UniQure now guides for initial AMT-060 clinical data in early January 2016, the first look at the company’s gene therapy for hemophilia B, arguably their most important pipeline asset in the near-term. I felt now would be an appropriate time to summarize my expectations for the initial clinical data. Critical to understanding the asset, however, is a look at a discrepancy in non-human primate (NHP) data from uniQure when comparing to previously published data from the St. Jude/UCL group. And, I’ll elaborate on why St. Jude/UCL’s scAAV2/8.LP1.hFIX.co and AMT-060 are two distinct biologic products, therefore, extrapolating results from the older St. Jude/UCL clinical program to uniQure’s AMT-060 trial is a fraught analysis. Specifically, there is a difference in magnitude of hFIX expression in rhesus macaques receiving AMT-060 versus scAAV2/8.LP1-hFIXco, though there could be a rationale – and overlooked explanation.
As many know, the key clinical proof-of-concept for AMT-060 comes from a small AAV8 (scAAV2/8.LP1.hFIXco) gene therapy trial initiated by St. Jude/UCL in 2010. However, there are important differences between AMT-060 and St. Jude/UCL’s scAAV2/8.LP1.hFIXco, which seem to have gone overlooked by the investing community:
St. Jude/UCL: 3’ deleted terminal resolution site (trs) ‘self-complementary’ LP1.hFIXco
uniQure: Baculovirus expression system
St. Jude/UCL: Mammalian HEK293 triple transfection
Those who have followed uniQure may be surprised to learn that the AMT-060 transgene cassette is actually different from that used by St. Jude/UCL. In past presentations, uniQure has emphasized the use of the same exact (expression) transgene as St. Jude/UCL did in 2010. In reality, it is only the expression cassette (hFIXco) that is the same. The key difference is the sequence of the terminal resolution site (trs) in the inverted terminal repeats (ITRs), which facilitate the transgene folding into the ‘double-stranded’ (self-complementary) form. I will discuss this in more detail later on, but first I want to assemble the previously published rhesus macaque AAV gene therapy data that supports the theory that AAV5- and AAV8-based gene therapies facilitate similar magnitude transgene expression in rhesus macaque livers.
In order to make the rest of this post easier to understand, it is appropriate to summarize the structure:
Studies supporting the similar transgene expression facilitated by AAV5- and AAV8-based gene therapies in rhesus macaques
UPenn’s AAV2/5 and AAV2/8 rhCG rhesus macaque data
St. Jude/UCL’s hFIX scAAV2/5.LP1.hFIXco and scAAV2/8.LP1.hFIXco rhesus macaque data
Comparing St. Jude/UCL’s scAAV2/8.LP1.hFIXco and uniQure’s AMT-060 gene therapy products and rhesus macaque data
My thoughts going into the initial AMT-060 clinical data release
UPenn’s Rhesus Macaque AAV Data
In 2006, Gao et al. published a study discussing their work comparing AAV2, AAV5, AAV7, and AAV8 transgene expression in rhesus and cynomolgus macaque livers . This is one of the earliest studies supporting the similar liver transgene expression in NHPs facilitated by AAV5- and AAV8-based gene therapies. For simplicity, I will focus mainly on the rhesus macaque data from this publication.
Rhesus macaques were administered 1E13 [vg/kg] of either AAV2/5.TBGrhCG or AAV2/8.TBGrhCG vectors (n=2 per dose cohort) containing a liver specific transgene cassette (TBGrhCG). The TBGrhCG transgene cassette consists of the liver specific human thyroid-hormone-binding globulin (TBG) promoter and the gene for the β-chain of rhesus chorionic gonadotropin (rhCG). As shown below, the rhCG transgene expression kinetics appear to be similar in magnitude in rhesus macaques receiving either AAV2/5.TBGrhCG (left) or AAV2/8.TBGrhCG (right) :
This similarity in transgene expression is further supported by the transgene expression kinetics in cynomolgus monkeys administered 1E13 [vg/kg] AAV2/5.TBGrhEPO and AAV2/8.TBGrhEPO (shown below) :
The key take away is that these data from Gao et al. support the theory that AAV5- and AAV8-based gene therapies facilitate similar magnitude transgene expression in NHPs (rhesus and cynomolgus macaques). This conclusion is further supported by data from Nathwani et al. at St. Jude/UCL that I will discuss below.
