Going viral

False colour image of herpes virus. 

That certain viruses cause or promote cancer has been known for decades, prompting the development of effective vaccination against human papilloma virus  and hepatitis B and curative  drug treatments for hepatitis C to protect against or eliminate these cancer-causing (“oncogenic”) viruses.  Conversely, viruses also have the potential to be useful allies in cancer treatment.

Destruction of tumour cells as a consequence of viral infection was first observed in the 1950s, leading to empirical, and largely unsuccessful, clinical experimentation. In the last 20 years, the capability to genetically modify viruses and culture them consistently in quantity has allowed the practical exploitation of tumour-destroying viruses to be revisited. 

A variety of common viruses (including herpes, measles, and polio viruses) are “oncolytic”, that is they can selectively infect and rupture cancer cells and, in doing so, usefully redirect the innate and adaptive immune responses towards the tumour. Lysis is also believed to reveal tumour antigens normally hidden from immune system recognition and can disrupt blood vessels essential for tumour survival.

The other side of immune recognition (and memory) is that prior encounters with the myriad  of viruses that we are naturally exposed to serves to blunt the effectiveness of oncolytic viruses, by either thwarting their spread within the tumour, or through neutralization before the virus reaches the tumour. The latter can be circumvented by administering the virus directly into the tumour, albeit not a convenient way of dosing, while substantial ingenuity has been applied to improving the effectiveness of virotherapy through chemically masking viruses from immune recognition or using viral strains not normally encountered by humans, permitting systemic rather than local administration.

Other enhancements aimed at improving the safety and effectiveness of virotherapy have included genetic modification to more efficiently target molecules expressed only by tumours, to promote viral replication within cancer cells, and to express proteins that boost anti-tumour immunity. Despite numerous clinical studies across a range of tumour types, including combination with chemotherapy or radiotherapy, consistent and compelling efficacy data has largely eluded virotherapy. 

This might be set to change. Virotherapy has the very useful side effect of upregulating immune checkpoint inhibitor expression, opening up prospects for improving clinical response rates in combination immunotherapy.

This week saw Merck take a plunge into virotherapy with the acquisition of Viralytics, an Australian biotech that has successfully taken a therapy exploiting a common cold virus into the clinic. Merck have gambled $394 million on the Viralytics candidate being synergistic with their blockbuster PD-1 immune checkpoint inhibitor, Keytruda®.

Amgen, the first global biopharmaceutical company to venture into virotherapy in 2011 with the acquisition of BioVex and its lead development candidate, since rebranded as Imlygic™ (talimogene laherparepvec or “T-Vec”) has shown that the combination of Imlygic™ and the CTl4-A checkpoint inhibitor Yervoy® resulted in a doubling in clinical response rates over Yervoy® alone in melanoma patients. Amgen and Merck are co-sponsors of an ongoing Phase II clinical study evaluating T-Vec in combination with Keytruda® in sarcoma patients.

While the Merck deal offers encouragement for the raft of small and mid-cap biotechs pursuing virotherapy development, it remains to be seen whether these “living drugs” can hold their own against the multitude of more easily manufactured and administered biologic and small molecule immunotherapies also under evaluation in immune checkpoint combination studies.  That said, further tweaking could eventually establish virotherapy as a potent means of triggering innate and adaptive immune responses across a spectrum of solid tumours, irrespective of checkpoint inhibitor expression, immune infiltration or degree of tumour mutation.

Photo credit: Credit: NIH Image gallery. Bernard Heymann, Ph.D., NIAMS Laboratory of Structural Biology Research.

The pain paradox

Pain, whether short, sharp and temporary, or a constant misery, affects all of us. Around 20% of the adult population of Europe and the US lives with chronic pain (defined as "localised or widespread pain lasting three months or longer"), most commonly  arising from low back problems, osteoarthritis or rheumatoid arthritis, and from nerve damage (neuropathic pain), a common consequence of diabetes and surgery.

