Yesterday, Uniqure reported on the 24-month results of their Phase 1/2 clinical trial with AMT-130, a viral gene therapy trial for HTT lowering for the treatment of HD. See the links to the press release and investor call slides.
I wanted to review some aspects of the claims made by Uniqure, and put them in the context of the extensive preclinical data with AMT-130 and with other late-stage HD trials for HTT lowering. This article is meant to be educational and informative, and in no way represents any criticism of Uniqure or the data reported. I maintain my enthusiasm for this program and I will anxiously await the results of the Phase 3 trial in planning.
My concerns relate to the expectations of HD families, who traditionally over interpret press releases and early data presentations. Therefore, it is important for me to educate the public and highlight aspects that require further clarification. Unfortunately, press releases and investor meetings tend to color the way scientific and clinical data is presented in an over enthusiastic manner.
AMT-130 is an investigational therapeutic that consists of a microRNA targeting the HTT mRNA. The drug affects both the wild type and mutant HTT mRNAs, leading to their degradation. The drug also leads to the degradation of the smaller HTT mRNA termed exon-1a, which encodes a pure exon-1 protein product, known to be highly neurotoxic in mice. This product, first described by Gillian Bates, arises only in the presence of expanded CAG repeats in mHTT, due to a read-through mechanism through intron-1 of the human HTT gene. Although the functional significance of this presumed toxic product in humans is still under investigation and subject to significant controversy in the field, the fact that AMT-130 targets it for degradation makes this RNAi approach different from all other therapies in clinical development.
AMT-130 is delivered via an AAV5 viral vector, which infects brain cells after direct administration into the caudate and putamen after a neurosurgical procedure. The extensive preclinical data published by the Uniqure investigators showed profound suppression of HTT mRNA and protein expression, exceeding 75% of its normal levels in the caudate and putamen. These studies also showed a remarkable effect in regions of the brain connected by fibers to the caudate and putamen – namely deep layer cortical neurons (layers III and V), thalamic projection neurons, and the output nuclei of the caudate/putamen (the globus pallidus external and internal segments). Even though the virus does not infect the entire caudate and putamen, a significant region of both structures is being targeted efficiently. Therefore, this therapy targets deep basal ganglia structure regions, and contrasts with very regions most targeted by the ASO therapies being developed by Ionis/Roche and Wave.
It is important to keep this in mind when interpreting results – good or bad- from these clinical programs. Any potential deleterious effects of lowering HTT too much or of the therapeutic administered with ASOs cannot be used to predict or interpret effects of a basal ganglia delivered therapy, so be cautious about interpreting the disappointing results observed in the tominersen ASO Phase III study as evidence of a HTT lowering driven toxicity.
The effects of lowering HTT via AMT-130 in the basal ganglia represents a novel and very important step in defining whether HTT lowering is tolerated in these brain structures. The Roche study simply doe not inform of basal ganglia lowering of HTT. Similarly, the AMT-130 study presumably only targets a subset of all the cells of the cortex, whereas the ASOs target most cells in the cortex. These studies are not comparable.
AMT-130 Clinical trial design and outcome measures
The clinical trials for AMT-130 are being conducted in Europe and North America, with two doses of AMT-130 being administered (low and high doses, based on the injected volumen, eg the amount of viral particles administered, the high dose delivering 10x more genome copies of the AAV5 virus than the low dose). For a description of the trial, the primary endpoint (safety), secondary endpoint (durability of AMT130 in brain) and exploratory endpoints (biomarkers and clinical tests), and the patient enrollment criteria, please see the link at clinicaltrials.gov
The data reported by Uniqure shows the therapy is well tolerated up to 24 months, which was the primary endpoint. As I have explained before, this is a significant milestone for the field of HD. There was always concern about the tolerability of a direct injection into the degenerating basal gsanglia nuclei. So this is good news for this program and for other gene therapy approaches in the pipeline.
However, only 9 individuals with HD have completed the high dose 24-month endpoints. The number of people who have completed this stage of the trial is small. Uniqure has reported data on the 20 patients receiving the high dose (this includes patients who have ‘crossed over’ to the high dose), and in 13 receiving the low dose group. There are some adverse events which include headaches but these appear related to the surgical procedure, and do not differ significantly from the 10 controls in the study.
Importantly, additional exploratory outcome measures which are part of the study to monitor safety or efficacy of AMT-130, include: MRI, CSF measures of HTT, mHTT, NFl (neurofilament light chain), YKL40 and GFAP (markers of astrogliosis), and a range of HD-relevant functional, motor and cognitive scales.
Disease stabilization claims – biomarker results
The press release & investor slide deck from Uniqure only reports some of the relevant metrics, but not all, of the trial’s outcome measures. Importantly, they report on levels of neurofilament light, but not of astrogliosis markers for neuroinflammation. Similarly, they state that mHTT levels do not significantly change in the study – although the levels are not shown. Based on the preclinical transgenic HD pig model work – which expressed an expanded human HTT transgene; published – we knew that AMT-130 delivered via direct injection did not substantially alter HTT levels in CSF as compared with the level of suppression detected in the basal ganglia (13% in CSF vs 75% in caudate & putamen tissues). Even though this was expected, it would have been nice to see the mHTT levels in this critical clinical study.