St. Jude’s/UCL Rhesus Macaque AAV hFIX Data
First some background: The St. Jude’s/UCL group has published extensive NHP data exploring the use of AAV vectors to deliver the human Factor IX gene (hFIX) to treat hemophilia B. Here, I will focus on two key publications by Nathwani et al. that compare scAAV2/5.LP1.hFIXco and scAAV2/8.LP1.hFIXco in rhesus macaques [2,3].
In their 2007 Blood publication, Nathwani et al. treated seven rhesus macaques with 1E12 [vg/kg] of scAAV2/8.LP1.hFIXco or scAAV2/5.LP1.hFIXco (depending on AAV8 neutralizing antibody titers). For simplicity, I have compiled the data into a table below (left) :
As shown above (left), the steady state hFIX expression (prior to the occurrence of anti-hFIX antibodies or sacrifice) is similar in rhesus macaques receiving scAAV2/8.LP1.hFIXco or scAAV2/5.LP1.hFIXco (note: hFIX values taken as stated in the publication). Nathwani et al. followed this publication up with another published in Molecular Therapy in 2011 where the group explored the administration scAAV2/8.LP1.hFIXco in a larger cohort of rhesus macaques. For simplicity, I have compiled the data into a second table (above, right). There are a few things that need to be defined in these tables:
“Avg.”: Average expression from rhesus macaques in the specified dose cohort calculated from values given in the publication
“Calculated”: Average expression from rhesus macaques in the specified dose cohort calculated from my interpretation of each chimp’s hFIX expression profile
Corrected steady state (S.S) hFIX: corrected hFIX values compensating for the St. Jude/UCL assay overestimating hFIX by eight-fold (see below)
It is important to note the different assays utilized to assess vector genome titers in the 2007 Blood publication and the 2011 Molecular Therapy publication. In the Blood publication, the group used a quantitative slot-blot analysis. In the Molecular Therapy publication, the group switched to quantitative PCR (qPCR). Based on the hFIX expression seen in both groups of chimps, the group believes the ‘qPCR corrected dose’ for chimps receiving AAV2/8 or AAV2/5 in the Blood publication is ~2E11 [vg/kg] [3,5]. The difference in vector titer assays is one of the reasons why it is very difficult to compare the hFIX expression profiles between publications. Despite this, there are some key takeaway points that can be established:
AAV2/5.LP1.hFIXco facilitates similar magnitude hFIX expression in NHPs as AAV2/8.LP1.hFIXco
Close to a linear dose-response effect
Small sample size
These data support the data produced by Gao et al. demonstrating the similar magnitude of transgene expression facilitated by AAV5- and AAV8-based gene therapies in rhesus macaque livers
UniQure’s In-House AMT-060 Hemophilia B NHP data
In preclinical studies, uniQure administered 5E12 [vg/kg] AMT-060 to three rhesus macaques. UniQure presented these data in a poster at the 2014 American Society of Hematology (ASH) Annual Meeting. When I started comparing these AMT-060 rhesus data to the Nathwani rhesus data, there was a clear difference in hFIX expression between the two. I discussed this difference with uniQure management, and they informed me that they believe – based on in-house data – that the hFIX assay St. Jude/UCL used in the past over-estimated hFIX levels by around eight-fold (Note: I have not personally seen this supportive data). From my understanding, uniQure is implying that the difference in hFIX expression in rhesus macaques receiving AMT-060 and scAAV2/5.LP1.hFIXco (or scAAV2/8.LP1.hFIXco) can be attributed to the assays used in FIX antigen measurement, and potentially, AAV vector titer.
With this in mind, I again compared uniQure’s AMT-060 rhesus data to Nathwani’s rhesus data. As exemplified above, this is not a simple task due to the studies’ inherent heterogeneities, such as the difference in assays and the complex hFIX expression kinetics. Therefore, the simplest way to compare these data is to look at a side-by-side view of the hFIX expression profiles of rhesus macaques who received similar doses (after correcting for the difference in hFIX assays) [3,4].
