Oxford BioMedica — CTL019 approval likely in pALL, ad-com positive

OXB provides a critical component (lentiviral vector) of Novartis’s T- cell therapy CTL019 (tisagenlecleucel). CTL019 is known as a chimeric antigen receptor T-cell (CAR-T), a new generation of therapies that aim to engineer a patient’s immune system to fight their cancer. This process is complex and requires both significant expertise and manufacturing capabilities. Novartis has submitted a biologics license application (BLA) to the FDA for the treatment of relapsed/refractory paediatric ALL (r/r pALL) patients (EU submission expected shortly). Additionally, Novartis expects to file a BLA in DLBCL in both the US and EU shortly. A potential FDA approval in pALL is expected in the autumn (PDUFA action date: 3 October), with an EU approval and approval in DLBCL (in both regions) potentially by year end. Key to this approval process was the recent FDA advisory committee meeting (ad-com), which invited independent experts to pass their guidance of CTL019 in pALL. The 10 person committee voted unanimously in recommending approval based primarily on data from the Phase II ELIANA trial (study B2020). Assuming a positive PFUDA meeting, CTL019 would be the first ever CAR-T approved.

OXB has developed and can manufacture engineered HIV (part of the lentivirus family, modified to be safe) to genetically modify a patient’s T-cells to target cancer. OXB provides this lentiviral vector to Novartis, which utilises it in its manufacturing process to treat patients. While the technology is in its infancy, the approach has had clinical success in blood cancers like acute lymphoblastic leukaemia (ALL) and diffuse large B-cell lymphoma (DLBCL).

One of the core focuses of the ad-com panel was to better understand the long-term safety of a genetically modified cell. Two points raised in the ad-com were insertional mutagenesis and the creation of replication competent lentiviral vector. The lentiviral vector technology that OXB utilise is based on a third-generation vector system where accessory genes are removed and the genome is split among plasmids. This makes the vector replicant incompetent, preventing the virus used to modify the T-cells from spreading to healthy cells in the body. In 13+ vector lots and 72+ patient-specific CTL019 lots that have been tested there has been no positive reading for replicant-positive lentivirus, validating OXB’s technology as safe.

Insertional mutagenesis, where insertion of the gene may cause a change in activity of host genes through mutation, disruption or changes in activity is a long-term cause for concern as it could lead to new cancers. OXB has designed its vectors in a way to minimise this possibility; they lack enhancer sequences and as such are less likely to activate nearby genes. To date there has been no vector-associated leukaemia with CTL019 or any other vector-modified products. Based on data to date, we believe that OXB’s viral vector technology in CTL019 is robust and, while a long-term study (Novartis plans a 15-year registry study) will monitor for these side effects, we believe they are not of significant concern, a point highlighted by the unanimous vote in favour of CTL019 by the advisory committee.

The efficacy of CTL019 (tisagenlecleucel) was not a core focus of the ad-com, as its efficacy is clearly evident. In the ELIANA trial, the overall response rate (set at three months) was 82.5% (n=52/63) (95% CI: 70.9, 91.0), with 40 of those patients experiencing a complete remission (CR) with complete blood count recovery and the remaining 12 experiencing complete remission without blood count recovery (CRi). All complete responses were minimal residual disease (MRD) negative in the bone marrow, indicating complete eradication of the cancer cells. Median duration of response (DOR) has not yet been reached (median follow-up of 4.8 months). The estimated relapse-free rate among responders at six months was 75.4% (95% CI: 57.2, 86.7). At the data cut-off date (23 November 2016), 11 of the 52 subjects relapsed after treatment with CTL019 and a further two relapsed after treatment with both CTL019 and another cancer therapy. 22 remain in remission at data cut-off.

