Imugene — In-licensing deal for CF33 oncolytic virus

Imugene has agreed to acquire a worldwide exclusive licence to the novel, highly potent, oncolytic poxvirus known as CF33, which is poised to enter Phase I clinical trials in 2020.

The CF33 technology was developed at the City of Hope Cancer Center in Los Angeles, California (City of Hope). Imugene has agreed to pay City of Hope licence fees comprising an upfront payment, annual maintenance fees, development and commercial milestone payments, a royalty on net sales and sublicensing fees.

As part of the overall transaction, Imugene has entered an agreement to acquire 100% of the unlisted company, Vaxinia. As Vaxinia’s major shareholder is Paul Hopper, Imugene’s executive chairman, the deal is a related party transaction that requires shareholder approval at an extraordinary general meeting (date to be set). The Vaxinia acquisition includes an upfront cash payment of A$462,500 and the issue of Imugene shares valued at A$1.6m. The Vaxinia shareholders will also be eligible for equity payments on the achievement of development milestones including granting of the Investigational New Drug (IND) by the US FDA, dosing the first patient in a Phase I clinical trial and the Phase I trial demonstrating safety.

The proposed clinical development strategy for CF33 (combination therapy with immune checkpoint inhibitors) is similar to that pursued by Viralytics, which led to its acquisition by Merck for A$502m, announced in February 2018. While the transaction gives Imugene a foothold in a space that has attracted a lot of pharma interest, it is also an area where there are a large number of competing products in development. Therefore, we suspect that Imugene will need to demonstrate superior efficacy in the clinic in order to attract a pharma partner.

Oncolytic viruses aim to infect and kill cancer cells. A large number of oncolytic virus candidates have been developed and tested over the past 30 years; while they have been shown to be safe, their efficacy to date has been modest.

In an attempt to improve on this modest efficacy, CF33 was developed through a process that aimed to maximise its potency against cancer cells while at the same time maintaining adequate safety. CF33 has been shown in preclinical studies to selectively target and destroy tumour cells without affecting healthy cells.

CF33 was created by researchers at the City of Hope by infecting a cell line with six different strains of the vaccinia virus (widely used in smallpox vaccines) plus raccoonpox, rabbitpox and cowpox viruses. The presence of the nine different virus strains in the cells at the same time allowed the genes from the different strains to be rearranged by homologous recombination to form novel chimeric daughter viruses that contained different combinations of genes from the parental virus strains. One hundred different chimeric viruses were purified and tested for potency against a panel of cancer cell lines. CF33 was the most potent of the 100 chimeric viruses tested.

Exhibit 1 shows that CF33 was more potent than each of the parental virus strains when compared across a panel of 54 solid tumour cell lines. Exhibit 2 shows the individual results from a subset of the panel, comprising four triple-negative breast cancer (TNBC) cell lines. Exhibit 2 shows that whereas none of the viruses was effective against the first two cell lines at the dose tested, CF33 (bright green bar) was more effective than any of the parental viruses against the BT549 and HS578T cell lines.

While the rationale behind the design of CF33 is the belief that the higher potency will give it an efficacy advantage in the clinic, the strategy has a number of other advantages. These include the fact that the new chimeric virus has generated fresh intellectual property (patent application filed) and that the lower number of virus particles needed to infect tumour cells will reduce the cost of manufacturing each dose of virus used to treat patients.

Like other vaccinia-based oncolytic viruses in development, CF33 has had its cancer selectivity enhanced through the inactivation of the thymidine kinase (TK) gene. The inactivation of TK increases tumour selectivity and improves safety, as it means that the virus can only replicate in rapidly dividing cells (such as tumour cells) which have a large pool of available nucleotides, including thymidine. Preclinical studies have shown that inactivating the TK gene does not affect the growth and cytotoxic potential of CF33 in cancer cells.

In addition, CF33 has been engineered to express the human sodium iodide symporter (hNIS) gene. hNIS is a cell surface protein that transports iodine into the thyroid gland for thyroid hormone synthesis. When expressed in tumour cells, hNIS transports radioactive iodine (131I) or rhenium (188Re) into the cell for tumour imaging or therapy.

CF33 has shown promising efficacy in a range of mouse models of cancer. Exhibits 3 to 5 show that CF33 was able to either shrink tumours or arrest the growth of both injected and non-injected tumours, when administered by either intratumoural or intravenous injection. The studies shown in Exhibits 3 to 5 were carried out in standard xenograft models of breast, pancreatic and colorectal cancers, in which human cancer cell lines were injected into immunocompromised ‘nude’ mice, which do not reject the transplanted human cells.

Exhibit 8, on page 7, shows that CF33 is also effective at low doses in an animal model of lung cancer.

Some of the preclinical studies were conducted using CF33 virus that has been modified to express the firefly luciferase gene. Firefly luciferase is an enzyme that produces a bioluminescent light signal in the presence of an appropriate substrate. The light signal allows real-time imaging of CF33-infected tissue in live animals. The modified virus is known as CF33-Fluc.