Pain medication is a daily and essential  part of life for millions of individuals but we remain heavily reliant on a relatively small number of drug types, several of which were introduced into common use over a century ago. The origins of aspirin and purified opiates, both available commercially since the start of the 20^th century, go back thousands of years,  to the use of willow leaf or bark teas and extracts of poppy seeds and juices.

Paracetamol (acetaminophen), was introduced in the 1950s (some 50 years after its discovery) and “non steroidal anti-inflammatory drugs” (NSAIDS), a mainstay for arthritis sufferers, during the 1970s and 1980s. Synthetic opioids and opioid reformulations, such as  fentanyl and extended release oxycodone,  have acquired notoriety through their association with epidemic levels of abuse and dependence but these drugs remain invaluable in the relief of otherwise intractable pain (blame the system, not the product).

An analysis published by the Biotechnology Innovation Organization (BIO) highlights a surprising lack of innovation in pain drug development. While the BIO report focuses on the US, the picture is not radically different in Europe and Japan.

Almost all pain medications approved for sale in the  US during the past decade have been either reformulations of previously marketed drugs or have involved drugs previously approved for other indications. Only two novel pain medications have been approved since 2009, of which one, milnacipran, had previously received approval  as an antidepressant. Industry’s last great foray in pain drug development ended in the early 2000s,  as a class of anticipated blockbusters, the cyclooxegenase-2 (COX-2) inhibitors such as Celebrex® and Vioxx®,  turned out to be potentially unsafe in a significant subset of patients.

Pain drug development activity is not at a standstill, but is hardly thriving. The BIO analysis identified 220 clinical studies for pain indications, of which over half involve novel drugs. This seems respectable until you consider that, in total,  this clinical development activity is less than one-tenth of the activity devoted to cancer drug development, and that the majority of cancer studies involve novel agents.

Investment serves as a rough and ready barometer of relative commercial attractiveness. Novel pain drug development pulled in $576 million in venture funding during the period 2007 to 2016: in comparison, $10.3 billion was showered on novel cancer drug development during the same period. As an aside, despite the uncontested economic and societal impact of drug addiction, novel substance abuse treatments have attracted only $16 million in venture investment in a decade. 

Given the number of individuals in dire need of safer and more effective pain medication, why is industry and the investment community apparently disengaged from pain drug development? Perceived risk would appear to be a major factor, with the probability of FDA approval of a novel pain drug being around 2%, well below the 10% or better for drugs addressing other indications.

Reimbursement is another key issue. The pain drug market is largely generic and healthcare payers demand hard evidence of superiority or cost-effectiveness of new treatments over established medications, something that’s not easy to establish given  the difficulty in capturing and quantifying the subjective impact of pain in clinical studies.

Even when there is a clear basis for potential advantage over existing drugs, such as with the close to market calcitonin gene-related peptide (CGRP) inhibitors which have been shown to be highly effective in migraine prevention, the inability to predict which patients are likely to respond to treatment could make payers reluctant to meet the expected $8,500 per patient drug cost.  

Pain is biologically complex  and novel targets amenable to drug discovery are, to date, few and far between. As in oncology drug development, across the board innovation, combining AI analytics, population-based studies, biomarkers, predictive animal models and better means of patient data capture will be needed to find more prospective winners and to tempt industry and investors back to pain drug development.   

Photo credit: Trying2

ASCO Clinical Cancer Advances 2018. And the winner is...

T cells ganging up on a tumour cell 

The American Society of Clinical Oncology’s (ASCO) annual review always makes for interesting reading. Immunotherapy continues to feature large [ASCO 2017 Annual Report again picks immunotherapy as “Advance of the Year"] , with adoptive cell immunotherapy being named “Advance of the Year”.

This bespoke treatment involves genetic manipulation of the patient’s own T-cells, in which an engineered virus is used to insert DNA encoding a protein capable of recognising unique molecules present on the surface of tumour cells. Transformed cells expressing chimeric antigen receptors (CAR) are cultured in the laboratory and then infused back into the donating patient. CAR-T cells then seek and destroy cancer cells. Being a “living therapy”, CAR-T cells continue to multiply to exert a persistent anti-tumour effect.

