Grade 4 Liver Enzyme Elevation in Intellia’s Phase 3 ATTR Amyloidosis Trial

 Last night, Intellia filed a material event report (8-k) with the SEC.  In there, they revealed a case of very high, grade 4 liver enzyme elevations in a single subject treated with NTLA-2001 (nex-z) in the ongoing MAGNITUDE phase 3 ATTR-CM trial.  This case appears to be resolving without any hospitalization or medical intervention.

For context, Intellia has now dosed around 400 subjects with its particular LNP-mRNA formulation across the ATTR and HAE trials.  This is the first such case to be reported.  Earlier cases of liver enzyme elevations were rare.  There were 2 milder, but significant AST elevations in the first month following dosing among the 36 subjects in the phase I/II portion of ATTR-CM development of nex-z (Fontana etal, 2024).  These cases similarly resolved within days.

Liver safety is a key consideration in the development of systemically administered LNP-delivered nucleic acids.  Following delivery, the liver soaks up these LNPs containing non-natural lipids that could insert themselves into normal lipid biology.  It is therefore important that they degrade and get removed from the body- the sooner the better.

There has for example been a case of so called Hy’s Law with Alnylam’s LNP-formulated Patisiran (RNAi) in its phase 3 trial in ATTR-CM (APOLLO-B), meaning that bilirubin was concurrently elevated (excerpt from the APOLLO-B Briefing Docs):

  

Clearly, more context, including the temporal association with nex-z administration, any changes in bilirubin, and the general health and behavior of the subject involved, need to be eventually provided by Intellia for better judgement of the event (it is a blinded study).  It needs to be remembered, too, that nex-z (or most other CRISPR-LNPs) is administered only once which allows for close monitoring in clinical practice.  The fact that the trial is allowed to continue is a positive sign.

Update (5 June, 2025): Since the initial 8-k came out, the company had meetings with analysts during which it emerged that the liver enzyme elevations occured and waned in week 4-5 following administration.  This is inconsistent with acute LNP toxicity as I speculated.  For example, VERVE-101 triggered such an acute response with ALT peaking in the first week.  

This leaves a rare adaptive immune response to the CRISPR editing enzyme (Cas9) or a delayed response to the editing mechanism (double-strand breaks) as the two main other plausible mechanisms.  Pre-existing immunity to various Cas9 enzymes is quite common, so that might be a line of investigation and lead to future adaptations of use.  If hepatocytes that highly express Cas9 for a prolonged period of time got preferentially attacked by cytotoxic T-cells, there should be a decrease in editing levels following the immune attack.

There was also the disclosure that the subject had taken 3g paracetamol for 8 days prior to receiving therapy, a medicine known for occasional severe hepatotoxicity (yes, even common, over-the-counter medicines can cause grade 4 and higher liver enzyme elevations).  Keeping your liver happy around systemic LNP-RNA administration, for example by abstaining from alcohol should be good practice in any case.

1 1/2 weeks following the hepatotoxicity disclosure, no clinical halt has been placed on NTLA-2001 by any of the global regulatory bodies involved in the phase 3 trial, giving extra comfort around its overall safety profile.

RNAi Therapeutics Now a Commercial Reality

Last Friday, Alnylam received notification from the US Food and Drug Administration (FDA) that ONPATTRO (aka Patisiran) has been approved for the treatment of hATTR-related polyneuropathy.  This marks the culmination of an almost picture-book translation of a brand-new biotechnology into therapeutic reality.  20 years from worms to patients is nothing.

ONPATTRO development path

Much has been made in the press about the great uncertainty, reflected in supposedly unusually long timelines, of whether RNAi can be a therapeutic modality at all.  Of course, the challenges of delivering RNAi triggers in WoMan, staving off unwanted innate immune responses, and the question of how the transcriptional noise stemming from slightly modulating dozens of unrelated targets will affect safety were all daunting at the beginning.

But looking back, with the exception of a 2-3 year delay due to having to change from an insufficiently potent LNP chemistry to the MC3 lipid-based LNP underlying ONPATTRO and having to institute steroid pretreatment to minimize ‘infusion reactions’, all these challenges more or less dissolved in the development path of ONPATTRO.

A special shout-out here to Ian MacLachlan and his team at formerly Tekmira for solving the critical delivery challenge first.

A best-case scenario would therefore have seen an approval in 2015-6.  2018 is therefore not that bad at all.  Here a quick run-down of the milestones leading up to the approval:

1998: discovery of double-strand RNAs (dsRNAs) being the trigger for RNAi in worms

2001: finding that RNAi can be triggered in WoMan by short dsRNAs

2002: first demonstration of RNAi in mice

2005: first therapeutically relevant demonstration of RNAi in monkeys

2009: initiation of first clinical trial of first-generation LNP-RNA (SNALP-ApoB, aka PRO-040201; ALN-VSP02)

2012: start of ONPATTRO clinical development

2012: first solid clinical proof-of-concept for RNAi in WoMan from ONPATTRO phase I study

2012: start of ONPATTRO phase II study

2013: start of ONPATTRO phase III study (APOLLO)

2017: positive data read-out from APOLLO

2018: EU and US marketing approvals for ONPATTRO  

ONPATTRO is just the beginning

Since ONPATTRO had been conceived starting about a decade ago, there has been, of course, considerable progress in RNAi trigger design (safety, efficacy) and (conjugate) delivery technologies.  As a result, a slew of other RNAi drug candidates by Alnylam, including Givosiran for Acute Intermittent Porphyria (likely 2019 approval), Inclisiran for cardiovascular disease (likely 2020 approval), and Lumasiran for primary hyperoxaluria (likely 2020 approval), are about to be approved.  Especially with the adoption of off-target-minimizing chemical modifications, I am confident that the success rate of RNAi drug candidates for tissue types for which delivery is robust will only increase.

