BioByte 088: AlphaProteo, Merck's Blockbuster Recipe, Dark Genome and Grand Challenges in Healthcare
Welcome to Decoding Bio’s BioByte: each week our writing collective highlight notable news—from the latest scientific papers to the latest funding rounds—and everything in between. All in one place.
What we read
AlphaProteo generates novel proteins for biology and health research [Deepmind, September 2024]
Deepmind has released AlphaProteo, its first system for designing high affinity protein binders. How does AlphaProteo work? First, a target protein is given and potential binding sites are specified, then, AlphaProteo produces a candidate protein binder to the target at the suggested locations. Whilst important for drug discovery, this model does still require an understanding of potential binding sites, meaning the classically ‘undruggable’ targets with a less defined structure may still be out of reach.
Results from AlphaProteo are mixed, with high in vitro success is some therapeutically relevant cases but in terms of binding affinity, a AlphaProteo designs would need to be supplemented with current methods.
Using AI and brain implants, researchers reconnect ALS patient with their lost voice [Conor Hale, Fierce Biotech, August 2024]
Significant strides are being made with BCIs as of late. In July 2024, researchers at the UC Davis Neuroprosthetics Lab enabled a patient who had lost his ability for speech to ALS four years ago to once again communicate in his own voice. This feat was achieved using neural sensors developed by Blackrock Neurotech which were implanted in the motor cortex and AI text-to-speech software trained on prior voice recordings of the patient. While the utilization of BCIs for non-verbal communication is not novel, this breakthrough illustrates a marked improvement for accuracy. Researchers reported a near-perfect accuracy rate of 97%, constituting a substantial improvement from prior thought to speech technologies which were plagued by frequent word errors. This success story is one of many arising as a part of decades of work being done through BrainGate, a clinical trial consortium implementing BCIs to restore communication and mobility for individuals with paralysis.
The Startling History Behind Merck's New Cancer Blockbuster [David Shaywitz, Forbes, 2017]
While Keytruda is not a new cancer blockbuster anymore, it has been Merck's most important revenue driver over the past years and is on track to reach $30 billion in sales by 2026 (Reuters, 2024).
It's a fascinating story that is still worth recounting. Here are the major highlights of the drug’s history:
The original program was initiated in Oss, Netherlands in the lab of Andrea van Elsas as part of the company Organon, part of the conglomerate Akzo Nobel.
The idea was to develop PD1 antibody agonists to treat autoimmune disease (such as rheumatoid arthritis). The PD1 program ultimately yielded no good agonists but exceptional antagonists.
The team explored various applications of PD1 antagonists, such as anti-infectives, but decided to focus on immuno-oncology. As the team continued the IND-enabling work, Organon was acquired by Schering Plough in 2007. The programme was close to being scrapped, but van Elsas was permitted to continue to prove its value to the new executives.
As the programme started “to gain momentum, Schering itself was acquired by Merck in 2009, initiating, of course, a new reprioritization process for the integrated companies.” This time the PD1 programme was halted and placed on the out-license list. A term sheet was drafted to sell the asset at negligible cost.
However, by mid 2010, BMS posted a successful Phase III study “in refractory metastatic melanoma involving ipilimumab (Yervoy), a drug targeting the different but related T-cell inhibitor molecule CTLA4, pointing to the potential efficacy of the so-called “checkpoint inhibitor” approach.” and secondly that “word began to filter through to Merck that nivolumab, or nivo (brand name: Opdivo), an earlier-stage molecule targeting PD1 that BMS also acquired though its recent acquisition of Medarex, was showing signs of promise”
Merck decided to activate work on pembro. In its Phase I study, UCLA PI Antoni Ribas said “Among the first 7 patients we enrolled, 6 patients had objective responses. I realized I was probably lucky and this high rate of responses would not hold up forever, but also realized that [pembro] was different from all the other cancer immunotherapy we had tested to date, all of which had response rates in the 10-15% range.” and he worried that Merck was “going too slow in recognizing how important this was.”
Merck received breakthrough designation status given its results in advanced melanoma. Merck decided to expand to NSCLC.
As Merck was years behind BMS, “only chance for rapid approval would be to demonstrate unusually high response rates in second-line therapy. The challenge was how to achieve this–and they turned to a biomarker.”. They chose high expression of PDL1 in tumors as the marker. Roger Perlmutter, who replaced Peter Kim as head of Merck Research Lab, continued to back the use of the contentious PDL1 assay in the NSCLC second-line programme and in the new first-line treatment for the same indication.
In 2016, Merck’s study was reported to be strongly positive for first-line treatment, whilst BMS’s study, which used a lower threshold for its biomarker, did not.
Perlmutter’s key impact was to recognise the value of the asset and virtually guide all work at Merck towards this program. “When you realize the potential of something like this,” he explained to me, “you need to bet your career on it.”
How Machines Learned to Discover Drugs [Dhruv Khullar, The New Yorker, 2024]
In an inspiring piece of journalism, physician and writer Dhruv Khullar traces the discovery of potent antibiotic vancomycin among other antibiotics and medicines. He highlights soil discovery and rote drug synthesis by studying bacteria before exploring how machines and computation is speeding this process up.