While the number of patients who have received CAR-T therapy to date is still in only in the hundreds and largely limited to certain haematological cancers, clinical response results certainly justify the ASCO accolade.  Kymriah® (Novartis), the first adoptive cell immunotherapy to receive FDA approval, achieved unprecedented response rates in children and young adults suffering from relapsed or refractory acute lymphoblastic leukaemia (ALL).  Over 80% of Kymriah® treated patients went into remission within 3 months, with 75% remaining still in remission after 6 months. 

Impressive study results have been reported for another hard-to-treat blood cancer, relapsed or refractory diffuse large B-cell lymphoma (DLBCL). Response rates to available treatments are less 10%, with survival measured in months. Complete response rates were 40% and 30% after 1 and 6 months, respectively. Moreover, treatment raised the probability of being relapse-free at 6 months to almost 74%. Another FDA approved CAR-T therapy, Yescarta® (Kite Pharma, recently acquired by Gilead) has also shown achieved complete and durable complete responses in DLBCL and other forms of B-cell lymphoma.

The major challenge of CAR-T treatment is management of potentially life-threatening adverse effects, which are a direct consequence of the engineered anti-tumour response, with the most serious being “cytokine storms” resulting from release of potent inflammatory molecules (including IFN-γ, IL-10 and IL-6). Some degree of neurotoxicity is common in recipients, although the long-term consequences of this are unknown.

Other significant challenges are logistics and cost: as personalized living therapies, cell collection, transformation, expansion and delivery to the patient is a complex series of steps, with consistent manufacture of a safe and reliable product being the most demanding. While the list prices of Kymriah® and Yescarta® are $475,000 and $373,000, respectively (Novartis offers a rebate where patients do not respond to treatment), side effect management and other support costs can bring the total medical bill close to $1.5 million per patient. This, and complex reimbursement issues, are at least in part, factors in the low uptake of CAR-T therapy. Better definition of cost-effectiveness is needed.

Despite these challenges, a variety of investigational adoptive cell immunotherapies is in the works for cancers other than B-cell malignancies, with studies underway in multiple myeloma (Bluebird Bio, Kite Pharma); ovarian cancer (Juno Therapeutics), and glioblastoma (Mustang Bio, Aurora Biopharma). Several “next generation” candidates incorporate drug-sensitive “safety switches” that allow the anti-tumour response to be turned on or off.

Moving from bespoke to “off the peg” therapy using cells collected from healthy donors (“allogeneic” cell therapy) could substantially reduce the overall cost of cellular immunotherapy. Early clinical studies with an allogeneic CAR-T cell product developed by Cellectis are underway.

Cellular immunotherapy development has proved to be a rocky road. Clinical study deaths have resulted in the abandonment of a CAR-T candidate by Juno Therapeutics [CAR-T: A wheel falls off, but still rolling], and clinical trial holds imposed on Cellectis (although recently lifted) and as of this week, Bellicum Therapeutics, with three unexpected deaths occurring in subjects under treatment for haematological cancers. Growing experience may allow early identification of study subjects at increased risk of severe adverse events or development of safer investigational treatment regimens.

Commercial success could prove elusive for several developers with high product cost, small addressable patient populations, safety issues, and competition from both cellular and non-cellular immunotherapies, particularly in the B-cell malignancy segment, constraining uptake. Giants such as Novartis can afford the necessary infrastructure, flexible pricing and slow rate of uptake, but smaller players may join Kite Pharma and Juno Therapeutics in being acquired by companies with deep pockets.  

Photo credit: NCI/NIH, Alex Ritter, Jennifer Lippincott Schwartz, and Gillian Griffiths

A single shot

Each year, healthcare agencies undertake the crucial task of choosing which particular influenza virus strains will be included in vaccines to be manufactured and deployed in time for the next flu season. Selection is assisted by history and epidemiological surveillance but, as in this, the 2017-2018 flu season, mismatching of vaccine composition and the actual infecting strains greatly reduces the impact of vaccination.

