considering the low sample size, and come from a relatively small 50-patient single site study outside the US (in Bangladesh of all places – see background below). Although I’m relatively optimistic given the scientific rationale for C1 inhibition and the consistency of improvement across endpoints in Phase 1b, I'm assigning the ongoing Phase 3 a 46% base rate probability of success (defined as successful BLA approval) based on the historical success rates for this disease area and stage of development.

ANNX has fully enrolled its Phase 3 pivotal trial of ANX005 monotherapy in GBS, with enrollment completed in 2H2023, and is on track for topline data in the current quarter (literally any week now). The study enrolled a much larger sample size of 241 GBS patients, randomized 1:1:1 to placebo, 30 mg/kg ANX005, and 75 mg/kg ANX005, with a primary outcome of GBS-DS at week 8 (supported by Ph1b data showing 30% of patients with >3= points of improvement on GBS-DS at week 8, versus 0% on placebo). The nearly 5X increment in sample size will leave this study robustly powered to demonstrate statistical significance versus placebo if Ph3 effect sizes are anywhere near those reported in Ph1b.

A core reason that the Ph1b and Ph3 placebo-controlled trials have taken place in less-developed nations outside the US and Europe (Bangladesh and the Philippines) is that in wealthy nations that can afford it, the current approach to GBS involves multiple rounds of IVIG infusions or plasmapheresis (removal and replacement of a patient's blood plasma using specialized apheresis machinery). The efficacy of these approaches may be suboptimal, with many GBS patients dealing with severe and long-lasting weakness, but until Annexon's studies, their presence had essentially led to no placebo-controlled studies being conducted in GBS for many years (apart from a small Japanese study of eclizumab that appears to have failed).

Having secured Fast Track and Orphan Drug designations from the FDA, the agency agreed that their pivotal study evaluating ANX005 as monotherapy can be sufficient for BLA approval, and that the BLA application can include analysis of real-world evidence data from IGOS, a global, prospective, observational, multicenter cohort study that has enrolled 2,000 patients who were followed for one to three years. This will effectively allow the sponsor to gather very clear evidence of significant benefit in a placebo-controlled study outside the US, while also being able to compare their outcomes data against real-world-evidence on IVIG and plasmapheresis. This is interesting because not only do IVIG and plasmapheresis require multiple rounds of treatment, many patients continue to deteriorate while on current treatments, with ~40% requiring ICU admission, ~25% requiring mechanical ventilation, and ~10% becoming permanently disabled. The improvements in clinical function with ANX005 vs placebo seen in Phase 1b would unquestionably be meaningful if borne out in Phase 3, and potentially largely supplant IVIG and pheresis if comparative outcomes are favorable.

Directional valuation of pipeline programs

To take a straightforward and conservative path to valuing this as an option, I'll focus entirely on the company's two late-stage programs (the topline readout of the ongoing Phase 3 study of ANX005 in GBS, and two Phase 3 pivotal studies being initiated this year in geographic atrophy or advanced dry AMD) and assign zero value to all other pipeline programs.

ANNX has four other pipeline programs in addition to these (ANX1502 in cold agglutinin disease and other autoimmune indications, ANX009 in lupus nephritis, and potential development of ANX005 in additional indications including Huntington disease and ALS). In reality, some of these programs could potentially have substantial additional value – there are a wide range of autoimmune and degenerative disorders that could potentially be addressable by targeting complement activation, and the market could quickly become more excited about the potential across the pipeline if seeing near-term late-stage success from ANX005 or ANX007.

ANX005 in GBS

To arrive at a conservative expected value estimate for the impending readout of ANX005 in GBS, I'll use the historical data on rates of success by study phase and disease area generated by the Biotechnology Innovation Organization (BIO), based on analysis of a decade of longitudinal data on 9704 drug development programs. Based on these data, the base rate likelihood of approval for programs in neurology indications currently in Phase 3 stands at 46%.

The company's claims data analysis and market research conducted by ClearView Health indicate incidence of hospitalized and treated GBS at 7000 patients annually in the US and 15000 hospitalized and treated patients annually in Europe.

The potential price point for a new standalone disease-modifying treatment for GBS is interesting, with significant upside but no clear recent analog to point to. Since this is a severe and potentially fatal orphan disorder with significant unmet needs, and current standards of care often require protracted ICU admissions, mechanical ventilation, rehabilitation, and long-term disability (which in addition to being very undesirable for patients are extremely costly over and above the costs of IVIG and pheresis), it's not out of the question that this one-time treatment could be priced very aggressively in the US (perhaps much higher than the $50k benchmark I'll use) and still be a net favorable proposition for payors. For a directional back-of-the-envelope estimate of market sizing, assuming market share of 0.3 and pricing of approximately $50k in the US and half that in Europe would equate to annual revenues of $217MM, or a valuation of approximately $650MM at ~3X sales. Handicapping this at a 46% empirical probability of success gives an expected value contribution of +$300MM.