The abscopal effect refers to a phenomenon in the treatment of metastatic cancer whereby shrinkage of untreated tumours occurs following the localised treatment of distant tumours by methods such as radiation therapy or intratumoural injection. Instances of abscopal effects have been reported for other oncolytic viruses, including for Viralytics’ Cavatak.1

Published reports of preclinical studies conducted with CF33 data show a number of examples of the virus spreading to non-injected tumours and causing tumour shrinkage, in human cancer cells implanted in immunocompromised ‘nude’ mouse models. Exhibit 6 shows an example of this in a pancreatic cancer model.2

Athymic nude mice are used in these studies because they have impaired immune systems and do not reject the transplanted human cell lines. The impaired immune system also means that it is easier for virus to spread in nude mice than it is in animals with intact immune systems. Therefore, it is important to note that Imugene’s slide presentation on the CF33 deal states that it has preclinical data showing regression of tumours that did not have viral spread (ie presumably due to an adaptive immune response), and increased infiltration of both injected and uninjected tumours by CD8+ T cells and other immune cells.

Furthermore, cells infected by poxviruses such as vaccinia secrete two forms of virus, an extracellular enveloped virus (EEV) form, which is surrounded by a host-derived membrane and is largely invisible to the host’s immune system, which makes it easier to spread within the body.

These two factors may enable CF33 to generate abscopal responses in clinical studies in cancer patients.

The safety of vaccinia virus, which contributes the majority of the CF33 genome, has been confirmed over a long period of time via its use as a smallpox vaccine. However, given that CF33 also contains components of the cowpox, rabbitpox and raccoonpox viruses, less is known about the safety characteristics of the chimeric virus.

To date, more than 500 mice have been treated with derivatives of CF33. More than 50 mice have been treated with doses of up to 10 million infectious virus particles (pfu), both intravenously and intratumourally, without any sign of toxicity. Exhibit 7 shows that when immunocompetent mice were injected intratumourally with 10m pfu of two different CF33 derivatives, no virus was detected in any other organ seven days after the injection, whereas virus was detected in the tumour.

Furthermore, as the chart labelled (d) in Exhibit 6, above, shows, in the immunocompromised nude mice tumour model, CF33 virus counts were at least 10,000 fold higher in tumour tissue than other organs. This shows that CF33 is highly selective for replicating in tumour cells rather than in normal tissue.

The potency of CF33 was compared to that of Amgen’s marketed oncolytic virus, Imlygic (T-vec) and a vaccinia-based virus that is in clinical development, GLV-1h68 (GL-ONC1, Genelux) that could be potential competitors to CF33. The study in cancer cell lines showed that CF33 was two to three orders of magnitude more potent than the other two viruses. That is, a 100-fold to 1,000-fold lower dose of CF33 virus particles was required to kill tumour cells in culture, compared to the two other oncolytic viruses (Exhibit 8, left hand panel).

CF33 was also tested against T-vec and GL-ONC1 in a mouse mode of lung cancer. The two panels on the right in Exhibit 8 show that an intratumoural injection of CF33 results in greater tumour shrinkage than an injection of either T-vec or GL-ONC1, both in the injected tumour and in a non-injected tumour in the same mouse.

Imugene proposes to commence clinical studies of CF33-hNIS in 2020. The plan includes a pre IND (investigational new drug) meeting with the FDA planned for H219, and submission of an IND application in H120 to gain permission from the FDA to conduct the study in the US.

The proposed clinical development plan is to start a US-based Phase I study in lung, TNBC, melanoma, bladder and gastrointestinal cancers in 2020. The proposal is that 15 patients would be treated with intratumoural (IT) CF33 as a single agent at up to three dose levels. A further 15 patients would be treated with IT CF33 in combination with an immune checkpoint inhibitor (ICI, to be selected) at two dose levels, as shown in Exhibit 9. The estimated trial cost is ~A$5m.

While oncolytic viruses have consistently killed tumour cells in preclinical studies, few have been able to stimulate the immune system to the degree required to generate meaningful clinical benefit. Imugene aims to overcome this challenge by using CF33 in combination with an immune checkpoint inhibitor.

This strategy of combining oncolytic virotherapy with a checkpoint inhibitor is sometimes described as putting your foot on the accelerator (stimulating greater recognition of the tumour by the immune system) at the same time as releasing the brakes (overcoming the immunosuppressive tumour microenvironment by blockade of inhibitory checkpoints).

The intention is for the Phase I study to be followed by a Phase I/II study of CF33 plus an ICI in four tumour types selected from the Phase I cohorts. The plan would be to enrol four cohorts of ~30 patients each, for a total of 100–120 subjects. It is envisaged that the Phase I/II trial could potentially start in H221 (estimated cost A$15-18m).

The manufacture of GMP virus material for the Phase I clinical studies has commenced. An experienced team with prior oncolytic virus development experience at Viralytics will be joining Imugene to guide the development of CF33.

A patent covering the composition of matter and use of chimeric pox viruses was filed under the international Patent Cooperation Treaty in 2017. If granted, its estimated expiry date would be in late 2037.