That flu viruses regularly undergo changes that render vaccines ineffective has been known since the introduction of large scale flu immunization campaigns in the 1950s, leading to the World Health Organization setting up a global influenza surveillance and response system.

Seasonal flu infection is largely due to influenza Type A, and to a lesser extent, the generally less severe influenza Type B. Type C influenza strains cause only mild and sporadic infection. New “pandemic” strains, to which there is no widespread immunity, can pop up with devastating effect. The “Spanish Flu” of 1918-1919 may have caused 20-50 million deaths (more than in the Great War), while the 2009 “swine flu” pandemic may have caused more than half a million deaths.

Flu vaccines work by inducing a neutralising antibody response to haemagglutinin (HA), a protein expressed on the surface of the virus essential for infection and spread. HA can, unfortunately, undergo regular “antigenic shift”, necessitating annual adaption of seasonal flu vaccine composition to match the characteristics of the infecting strains.  

The logistical challenge of getting the right vaccine ready at the right time (a task still almost completely dependent on growing the selected virus strains in hen’s eggs), along with the need to be better able to deal with future flu pandemics, are powerful incentives to develop so-called “universal” flu vaccines, capable of inducing long-lasting or even lifelong protective immunity which is not compromised by the mutability of the HA protein. Moreover, recombinant protein vaccines would simplify large-scale manufacture and speed up vaccine availability in the face of a pandemic. 

Government and industrially funded research is pursuing a variety of routes towards a universal vaccine. A leading contender being developed by Vaccitech, an Oxford University spin-out (backed in part by Google’s venture fund), combines two highly-conserved core proteins (nuclear protein and matrix protein 1) that are naturally expressed by influenza A strains. A Phase II clinical study, which aims to eventually recruit over 2000 subjects aged over 65, is now underway, with recruitment of the initial tranche of volunteers announced earlier this month. It’s hoped that the vaccine will elicit both antibody and cellular immune responses to generate long-lasting protection.

A not dissimilar approach is being pursued by BiondVax. Various antigenic sequences (“epitopes”) present in HA, nuclear protein and matrix protein have been selected for their ability to elicit both antibody and cellular responses and knitted together in a single recombinant protein. The company hopes to initiate a Phase III study in Europe later this year, involving 7,700 subjects aged 50 years or older older, with at least half of participants being over 65 years of age.

Promising pre-clinical candidates include synthetic nanoparticles incorporating multiple copies of a conserved matrix protein developed at Georgia State University, while another Georgia group (in collaboration with Sanofi Pasteur) has used computational analysis to cherry-pick and combine different epitopes from HA proteins to induce antibodies broadly protective against one particularly important flu strain and its variants.

DNA vaccination, in which a piece of flu virus DNA is injected and then expressed as an immunising protein by the subject’s own cells, has been shown to reduce the effects of flu infection in primates. An advantage of DNA vaccination is that the “immunising” strand can encode several different conserved flu proteins to give broad protection. On the other hand, despite the wide optimism over the utility of DNA vaccination expressed during the last 25 years, only a handful of veterinary DNA vaccines have obtained regulatory approval.

"All done. See you again in five years"

While there’s no shortage of ingenuity and endeavour, a truly “universal” flu vaccine remains a good way off. Progress has so far been largely confined to influenza A viruses and the ideal universal vaccine will need to provide protection against influenza B (and ideally, pandemic strains and those of animal origin).

Science is only one barrier. As with other vaccines, large scale studies will be required to establish efficacy over conventional vaccines and safety, particularly in those at most risk from flu infection (young children and the elderly). Meaningful evaluation and deployment of a universal flu vaccine is likely outside the capacity of industry or national agencies and will require regional, if not global, co-operation and co-ordination if we are to finally attain adequate protection against “la grippe”.

Photo credit: CDC and Doug Jordan, M.A

The C.difficile epidemic. Sugar rush?

Clostridium difficile

Clostridium difficle features high on the list of public (health) enemies, being a notorious cause of life-threatening hospital-associated infection, particularly in the elderly.