ANX007 in GA

Geographic atrophy is an advanced form of dry age-related macular degeneration (AMD), a leading cause of blindness with prevalence that increases rapidly with age, which is becoming increasingly common as the median age of the global population continues to advance. It is also preclinically and clinically well established that overactivation of the classical complement cascade is implicated in GA pathogenesis, with recent clinical validation supported by successful development (and rich >$5B valuations) of other complement component inhibitors as described below.

As a relevant comp, Iveric Bio developed an inhibitor of complement C5 (another component of the complement cascade downstream of C1) called avacincaptad pegol (Izervay), which was approved for GA in August 2023, and was acquired by Astellas Pharma for nearly $6 billion.

Apellis (APLS) has a inhibitor of complement protein C3 called pegcetagoplan (Syfovre), which is approved for GA and an ultra-rare autoimmune disorder called paroxysmal nocturnal hemoglobinuria (PNH). Relatively soon after its March 2023 launch date, Syfovre recorded net US revenues of $275MM in FY2023, and this was considered to be a relatively underwhelming initial launch by Apellis compared to the large market potential in GA. APLS' current market cap is $5.1 billion, largely supported by the long-term market potential for Syfovre.

There could clearly be significant value for a new GA treatment targeting the complement cascade upstream of C3 and C5 – but Annexon's Ph2 data with ANX007 in GA have admittedly been mixed. In the Phase 2 ARCHER study, monthly injections were associated with a modest numeric reduction in rate of measured GA lesion growth versus placebo (6.2% on monthly treatment) that did not reach statistical significance and was underwhelming compared to other complement agents.

Importantly though, neither Apellis nor Iveric have been able to establish that their drugs’ effects on geographic atrophy lesion size translate to preservation of vision, which clearly is more directly meaningful to the lives of patients.

In ARCHER, a secondary efficacy analysis showed patients who received a placebo were more likely to lose the ability to read at least 15 letters on a standardized eye chart after one year than patients randomized to ANX007. Treated patients had a 59% lower risk of a 15-letter-loss overall, with a greater risk reduction in those who were treated more frequently.

This could be dismissed as the company cherry-picking the data to present comparisons favorable to them – and this certainly happens in biotech. However, there is a significant degree of overall clinical validation behind targeting components of the complement cascade in GA, and it's interesting to note that C1 is upstream of both C3 and C5 (the Apellis and Iveric targets). In interrupting the classical complement pathway at the source, it may be possible that targeting C1 could more completely abrogate consequences of complement activation in a way that could not be fully accomplished by the other agents, potentially leading to a meaningful advantage in vision preservation that doesn't necessarily track with lesion size.

In addition, there have been some late-emerging safety issues reported with C5 inhibition that have not been encountered with C1 inhibition with ANX007. Following approval, Apellis reported at least ten cases of occlusive retinal vasculitis in treated patients, a rare but potentially blinding type of inflammation of retinal blood vessels that can cause blindness.

In 2H2024, ANNX is initiating a Phase 3 placebo-controlled trial designed to confirm vision results seen in the Phase 2 ARCHER trial, and a second Phase 3 head-to-head study against Syfovre. Overall, the jury is certainly still out on Ph3 prospects in GA, but the vast and attractive market represents significant expected value for even a low probability of Ph3 success. Ophthalmology indications also have a lower than average probability of success, with a benchmark rate of 46.7% from Phase 3 to approval. To be conservative, I'll slash this base rate by more than half to 20% considering the mixed Ph2 data, and use the lower of the two valuations of Iveric and APLS (call it $5B), for a measly +$1 billion in expected value at a 20% chance of success.

Summary and conclusion

ANNX excecuted a $125MM secondary offering at $2.88 in December of last year, which leaves them with cash and equivalents of $264.9 million as of the latest quarter ending March 31, with a generous anticipated runway into mid-2026 and no near-term fundraising needs.

The PPS has been creeping up as I've been writing this, but I think it's a reasonably attractive positive expected value bet which could create attractive uncorrelated IRR heading into the near-term readout in GBS. At 92MM shares + 34MM warrants = 126MM, the fully diluted market cap stands at $615MM (a fraction of ridiculous charades like SAVA), $265MM which is cash. If you're inclined to believe my expected value SOTP using 46% and 20% probabilities of success for their two lead programs only, $1.3B expected value is roughly 2X the diluted market cap giving no credit for cash.

Catalyst

Impending Phase 3 readout in GBS