Member Info for Burble

Bermuda

Catching up on your question early regarding Exhibit 6 in the Trinity Delta research notes (https://www.trinitydelta.org/research-notes/a-sweet-source-of-financing/).

Basically, any biological product you put into a human can result in an immune response being raised against that *IF* the immune system thinks of it as foreign. In the case of antibodies, if you generate a mouse antibody and inject it into a human, the human immune system will attack it as it thinks of it as a foreign invader. Hence, scientists aim to engineer a mouse antibody to look as close to a human antibody as possible, whilst still functioning.

There are a range of ways they can do this. If we refer back to this diagram (https://bit.ly/3hbJNn1) the part of the antibody which binds the target region of interest (the epitope) is found in the VH and VL domains (variable heavy and variable light domain). Within each of these domains you have three loops known as complementary determining regions (CDRs) which do the actual binding. These loops vary between different antibodies, both in length and in amino acid sequence. Hence giving the body nearly infinite diversity to raise an immune response against any pathogen it encounters. The exact sequence of these loops, coupled to the sequence of regions just outside of these within the VH and VL domain result in the loop adopting a specific shape which is complementary to the surface of the target a bit like a lock and key.

A chimeric antibody consists of a human antibody, but the VH and VL domains are replaced by the mouse versions (https://bit.ly/35hGIPX). So 4 out of the 12 domains that make up the antibody are mouse (33% mouse). A humanised antibody consists of a human antibody, but with the CDR loops replaced by mouse versions (so approximately 3% mouse). The closer you get to a fully human antibody, the less problems you have with recognition and rejection of that molecule. Now the problem you have is that the flanking regions either side of the CDRs are also key to how each of those loops are shaped. These flanking regions vary between mouse and human. So sometimes, ideally you may want a humanised antibody (CDR loop grafted), but you lose the ability to bind the target antigen, so you have to include more of the mouse flanks to ensure the CDR loops work. Occasionally though, you have to use the whole mouse VH/VL domains to maintain target binding.

Implications, ideally you want a humanised or fully human antibody, as this prevents the body from generating a human anti-mouse antibody (HAMA) response. Of these four antibodies humanised would be the more preferable, but there are some chimeric antibodies on the market (Dinutuximab - neuroblastoma, Rituximab - autoimmune diseases, Infliximab - Crohn's disease to name a few.)

Balerno

As per Inan's response, I don't anticipate you would see much benefit synergistically between glycan mAbs and keytruda. Keytruda works by unmasking cells to the immune system, leading them not to recognise the cancer cells as 'self'. Glycan mAbs work by binding sugar molecules on the surface of the cancer cell. In doing so they label they can work through two mechanisms - the traditional route whereby the antibodies in essence act as markers to the rest of the immune system, which comes along and kills the labelled cell through two mechanisms. Either by complement-dependent cytotoxicity (a mechanism which recruits proteins in the immune system to form holes in the cell which then kills it) or by antibody-dependent cellular cytotoxicity (a mechanisms which recruits an effector cell (which could be a natural killer cell, macrophage, neutrophil or eosinophil), which comes along and kills the labelled cell by secreting a number of cytotoxic molecules onto the cancer cell).

The Avidimab addition to these anti-glycan mabs (and any mAb if engineered with Avidimab in) allows the direct killing of a cell labelled with the antibody WITHOUT the need for either of the two mechanisms above. As the effector functions are retained even with Avidimab added, you may get all three mechanisms working in parallel.

To a point a benchmark has been set, but this is in HPV positive cervical cancer. Melanoma is at times a different beast all together. So whilst there are similarities between the two (DNA vaccine combined with Keytruda) lets not suggest that they are entirely alike.

The 50% is interesting, because more recently, the highly toxic combination of ipilimumab and nivolumab was also approved for the treatment of advanced melanoma. This has an overall response rate of about 60%, though despite this PD-1 monotherapy remains a key treatment for patients with advanced melanoma. Mainly due to its relatively low toxicity and acceptable response rate.

Pembrolizumab (keytruda) or nivolumab (Opdivo) are both mAbs targetting PD-1 receptor. Both show a response rate of 30-40%.

Hence, SCIB1 combo trial needs to be better than monotherapies alone. If response rates are near those seen in the ipilimumab and nivolumab combo, then we need to have a better safety profile. If response rates are better than those seen AND has a better safety profile then that'd be an incredible result.

2/2/end

Other mechanisms which may result in us not seeing synergy involve lack of interferon-gamma responsiveness in tumour (i.e. CD4+ (Th1) T-cells being unable to do their job) - Loss of the interferon-gamma responsiveness also would affect CD8+ cells and could render SCIB1 ineffective entirely), other mechanisms include things like changes to the tumour microenvironment to name a few.