The rise of C.difficile infection from being an occasional complication of hospitalisation to full-on epidemic in Europe and North America in less than a decade coincides with the appearance of more dangerous strains, readily identifiable through differences in the genes that encode for ribosomes, the macromolecular complexes that translate genetic information into proteins.

Ribosmal gene fingerprints (“ribotypes”) indicate that, from the early to mid-2000s, the appearance of strains of two particular C.difficile ribotypes, 027 (also known as NAP1) and 028, correlated with a steep increase in severe C.difficile infection and death. These so-called “hypervirulent” strains are much studied and while certain characteristics explain their propensity to cause severe infection, including toxin production, greater adherence to human gut cells and prolific spore formation, the reason for the sudden appearance of the 027 and 028 strains remains an epidemiological mystery.

An international research group centred at Baylor College of Medicine has raised the intriguing possibility that the C.difficile epidemic may be in fact self-inflicted, with the rise in hypervirulent infection being fuelled by an  increase in dietary trehalose, a food additive which hit the mass market in the early 2000s.

Trehalose, a natural sugar, has the handy properties of being able to survive high temperatures without browning while preserving cell structure on freezing. Low cost manufacture of this sugar resulted in trehalose becoming ubiquitous in processed foods, jam, fruit juice and ice-cream (with some brands containing over 10% trehalose).

  

The hypervirulent 027 and 028 C. difficile strains are capable of thriving on low concentrations of trehalose as their only source of carbon. Sequencing has revealed a genetic variation in 027 strains and the presence of additional genes in 028 strains, each of these differences from “normal” C.difficile strains conferring the ability to utilise low concentrations of trehalose.

Knocking out the genetic variant in a 027 strain reduced its virulence in mice, while adding trehalose to the diet of infected mice resulted in increased mortality, possibly through increased toxin production rather than an increased rate of bacterial overgrowth. Fluid collected from the small intestines of three volunteers on normal diets contained sufficient trehalose to trigger expression of the trehalose-metabolising gene in 027 strains, but not in other C.diffcile strains.

Although circumstantial, the discovery that hypervirulent C.difficile stains are uniquely adapted to make good use of low trehalose levels, together with the close temporal fit between the spread of these strains and the upswing in dietary trehalose makes for a compelling story.

On the other hand, increased trehalose consumption alone does not explain national differences in the appearance and dominance of hypervirulent strains, nor the rapid decline of these strains in the UK and other countries. A history of antibiotic use is an established risk factor for severe C.difficile infection and it’s possible that antibiotic-induced changes to the gut microbiome confer a selective advantage to hypervirulent strains.

And it’s far too early to start worrying rationally about trehalose, although minimizing intake in those with established risk factors for C.difficile infection might be worth consideration.

Photo credit: CDC/ James Archer

Could bugs beat broccoli in cancer prevention?

Generations of parents (including this writer) have berated their offspring with the message that green stuff, while looking and tasting yucky, is really, really good for you. And, for all sorts of reasons, including a reduction in lifetime cancer risk, it’s a message worth heeding.

Green is good...

Many fruits and vegetables are rich in metabolites which act directly or indirectly on pre-cancerous or cancer cells to constrain or reverse tumour formation, albeit with limited effect. Daily consumption of a large quantity of fruits and vegetables is required to get even close to preventative metabolite levels, something that does not sit conveniently with most modern lifestyles.

In a neat piece of lateral thinking, a cross-disciplinary research group at the National University of Singapore has engineered a strain of one of the commonest gut bacteria, Escherichia coli, to latch onto colorectal cancer (CRC) cells and express an enzyme that converts an abundant dietary metabolite into a potent cancer inhibitor.

The metabolite of interest, glucosinolate, is found in cruciferous vegetables (brassicas to us Brits), including broccoli, the unfairly maligned Brussels sprout, cauliflower and rocket. Broccoli and its ilk are rich in glucosinolate, a precursor of sulforaphane, a compound known to inhibit cancer in a variety of ways. Sulforaphane production is catalysed by  myrosinase, a plant enzyme that's largely lost during cooking. Gut bacteria can  break down glucosinolate, although not with sufficient efficiency to  maintain useful sulforaphane levels.