When looking at approved FDA therapies for r/r ALL we see how significant the efficacy of CTL019 is. Most recently Blincyto (blinatumomab), a bispecific CD19-directed, CD3 T-cell engager (BiTE) was approved in 2014. An open-label study (MT103-205) tested Blincyto in paediatric patients who had relapsed or were refractory to a range of treatments. It demonstrated a complete response (CR) of 17.1% (n=12/70, 95% CI: 9.2, 28.0). An additional 11 patients (15.7%) had complete remission with partial blood count recovery (CRi). In total 32.9% of patients had a CR or CRi (n=23/70, 95% CI: 22.1, 45.1). Median duration of response (time before relapse) was six months (range: 0.5-16.4) in the combined group. Blincyto prescribing information includes a black box warning for cytokine release syndrome and neurological toxicities.

While trial comparisons should be undertaken with caution, the substantial difference in complete responders (either with or without complete blood count recovery) between Blincyto and CTL019 is evident. Going forward, duration of response will be a key metric for CTL019, which will likely inform both pricing and treatment paradigms.

While efficacy has been remarkable, severe adverse events (SAEs), predominantly cytokine release syndrome and neurotoxicity, have been a serious cause for concern in many patients. In the ELIANA trial 69% of patients had a SAE in the first eight weeks after CTL019 infusion. While dose, disease burden and CAR-T construct (antibody fragment or co-stimulatory domain used) is theorised by some to predict the severity of side effects (and also the efficacy), Novartis clinically has no evidence that could point to any one component as being a key factor. Unlike some other CD19 CAR-Ts, notably the now cancelled JCAR015 from Juno Therapeutics, no cases of cerebral oedema have occurred with CTL019. It is thought that this may be due to the selection of the 4-1BB co-stimulatory domain by Novartis, while competitors Kite Pharmaceuticals and Juno utilise a CD28 domain. However, cases of cerebral oedema have occurred in patients treated with CAR-Ts containing the 4-1BB domain, indicating that a combination of factors is likely to be at play, including cancer type, starting T-cell populations and viral vector utilised.

Cytokine release syndrome (CRS), the other common adverse event associated with CAR-Ts, is a systemic inflammatory response that can be severe and sometimes fatal. However, it is significantly more manageable than neurotoxicity. One approach which is gaining traction is the utilisation of tocilizumab (an antibody that targets inflammatory cytokine IL-6) to dampen down CRS. Importantly, Novartis noted that it could be utilised prophylactically to minimise CRS without any noticeable effect on efficacy.

For both Novartis and OXB, manufacturing will be a key limiting factor in delivering CTL019. Never before has a commercial medical product (if approved by the FDA and/or EMA) required the removal, modification and re-administration of a patient’s immune cells. This is a labour-intensive technique that requires significant capabilities at all stages. Novartis is aiming for a 22-day manufacturing process (multiple patients have died while waiting for injection of CTL019, further highlighting both the need of the patient population and ideally a shorter manufacturing time) from extracting a patient’s cells to re-administrating them. To enable this, Novartis needs OXB’s lentiviral vectors, which take several months to manufacture. Once made, these are delivered to Novartis, which stores them to use later to modify multiple patient cell batches. Three vector batches were utilised to treat all 88 patients in the ELIANA Phase II trial. These vectors were produced via its Process A procedure and would not be able to produce the required vector volumes to meet any potential future commercial demand beyond paediatric ALL. As such, OXB has invested in and is gradually moving its manufacturing over to its Process B procedure. This involves the use of bioreactors, which the company states will improve productivity at least tenfold. This will put OXB at the forefront of lentiviral manufacturing, as few if any companies globally possess the ability to manufacture on that scale.

An internal review in April 2016 led to the prioritisation of three internally developed pipeline assets: OXB-102 (Parkinson’s disease – Phase I/II), OXB-202 (corneal graft rejection – Phase I/II) and OXB-302 (cancer, multiple types – preclinical), which could deliver the best potential economic returns. The goal for each is to be advanced to at least proof-of-concept in humans via out-licensing or through the formation of externally funded SPVs. OXB will look to obtain value through upfront payments, equity stakes or developmental milestones and from royalties on sales.

Product candidates that fall outside the priority programme (OXB-201 for wet age-related macular degeneration and OXB-301 for multiple cancers) will only be progressed once suitable opportunities, like partnering, enable reduced investment from OXB.