The levels of neurofilament light (a marker of axonal damage) initially increase due to the surgical procedure (which was expected), and the levels decrease over the 24 months of the study to baseline. This is also good news, given the contrast with the tominersen trial, where Roche reported a transient elevation after 4 months of dosing the ASO, and this correlated with the other adverse findings in the study that led to its discontinuation.
Uniqure makes the statement that the levels of NFl decrease in the CSF of the 19 treated individuals in the high and low AMT-130 cohorts as compared to the control natural history group. The data are shown as percent change over baseline of the same subjects at the beginning of the trial. They compare this to pre-existing data obtained from the Enroll-HD database (using a cohort of 19 individuals from the ‘Enroll-HD control natural history group’). It is unclear whether these measurements were done by a different lab, and whether they used identical methods; this might be important, as significant variability can occur when running these assays on different days, platforms, and laboratories.
This is worrisome, and in general, the use of natural history data as a control for the outcome measures poses some potential problems. For example, I do not know how variable the baseline data is for the ‘control’ Enroll-HD cohort vs the cohort in the trial. Uniqure should show these data to ensure baseline levels are similar. It is also interesting to see that they pooled the data from the high (n=9 individuals) and low (n=12 individuals) AMT-130 cohorts, probably to reach statistical significance, but they do not show whether these groups can indeed be pooled. They also do not show the Nfl baseline levels of the control individuals enrolled in the clinical trial. The changes in NFl in individuals enrolled in clinical trials, with repeated measures (as compared to studies such as Enroll-HD, where no therapeutic intervention is included) is unknown. We simply do not know if placebo effects manifest in changes in CSF fluid biomarkers such as HTT or NFl.
Therefore, I am cautious about interpreting this result as evidence of neuroprotection. It is also surprising that Uniqure does not report on structural MRI changes, which is by far the best characterized measure of disease progression and neurodegeneration in HD. We will have to await for these results.
Disease stabilization claims – clinical outcome measures
Uniqure has shown the results of the clinical outcome measures, including scores in the composite UHDRS (which includes scores in motor outcomes (total motor score or TMS), the total functional capacity (TFC), and the cognitive tests such as the simple digit modality test (SDMT) and the Stroop Word Reading Test (SWRT). The cUHDRS score correlates well with the current gold standard to measure progression, structural MRI (see reference here for the validation of the cUHDRS scale).
Below you can see the data presented by Uniqure, showing a significant stabilization of the cUHDRS score in the high dose group, particularly noticeable after 12m of treatment. When analyzing the individual test scores, it becomes apparent that the most significant effects are seen in the cognitive scales, SWRT and SDMT. This is far less clear when analyzing the changes in the TFC and the TMS. To my knowledge, this is the first instance of a therapy affecting the cognitive scores in HD, which if true and replicated with a larger cohort in the upcoming phase 3 trial will be transformative for patients.
In the cUHDRS, a change in score of 1.2 is deemed clinically meaningful. For the TFC, a score change of 1.0 is deemed clinically meaningful. Note that in the 24-month trial, the change in the cUHDRS is around -07-0.8, and for the TFC score it is -0.2-0.3 -; that is, not clinically meaningful. We will have to see what patients fare during the follow-up period (up to 5 years is planned).
One final note of caution relates again to the use of the ‘control natural history group’ .
When reviewing the tominersen Phase 3 trial paper published in the NEJM and the PTC518 investor presentation from PTC therapeutics, the following was reported:
1. In the tominersen trial, the placebo control group shows a decline in cUHDRS of approximatety -1.1 points in the 69 week study (see panel A below). Compare this to the change in cUHDRS after 104 weeks = 24 months in the Uniqure AMT-130 trial, of -0.8-0.9 points. In the PTC-518 trial, the placebo control group showed a decline of -0.91 points in this trial with a duration of only 12 months (52 weeks). Therefore, the decline reported in the Enroll-HD control cohort used in the Uniqure AMT-130 trial appears much slower than in the other two trials.
2. Similar observations are made when monitoring the scores using the TFC. In the tominersen trial, a -1.2 change is reported (see panel B below from the NEJM paper). In the Pivot-HD PTC study, a change of – 0.80 points is reported. Compare this to the Uniqure natural history control group of a change of -0.7-0.8, which again suggests a slower decline in the control group compared to these other interventional control groups.
3. The same observation holds for the TMS score changes in the Uniqure AMT-130 vs the PIVOT-HD PTC518 trial.
One wonders if a placebo effect is not driving these changes in functional outcomes in individual controls who participate in clinical trials vs in observational studies like Enroll-HD. Therefore I would strongly suggest caution and await longer term monitoring in the Phase 1/2 trial and the upcoming Phase 3 trial. I hope Uniqure will share the rest of the results from the trial in due course, particularly the MRI data.
2 Responses
Nacho, thank you for taking your time to write this extensive review of the latest uniQure results, and how they were presented. One suggestion I have, though, is that you add a summary (or list of bullet points) of your key observations about the uniQure data and their natural history data, so that HD families reading your article can have a clearer idea of your points.
Best regards, Bill K.
Thank you for sharing your insights, Nacho.