I therefore believe, only time will tell when it comes to trials as to whether it is a success or not. From a commercial side of things, if a combination trial failed (especially if it lead to patient deaths) that most likely could be the beginning of the end for the ImmunoBody platform – even if technically it was caused by the Keytruda and the underlying immunobody platform was safe and well tolerated, I think investor confidence would be knocked significantly which would harm the business as a whole. With all pharma companies their cash runway is what will keep them in business long enough to either get taken over or take their product all the way through trials alone and into the clinic.

Inan

Pembrolizumab targets the PD-1 receptor, binding and blocking it. This receptor normally prevents the brakes on the immune system from working and so allows the immune system to attack it’s own cells. Hence, allowing the immune system to target and destroy cancer cells. Tumours tend to generate a range of mutated proteins which serve as markers on the cell, highlighting to the immune system that they are in fact cancerous. Normally the immune system would be prevented from clearing these cells through the self-checkpoint system (i.e. PD-1 pathway). By blocking this, the immune system is able to recognise these ‘non-self’ cells (e.g. cancerous) and destroy them.

SCIB1 induces a specific and potent T-cell response that is directed to cancers showing epitopes from the proteins gp100 and TRP-2. These epitopes can also be thought of as markers present on the cell surface. So in essence, you have a cell showing markers of cancer (gp100, TRP-2 and a range of other mutated proteins), you have T-cells which recognise gp100 and TRP-2 specifically and you have PD-1 inhibitor (Keytruda) which now unmasks the immune system further. One would assume that you would get some form of positive effect from combining these two drugs.

However, as with all combinations you can run into problems. There are a range of adverse effects which are seen with Keytruda, these include severe immune-related adverse events including lung inflammation, inflammation of endocrine organs, inflammation of the gut, liver and kidneys. Treatment has also lead to fatal cases.

So far SCIB1 has induced dose-dependent T-cell responses in nearly 90% of patients with no serious adverse or dose limiting toxicities. This is without removing the brakes to the immune system. A worry, is that SCIB1 and Keytruda combination therapy will result in failure, not because of SCIB1 but because of the adverse and serious adverse events caused by Keytruda that haven’t been seen in the mice models. Combo trial would have to demonstrate an incredible benefit to patients, to offset any adverse events that may arise from using both together.

We also know that PD-1/PDL-1 inhibition therapies become ineffective in a significant percentage of patients, some who respond early develop resistance to these with relapsed disease. We also know that checkpoint blockade is only effective in a subset of patients and may are only partial responders. Mechanisms for both types of resistance are complex and caused by multiple reasons. The hope is that in patients who do respond, the combo clears the cancer quick enough to prevent resistance occurring. Sadly, if this isn’t the case, then you will not see synergy between these two, but more you’ll see SCIB1 working and PD-1 inhib not.

1/2

RP

When you say 'whether it is Lindy's approach or TCR or CAR-T', these are not all the same thing technically.

Tumour-infiltrating lymphocyte (TIL) therapy involves harvesting naturally occuring T-cells that have already infiltrated a patients tumour. Then in the lab activating them, expanding them in culture, then re-infuse these back into a patient whereby they seek out and destroy the tumour.

TCR therapy is slightly different but involves taking T-cells from patients, equip them with a new T-cell receptor (which enables them to target specific tumour antigens), activating them, expanding, then re-infusing. Allowing doctors to choose optimal targets for each patient tumour and distinct types of T-cell to engineer.

TIL/TCR therapies can only target and eliminate cancer cells that present the target antigens bound by the major histocompatability complex (MHC).

CAR-T technology allows scientists to equip a patients T-cells with a synthetic receptor (chimeric antigen receptor) but when re-infused allows the T-cells to bind to cancer cells even if the antigen is not present on their surface bound to MHC.

There are many bottlenecks in all of this is a) having enough healthy T-cells to harvest, b) successfully modifying enough of these T-cells in the test tube, c) being able to expand these enough to enable you to infuse them back into the patient, d) the modified cells not being destroyed by the immune system, e) the modified cells continuing to expand in the patient f) not losing the CAR-T or new T-cell receptor, g) the patient not losing the target antigen. All of this costs time, money, expertise and has additional risks to the patient.

SCLP does away with many of these bottlenecks with their cancer vaccine platform. ImmunoBody is quick and easy to produce as it's a DNA based vaccine (reduced cost and bigger target market), target antigens can be plug and play (allowing rapid development of new constructs) , you're not taking cells out of the patient (beneficial as you don't have to wait months for your CAR-T to be ready), you're vaccinating them and letting their immune system do the work (harnessing natural systems), you're not adding in a genetically engineered T-cell (lower risk of the immune system actually attacking the modified T-cells).