The group will continue to invest in earlier-stage gene and cell therapy concepts (eg in ocular, central nervous system and respiratory indications) with the aim of identifying new candidates for further development via out-licensing or spin-outs.

We note that SAR422459 (licensed to Sanofi) for Stargardt disease has progressed into Phase II development.

Our valuation has increased to 8.15p/share (£251.6m) vs 6.75p/share previously (£208.5m). This increase has been driven mainly by changes in our CTL019 assumptions and rolling forward our model. Our assumptions for OXB’s pipeline remain unchanged. We note Novartis’ ability to drive penetration of its treatment (mainly in the larger indication, DLBCL), its pricing of CTL019 and the royalty rate OXB receives all have a significant effect on our valuation. We believe significant upside for OXB could come from other partnerships as additional companies are likely to be attracted by OXB’s expertise and its ability to commercially produce the lentiviral vector. However, at this time the ability to predict the scope and timing of any deals is difficult and not included in our valuation.

For CTL019 we have updated our EU population numbers conservatively to include the top five countries (by GDP) as we feel these are more likely to be initially targeted. We have increased the age range in ALL to include young adults. Based on the impressive efficacy data in both the ELIANA (r/r pALL) and JULIET trial (r/r DLBCL), we have increased our peak penetration rates (35% in DLBCL and 50% in ALL). We expect lower penetration rates in DLBCL compared to ALL, as we anticipate increased competition (Kite Pharmaceuticals likely first to market). In pALL, competitors are at least 18 months behind and we expect Novartis to gain significant market share. We have increased the expected price of CTL019 to £300k based on market sentiment and have altered our royalty rate, which now at peak is 2%. We anticipate launch of CTL019 for pALL in autumn this year (PDUFA action date: 3 October 2017) and in DLBCL at year end. We note that DLBCL represents the majority of revenue in both manufacturing and royalties due to the significantly larger patient population. Our core CTL019 assumptions can be seen in Exhibit 2. A key unknown factor which materially affects our valuation remains the royalty rate OXB receives from Novartis. We note that a 1% peak royalty rate lowers our overall valuation to £182.7m (5.92p/share), while a 3% peak royalty rate increases our valuation to £320.6m (10.38p/share). Please see our previous notes for a more in-depth look into our overall valuation metrics.

We have updated our forecasts, which now reflect the creation of the new debt facility with Oaktree Capital Management, changes in our CTL019 assumptions and the $10m upfront payment in 2017 from the new supply deal with Novartis. We have adjusted our expectations for CTL019 (outlined in more detail in the valuation section), which has prompted us to revise our revenue expectations upwards in 2017, 2018 and 2019. Our 2018 and 2019 numbers are driven significantly by royalties and manufacturing of CTL019. In addition, we expect revenue from collaborations with Orchard Therapeutics and Green Cross LabCell to contribute meaningfully over this time period. The recently announced supply deal with Novartis consists of a $10m upfront in 2017 and an additional $90m committed over the next three years. Additional undisclosed CAR-T products are included in the deal and the supply agreement can be extended to five years with the agreement of both parties. We maintain our forecasts for R&D and bioprocessing costs, which we believe will see a significant reduction from 2017. This is a reflection of the near-term strategy to out-license or spin out the product portfolio (£21.5m in 2017, £18.5m in 2018 and £17.0m in 2019).

OXB recently announced the creation of a new $55m debt facility with Oaktree Capital Management. This facility has been used to redeem the existing debt facility with Oberland Capital Healthcare which, as last reported in the FY16 results, represented a net fair value of £34.4m. We estimate that c $10m is available to be drawn down when taking into account the original sum. Notably improved deal terms (potential cost of loan 11.5% compared to 15% for the previous Oberland facility) will aid OXB as it looks to become profitable in the near term. The loan is to be paid back by 29 June 2020 and has a 9% interest rate plus US$ Libor (minimum 1%). Subject to achieving certain conditions, the interest rate could drop 0.25% in both the second and third years. Additionally, the company has issued 134m warrants to Oaktree, which can be exercised over the next 10 years at a nominal share price of 1p.