Researchers have been exploring the utility of modified vaccinia viruses and other poxviruses as oncolytic virus therapies over the past 20 years.3 Exhibit 10 lists some examples of oncolytic viruses that have been investigated in preclinical studies. Each of these viruses has been modified using recombinant DNA technology to inactivate the thymidine kinase gene and express inserted transgenes to modify the virus properties, in order to improve safety and/or efficacy (in the case of CF33-hNIS, the inserted transgene is hNIS).

Three vaccinia-based oncolytic viruses are currently undergoing clinical studies (Pexa-Vec, GL-Onc1 and TG6002) and at least one other has previously been tested in Phase I (vvDD-CDSR), as shown in Exhibit 11.

The most advanced of these is Pexa-Vec, which is currently being studied in the Phase III PHOCUS trial in first-line liver cancer. Pexa-Vec is an oncolytic vaccinia virus engineered to express the granulocyte-macrophage colony stimulating factor (GM-CSF) gene to stimulate a systemic anti-tumour immune response.

PHOCUS is a 600-patient randomised, open-label study comparing overall survival (OS) in patients treated by Pexa-Vec followed by sorafenib (tyrosine kinase inhibitor) versus those treated with sorafenib alone. PHOCUS was initiated in January 2016 and was expected to undergo an interim analysis in H119. According to the record on clinicaltrials.gov (NCT02562755), the trial is expected to complete in December 2020.

In a previous Phase II dose-finding study of Pexa-Vec in liver cancer (n=30; 80% first-line), those receiving high-dose Pexa-Vec (intratumoural delivery) had a median overall survival of 14.1 months compared to 6.7 months for those on a low dose. However, the subsequent Phase IIb TRAVERSE study in second-line liver cancer was terminated early in 2013 as data from the first 80 events showed no evidence of overall survival benefit associated with Pexa-Vec.

Early stage clinical studies are underway of Pexa-Vec combined with the checkpoint inhibitors Nivolumab, Ipilimumab and REGN2810 in liver cancer, kidney cancer and solid tumours, respectively.

Pharma companies have shown considerable interest in oncolytic virus technologies, with a number of significant deals in recent years (Exhibit 12). This has included the acquisition of three oncolytic virus developers by pharma companies in 2018 alone, totalling US$770m (~A$1bn) in upfront payments and a further US$900m (~A$1.2bn) in potential milestone payments. The transactions included Merck’s acquisition of ASX-listed Viralytics for US$394m (A$502m) announced in February 2018, primarily for the purpose of studying Cavatak in combination with its checkpoint inhibitor, Keytruda. This was followed by the acquisition of two companies with preclinical oncolytic viruses, first Janssen (J&J) buying BeneVir for US$140m upfront and up to US$900m in potential milestones, and then Boehringer Ingelheim buying out its partner Viratherapeutics for ~US$236m.

The most notable licensing transaction is the December 2016 deal between Bristol-Myers Squibb and the unlisted British biotech PsiOxus for its preclinical armed oncolytic virus NG-348 (US$50m upfront and up to $886m in development, regulatory and sales-based milestones).

While the deals listed in Exhibit 12 demonstrate the high level of interest from pharma companies in oncolytic virus technologies, they also illustrate that it is a highly competitive space. While in the previous section we have discussed in some detail other vaccinia/poxviruses in development, we note that only two of the 15 oncolytic virus deals shown in Exhibit 12 involve vaccinia-based viruses (Transgene and Jennerex). This reflects the fact that several other virus species are also being developed as oncolytic virus therapies and are potential competitors for Imugene. To attract a Pharma partner, we suspect that Imugene will need to demonstrate that the high potency of CF33 at infecting tumour cells translates into improved efficacy in the clinic.

The CF33 licensing deal expands Imugene’s portfolio of immunooncology products, which is currently focused on B cell vaccines. It is the company’s second significant in-licensing deal in less than 12 months and follows on from the in-licensing of anti-HER2 and anti PD1 B cell vaccines from Ohio State University in June 2018.

If the CF33 licensing deal is approved by shareholders, Imugene’s pipeline will comprise the CF33 oncolytic virus plus two anti-HER2 B-cell vaccines that are in Phase II clinical development, plus a preclinical anti-PD1 B cell vaccine, as shown in Exhibit 13.

Imugene is exposed to the same clinical, regulatory and commercialisation risks as all biotech companies. The CF33 licensing deal is subject to shareholder approval and therefore it is not certain that the transaction will proceed, so we leave our forecasts and valuation unchanged for now. If the deal is approved, it remains to be seen whether the increased potency of CF33 in cancer cell lines and animal models will translate into improved efficacy in the clinic. The incorporation of genetic material from other poxviruses into CF33 increases the potential of unexpected safety concerns, compared to the long-established safety record of vaccinia viruses.

Given that there are a lot of other oncolytic viruses at various stages of development, we think it most likely that CF33 will need to demonstrate superior efficacy in clinical studies order to attract a pharma partner. On the other hand, it is possible that the anticipated long patent life for CF33 (if granted) could attract interest from Pharma even if efficacy only matches rather than exceeds that of competing oncolytic virus therapies, in our view. Imugene had A$21.0m cash at the end of March 2019, which should be sufficient to fund operations beyond our FY20 forecast horizon (excluding expenditure related to the CF33 transaction).