Pharma are piling cash in the billions into CAR-Ts across the world. Just look at the private company ImmunoCore managing to raise $320m in 2015, then come back for another $130m earlier this year. If SCLP and the ImmunoBody platform is successful, SCLP will completely blow the CAR-T world out of the water. Oh wait and then there's also Moditope, Avidimab, TaGmAbs etc along side this....

Interesting. I see their website says this 'ImmunoBodies™ are antibody-antigen fusion proteins. ImmBio has developed this technology and established a cross-license agreement for ImmunoBodies with Scancell Ltd.'

Would we get any royalties from this? or am I right in thinking one of the board are co-inventors of this with Lindy hence the cross-license agreement?

Inan,

Are Immbio still a thing? Their website looks upto date, they say they have passed phase 1 trials but unsure whether or not they're still a thing.

Avidimab from what I gather is both the increased avidity AND the direct cell killing. I suspect when LD transferred these residues into the human cells it was done on the belief the direct-cell killing would be transferred. I suspect the acidity increases are a separate but serendipitous discovery which happened to be seen through these experiments. Without looking at physical structures, I can’t comment on what these residues look like chemically bar the fact they are charged (but charged residues are found on the surface of proteins so charge alone isn’t that exciting to this discovery)

RP

The Avidimab discovery is detailed within the recently published paper (DOI: 10.1158/0008-5472.CAN-19-3599) https://pubmed.ncbi.nlm.nih.gov/32532823/.

Basically, an antibody is made up of four chains (two heavy, two light) (https://bit.ly/323owrk). The heavy chains are made up of four domains (VH1, CH1, CH2 and CH3) (picture this as four sausages on a string). Each light chain is made up of two domains (VL, CL). The CH2 and CH3 regions are known as the Fc region of an antibody.

Lindy et al. had identified mouse antibodies which were capable of directly killing cells. Human antibodies lacked this ability. Antibodies are relatively robust proteins and with a vast amount of commonality between species, so you can basically chop off parts of an antibody, then take the same part from an antibody from another species and replace it. The team, took the mouse Fc region and basically replaced sections of the human antibody, by grafting parts of the mouse antibody into this. Then tested to see whether the human antibody gains the ability to directly kill cells. If the human antibodies gained this function, then you generated another construct which used a slightly smaller section of graft. Test again. Repeat this a number of times and you end up identifying the exact residues which are capable of allowing antibodies to directly kill cells.

This domain shuffling (e.g. taking whole sausages from the mouse chain and replacing their corresponding human ones) or 'chain shuffling' (e.g. cutting the sausage into pieces, then swapping these internal pieces with the corresponding pieces in the other sausage) is even more impressive when you realise a protein as a long string of individual building blocks (amino acids) all folded up into a 3d structure and not as a sausage like cylinder. Taken another way, imagine a long piece of string bunched up into a ball (this could be one of these domains e.g. CH1, CH2, CH3...). If you coloured a patch on the surface of that ball of string, then unwind the string into a straight line, you realise that the coloured patch of the string isn't continuous in a long section, it's a number of coloured spots spread in multiple places along that length of string. The coloured patch only appears when the string is correctly folded back into the same ball shape it was before. Hence Lindy has not only been able to identify which general region of the Fc domain the direct killing is attributed to, but also the individual residues that form the patch. Hence, she identified two regions one within CH3 domain (the major contributor) and one within CH2 (the minor contributor) which map onto discontinous (e.g. non-sequential) residues 286-306 and 339-378.

Come full circle, take those now known patches, transfer them into the same location on any other antibody and you gain direct-cell killing whilst retaining the functionality of the rest of the antibody. Patent it as AvidiMab and you now have a new platform to

Hahaha nope! No amino acids here. I did realise that 'immunodominant' and 'epitope' can be quite over the heads of people. Hence an analogy is better! I think a lot of people think that injecting a vaccine encoding/with the whole spike protein leads to a better response - but we've shown that's not always the case. The protein I worked on for my PhD, we vaccinated two animals and obtained 4.2 x 10^13 and 6.2 x 10^13 antibody producing cells respectively from a 100 mL blood sample. We selected from these antibodies which targetted our protein of interest and then sequenced 192 of these and out of the 192 antibodies we sequenced, we ended up with around 20 different families of antibodies (similar DNA sequence), yet when you looked closer, these actually only bound 3 unique epitopes on the protein. So from a whole immune response from two animals, actually only three epitopes were immunodominant. The rest of the protein the immune system didn't 'see'.