After the liver, for which the current crop of marketing candidates is the target organ, the CNS should be the next huge opportunity as Alnylam gears up to bring a first candidate into that organ over the next year.  In addition, Arrowhead Pharmaceuticals and Alnylam have also started to talk about opportunities in the lung and cancer, although here I am still hesitant about whether these are near-term clinical opportunities or something that will have to wait another 3 years or so.

RNAi shatters antisense competition

As the world is abuzz about the first full-blown marketing approval of an RNAi Therapeutic (note: formal marketing authorization in the EU is still pending), I would be remiss not to display a key graph illustrating the power of RNAi gene silencing versus competing approaches such as gene silencing by mass action RNaseH antisense (TEGSEDI by Akcea Therapeutics) and trying to keep bad proteins from acting out with small molecules (generic tetramer stabilizer diflusinal):  

If you were diagnosed with TTR amyloidosis and had an average life-expectancy of 2-5 years, which treatment would you like to be on?

Stock market reaction

When Alnylam opens for trading today, it will likely be down due to disappointment that the FDA label, unlike the anticipated case in the EU, will only specify TTR-related polyneuropathy for which the APOLLO study had been designed, and not cardiomyopathy with none of the exploratory cardiomyopathy data from APOLLO included.  While this is somewhat disappointing in light of biomarker-based accelerated approvals of Eteplirsen in DMD where target modulation and therapeutic hypothesis were enough to win the day, this should only be a minor hiccup in the overall history of RNAi Therapeutic.  In light of the recent Pfizer small molecule Tafamidis cardiomyopathy data, it is very likely that RNAi trials will show a benefit for cardiomyopathy symptoms as well in a dedicated study (à ALN-TTRsc02). 

Until then, it is possible that the cardiomyopathy data are being withheld until Alnylam and the FDA can come up with a way to include those in an amendment, or as part of an accelerated approval submission.  It is also likely that TTR patients will now more readily be referred from cardiologists to neurologist for an intense neuro check-up such that primary cardiomyopathy patients can access ONPATTRO (and get paid for it) at the slightest signs of polyneuropathy.

TEGSEDI worse than tetramer stabilizer according to newly released EMA document

When Ionis presented data from the phase III NEURO-TTR study last year in Paris, they clung to numbers close to ‘zero’ to make the point that its TTR-lowering antisense drug TEGSEDI (aka inotersen) was 'stabilizing' and ‘halting’ disease progression.  According to a newly released document by the European Medicines Agency (EMA), this, however, does not seem to be truthful: TEGSEDI only delayed the progression of polyneuropathy compared to placebo, but patients on placebo still got worse over the 15 month study period.   

This not only widens the apparent distance in therapeutic efficacy between TEGSEDI and RNAi competitor ONPATTRO (which improved on disease parameters), but even puts it apparently behind generic tetramer stabilizer diflusinal.  Diflusinal also happens to be much better tolerated than TEGSEDI which has been plagued by platelet and renal issues.

EMA document suggests numbers were inflated

According to the ‘Summary of Product Characteristics’ document issued by EMA following its approval of antisense drug TEGSEDI for the treatment of TTR-related polyneuropathy, mNIS+7 after 15 months increased by +11 points vs 25 points for placebo.  By contrast, Ionis Pharmaceuticals (which has now licensed the drug to subsidiary Akcea Therapeutics) claimed a mere +5 point progression.  Curiously the placebo values haven’t changed.

This compares to an increase of +9.2 for diflusinal over 24 months and -6 for ONPATTRO over 18 months.

Similarly, on another measure of disease progression, the Norfolk Quality of Life questionnaire increased by only +0.99 per the Paris presentation last year, but by +4.38 per the EMA document.  Once again, the placebo numbers remained essentially the same.

Finally, what had been heralded as a TTR knockdown close to that of ONPATTRO, a median 75-79% TTR reduction vs 82% for ONPATTRO, now looks much different when considering that mean knockdown was only 68-74%, possibly reflecting the poor tolerability profile of TEGSEDI and missed doses.

I am sure that Akcea and Ionis will have eloquent explanations for the discrepancies which just so happens to  conveniently and selectively favor their drug when analyzed by them.  These new numbers, however, are not just minor adjustments, but represent substantial changes to the TEGSEDI narrative.

It should be noted that it is likely that tetramer stabilizers and TTR-lowering agents will be taken together by many patients.  The relative efficacy and tolerability numbers, however, put TEGSEDI in a very weak position with regard to direct competitor ONPATTRO, also as it comes to reimbursement decisions. 

ONPATTRO heart aches

In the phase III APOLLO study, patients treated with ONPATTRO were numerically less likely to die compared to those on placebo (~50% reduction in death rate).  Following the Paris meeting, I came away with the impression that the deaths in the ONPATTRO arm were largely due to cardiac failure and infection.

According tothe New England Journal of Medicine publication on the study, this seems to be a misunderstanding as infection was a main cause of death in the placebo arm while all deaths in the ONPATTRO arm were cardiac.  As has been pointed out by others on Twitter (@ionisdisrupts and @artkrieg), this could raise questions in the minds of regulatory bodies whether to include TTR cardiomyopathy applications on the label despite of ONPATTRO improving on related secondary endpoints.

In fact, considering that the recent study design agreement with the FDA for follow-up drug ALN-TTRsc02 also focuses on polyneuropathy endpoints, it is all but official that the label for the upcoming approval of ONPATTRO will be targeted at the polyneuropathy population only and that separate trials will have to address the patients mainly suffering from cardiomyopathy symptoms.

Disclosure: short AKCA, long ALNY.