Even after correcting for the difference in hFIX assays, there still appears to be a difference in both peak magnitude and steady state hFIX expression (prior to day 60). The rational explanation for this difference is the differences between AMT-060 and the Nathwani AAV5 & AAV8 biologics. This is where the difference in transgene cassette mentioned earlier becomes quite relevant: the difference in techniques to facilitate self-complementary folding by the transgene could be responsible for the data discrepancy. UniQure adopted the monomeric duplex technology due to existing IP on the single trs-deletion approach, and the inability of the single trs-deleted self-complementary approach to be used with their baculovirus manufacturing system. According to uniQure, the efficiency of monomeric duplex transgene packing into AAV capsids is ~75% (varies depending on the asset). This compares to the typical production yield of >90% dimeric (self-complementary) genomes from trs-deleted products .
This difference, in tandem with obviously skewed expression kinetics in rhesus monkeys, is worrisome. I would feel much more comfortable with uniQure’s AMT-060 program – with currently available data – if the biologic was EXACTLY the same as Nathwani’s scAAV2/5.LP1.hFIXco considering the similarities in outcomes with scAAV2/5.LP1.hFIXco and scAAV2/8.LP1.hFIXco.
It’s also worth mentioning that there are additional AAV5 data in mice and humans that leaves me with more questions than answers concerning AAV5’s efficacy in humans. For the sake of simplicity, I won’t get into the nitty gritty of these data [7-12].
My Expectations for the Initial AMT-060 Data Release (based on available pre-clinical data)
Based on available pre-clinical data and the day 56 corrected rhesus macaque steady-state expression, I believe AMT-060 may be 2-3 times less potent than St. Jude/UCL’s scAAV2/8.LP1.hFIXco. However, this comparison (and my prediction for the initial clinical data) is far from perfect and relies on a number of assumptions:
St. Jude/UCL’s hFIX assay overestimates hFIX by eight-fold
St. Jude/UCL’s and uniQure’s qPCR AAV vector titer assays are very similar
Assumes hFIX expression around day 56 is a suitable time-point to assess steady state hFIX expression in rhesus macaques
Assumes no effect on gene transfer (in these studies) of different levels of serotype specific neutralizing antibodies in rhesus macaques
Human translation: Assumes similar translation of AAV5 and AAV8 transgene expression kinetics from rhesus macaques to humans
Considering the AMT-060 low dose cohort (5E12 [vg/kg]) is dosed 2.5 times higher than St. Jude/UCL’s scAAV2/8.LP1.hFIXco clinical trial high dose (2E12 [vg/kg]), it’s not unreasonable to expect AMT-060 to facilitate ~5% steady state hFIX expression. However, the initial AMT-060 data release will probably not allow for a clear-cut interpretation of steady state hFIX expression because of the short-term follow-up. During this period, there will most likely be a recombinant hFIX “wash-out” period (similar to that seen in the St. Jude/UCL trial) which will make analyzing the amount of hFIX actually facilitated by AMT-060 difficult. Trying to assess steady state hFIX during the initial follow-up is further made difficult by the higher levels of hFIX expression typically seen in the first weeks after treatment. Therefore, I would not be surprised to see the treated AMT-060 patients show >5% FIX expression at <12 weeks post-gene transfer due to the combined hFIX expression from AMT-060 and recombinant hFIX.
In conclusion, I am certainly worried that AMT-060 is inferior to St. Jude/UCL’s scAAV2/5.LP1.hFIXco and scAAV2/8.LP1.hFIXco based off available pre-clinical data. Despite these differences in pre-clinical data and product characteristics, I’m cautiously optimistic about AMT-060 being safe and facilitating >5% steady state hFIX expression in humans.
Biotechnology stocks are speculative and not suitable for all investors. We invest in a limited number
of positions and employ strategies which may compound this risk. We are not market timers so generally we remain fully invested. Nothing on this website should be construed as a solicitation to buy or sell any security.
BTP is not registered with, or been approved by, any regulatory body. BTP is not available for public participation.