I highly recommend taking a few minutes to read this piece for its storytelling but if you’re in a rush, here are some fun points raised:
One gram of soil has up to ten thousand different types of bacteria
New Jersey’s official state microbe is streptomycin, an antibiotic found in the “heavily manured” soils of Jersey that is used to treat tuberculosis
No useful antibiotics have been found via high throughput screens of existing chemical libraries
Halicin was discovered with early AI-assisted screening methods at the Broad
Pharmaceutical companies don’t invest much in antibiotics because they are not taken for long durations
No new major classes of antibiotics were brought to market between 1962 and 2000
Restoring hippocampal glucose metabolism rescues cognition across Alzheimer’s disease pathologies [Minhas et al., Science, Aug 2024]
A new paper out in Science from Katrin Andreasson’s lab at Stanford helps elucidate the interplay between several disease mechanisms in Alzheimer’s disease. In addition to hallmark pathologies like the accumulation of amyloid-β plaques and tau neurofibrillary tangles, growing evidence suggests that metabolic dysfunction and neuroinflammation play crucial roles in AD disease progression. Impaired cerebral glucose metabolism is a well-established feature of AD, often detectable years before clinical symptoms manifest. Additionally, chronic activation of the immune system, particularly involving microglia and astrocytes, contributes to neuroinflammation and exacerbates neuronal damage. The intricate interplay between these factors – protein aggregation, metabolic alterations, and immune responses – creates a challenging landscape for understanding AD pathogenesis and developing effective therapies.
The study investigated the role of the enzyme IDO1 and its metabolite kynurenine (KYN) in AD pathology, focusing on their effects on astrocyte metabolism and neuronal support. IDO1 inhibitors have previously been investigated as cancer drugs because IDO1 is involved in inhibition of the immune system. In this study, the researchers discovered a novel function of IDO1 in the brain in which IDO1 activation by amyloid β and tau oligomers suppressed astrocyte glycolysis and lactate production. Conversely, IDO1 inhibition restored astrocyte metabolism, hippocampal glucose metabolism, spatial memory, and long-term potentiation in an AD mouse model.
The findings reveal a general mechanism of neuronal dysfunction that cuts across both amyloid and tau pathologies, suggesting that IDO1 inhibition could be a promising approach for treating AD and potentially other neurodegenerative diseases characterized by protein aggregation. Importantly, these data suggest that altered glucose metabolism is directly involved in protein aggregation. By restoring astrocyte metabolism and astrocyte-neuron metabolic coupling, IDO1 inhibition represents a novel therapeutic avenue distinct from current amyloid-lowering strategies.
Cranioencephalic functional lymphoid units in glioblastoma [Dobersalske et al., Nature Medicine, 2024]
Historically, the brain was perceived to be an immune-privileged organ with limited intra-cranial or trans-cranial immune trafficking. Recent work has continued to reveal how the brain is intimately connected with the immune system, and this interplay is critical for responses to brain tumors and various diseases. Here, the authors discovered a unique population of tumor-killing immune cells in the skull bones of patients with glioblastoma, the most aggressive form of brain cancer.
Using the radioligand 68Ga-Pentixafor, which binds CXCR4 in immune cells, they found significant labeling in the cranial bone of glioblastoma patients compared to controls. Fluorescent imaging further revealed aggregates of CD45+ immune cells forming solid arrangements around microvessels in the spongy bone tissue called the diploë. These findings indicated a process where immune cells accumulate in the tumor-adjacent skull bone.
To characterize the immune cell types, the researchers performed single-cell RNA sequencing on CD45+ cells isolated from fresh surgical bone samples, peripheral blood, and glioblastoma tissue. Surprisingly, while tumor samples were enriched in immunosuppressive myeloid cells, skull bone samples contained abundant T cell fractions, particularly CD8+ effector memory T (TEM) cells. Functional assays demonstrated that skull bone-derived CD8+ T cells exhibited increased tumor reactivity compared to peripheral blood and tumor-derived T cells. The researchers also noted shared T cell clonotypes between skull bone and tumor, suggesting recirculation of tumor-reactive T cells across compartments.
Further work revealed that skull bone-derived CD8+ T cells had increased proliferative capacity, the ability to reproduce memory and effector phenotypes, and sustained tumor reactivity compared to tumor-derived T cells. Single-cell transcriptomic analysis showed a wide developmental range of CD8+ T cell differentiation in the skull bone, with enrichment for activation and effector function gene signatures, indicating a locally-driven anti-tumor response. Preliminary clinical evidence supported these findings, as presurgical CXCR4 PET-CT imaging data revealed that radiotracer enhancement in the cranial bone adjacent to the tumor positively correlated with improved patient survival. The authors thus proposed that the skull bone serves as a unique niche for the differentiation of tumor-reactive CD8+ T cells in glioblastoma, forming "cranioencephalic functional lymphoid units", and suggest that preserving the integrity of these units during treatment may be crucial for supporting anti-tumor immunity and immunotherapy responses.
What we listened to
Notable Deals
What we liked on socials channels
Field Trip
Did we miss anything? Would you like to contribute to Decoding Bio by writing a guest post? Drop us a note here or chat with us on Twitter: @ameekapadia @ketanyerneni @morgancheatham @pablolubroth @patricksmalone
The Merck story is a great example of where serendipity (and luck) meets real leadership. It would have been easy to just run the programme at a "standard" pace as opposed to go all in. It takes someone of stature who is comfortable with the risk of failure to be able to see the potential and act accordingly and override the various governance committees.