The Singapore group was able to demonstrate that, in the test tube, an engineered E.coli strain, designated EcN, inhibited growth of CRC cells, but not cells from breast or stomach cancers or smooth muscle cells.

When tested in a mouse model of CRC, EcN was found to bind to tumour cells in the gut and substantially increase the plasma concentration of metabolites akin to glucosinolate. Tumour burden was reduced by 75%. One unwanted effect was an increase in bleeding in the gut compared with untreated mice, although this may have been related to the mouse cancer model, and not a direct consequence of tumour cell binding.

While raising more questions than answers, this research does hint at the future prospect of being able to boost the limited anti-cancer effect of diet through supplementation by engineered probiotics. Caution is warranted: our understanding of interactions within the gut microbiome is in its infancy and introduction of an engineered arriviste could have subtle and unpredictable consequences.  And, as with other dietary approaches to cancer prevention, generating robust efficacy and safety data presents a significant challenge.

Photo credit:  khumthong at FreeDigitalphotos.net

Gene therapy and fair value.

Retinal pigment epithelial cells in
RPE65-mediated retinal disease

Back in March, I wrote a piece regarding the coming impact of high ticket gene therapies on healthcare budgets. December saw the FDA approval of Spark Therapeutic's Luxturna™ (voretigene neparvovec), a unique treatment for biallelic RPE65-mediated inherited retinal disease, a form of Leber's congenital amaurosis, which results in early onset, progressive loss of vision.

Predictably, Luxturna's approval has reopened the debate around the cost of leading edge therapies. At around $850,000 to treat both eyes, Luxturna™ pricing is somewhat lower than the $1 million plus price tag anticipated by industry analysts, although it's still the most expensive drug in the US by list price.

Justifiable? Perhaps. Gene therapy product approval is not a guaranteed path to riches. As with other genetic disorders, the potential treatment population is small, being around 1000-2000 sufferers in the US, with around the same number in Europe. Moreover, Luxturna™ is a one-time treatment. While even modest uptake should cover Spark's development costs, the overall return to Spark will be, by pharma standards, unremarkable.

Spark appears pragmatic in its approach to reimbursement, offering insurers rebates should patients fail to achieve an agreed degree of benefit, although with only limited and short-term study data available, defining a improvement for rebate purposes will not be easy. Spark is also thought to be considering an annuity model, allowing insurers to pay over time [see update of 12th January below]. 

So much for cost, but what about value? Although not an easy calculation, tallying the lifetime benefit accruing from reduced direct and indirect medical costs, increased individual economic activity and quality of life improvement, might come close to justifying Luxturna’s price tag.

A draft assessment published by the Institute for Clinical and Economic Review published just prior to Luxturna's approval concluded that, although likely to result in better outcomes than standard care, Luxturna would probably not prove to be cost-effective at an assumed acquisition cost of $1 million.  Another crank of the spreadsheet incorporating the actual drug price and post-approval efficacy data, particularly the durability of benefit, could tip the balance in Luxturna's favour.

The UK's National Institute for Health and Care Excellence (NICE) recently concluded that, compared with the cost and risk associated with stem cell transplantation for the treatment of adenosine deaminase deficiency–severe combined immunodeficiency (ADA-SCID or "bubble boy" syndrome) GSK's gene therapy, Strimvelis™, provided both the best treatment option and value for money, despite its  €594,000 (around £505,000) price tag.

Although invariably flawed, cost-effectiveness analysis needs to be at the centre of gene therapy pricing and adoption debates. Such analyses may not always prove favorable, but without an objective means of establishing fair pricing and reimbursement, gene therapies could become out of reach for many patients. The commercial abandonment of Glybera™,a gene therapy for lipoprotein lipase deficiency and announcement of GSK's intention to abandon Strimvelis® (and rare disease therapy development in general) are portents that should not be ignored.

Photo credit: National Eye Institute, National Institutes of Health.