Bunsie

If you take the Oxford vaccine, then you almost definitely would be able to take Covidity. The Oxford vaccine ChAdOx1 nCoV-19 vaccine (AZD1222) consists of a modified simian adenovirus vector which is incapable of replicating inside the body.
The vector contains the full-length structural surface glycoprotein (spike protein) of Covid-19.

Part of SCLPs Covidity vaccine contain DNA which codes for particular epitopes taken from this spike protein. The other part contains epitopes which target the nucleocapsid protein. An epitope is a patch on a protein that the immune system sees and generates an immune response against - though not all parts of a protein generate an immune response.

An analogy could be, imagine the full length spike protein is your left arm. Oxford vaccine contains many copies of this protein (i.e. many left arms). The immune system will recognise epitopes on the surface of that protein, hence will raise an immune response (antibodies and T-cells) to attack this. These epitopes could be each of the knuckles, the finger nails, the wrist joint and the elbow joint. Different people will respond differently to this style vaccine and generate an immune profile which recognises different parts of the protein. Generally you will have an immunodominant epitope (i.e. one that most people generate a response against). So most people may generate an immune response say against the thumb nail, with some lower level immune response against the knuckles, joints etc.

Scancell's Covidity encodes a number of these epitopes individually, rather than the whole protein and will have been chosen as they will be the immunodominant ones. So instead of your immune system seeing the entire arm, Covidity when transcribed would only show the immune system say the thumb nail and the elbow. Hence your immune system will only generate a response against these particular epitopes. It won't produce a response against other parts of the virus not encoded within the Covidity vaccine.

If you're infected by the virus (imagine a whole human body), your immune system will have antibodies and T-cells raised against the left arm (spike protein). People immunised with the Oxford vaccine will have an immune response targetting a number of the epitopes found on that left hand. These most likely will overlap with those epitopes covered by Covidity. If you're immunised subsequently with SCLP vaccine, you may not get a gain of response against the spike protein, but you'll get a gain of response against the nucleoprotein. Furthermore, you may benefit from the broader antibody response that the Oxford vaccine may product from their use of the full-length spike protein.

A caveat to all this. If C-19 mutates and it happens to change the epitope your immune system targets, you lose your immune response to that epitope. Hence SCLP targetting a number of epitopes and two separate proteins to prevent immune escape.

I personally think a lot of this will come down to Lindy’s vision. She Is the scientific powerhouse and brains behind a vast proportion of the scIentific output for scancell.

Many scientists are not interested in working for big pharma, it wouldn’t surprise me if the direction is one of Licensing rather than an all out sale, but I think ultimately it’ll come down to whether her driver is altruistic or intellectual. If it’s purely about curing cancer then it wouldn’t surprise me if it’s an all out sale, if they fee they have the power to take it to market alone, then that’s another option. If it’s purely about inventions and learning, then it wouldn’t surprise me if LD would rather keep SCLP developing and selling inventions.

The difference though in mutation rate between HIV and Covid-19 is the presence of a proofreading nonstructural protein 3’-to-5’ exoribonuclease. This is key to the high fidelity replication and recombination of the virus as it in a basic sense it proof reads what it has copied when replicating, ensuring that the copy it makes of itself is error free. HIV reverse transcriptase protein lacks this proofreading ability and is considerably lower than cellular RNA polymerases and far lower than other retroviral reverse transcriptases - hence when HIV replicates it makes mistakes, these mistakes result in mutations and these mutations can serendipitously confirm a selective advantage.

Exactly! If you think about it, the virus has limited reason to mutate the way it packages itself. So long as it does this efficiently, it has no reason to try and improve this. However, to be able to replicate a virus needs to be able to hijack another cell. Therefore, the ability of a virus to be able to survive hinges on its ability to get inside a cell and replicate. The S-protein is key to a virus's ability to latch onto cell surface proteins and gain entry into the cell. Thus, if it can do this better it has an evolutional advantage. Hence S-proteins are more likely to mutate.

Other proteins within the virus are also prone to mutation because they enable the protein to replicate quicker, package themselves quicker, integrate into the genome better etc. But the nucleoprotein has limited selection pressure. So long as it is able to assemble and package the viral genome safely, that's it. There is limited evolutionary pressure for the protein shell to change. Hence, why it is more likely to be a common protein across different coronaviruses and with limited diversity between strains.

Feel free to repost.

RP/AB

Something also to bear in mind. If your bodies immune response is targeting a single protein (i.e. spike protein), then you are putting a selective pressure on the virus. If a virus by chance happened to mutate that single protein which resulted in it escaping the immune response, then the virus can continue to infect, replicate and move on. Mutations are chance events, the vast majority are detrimental to the virus and so do not lead to this escape. Now, something like Scancell's Covidity, targets two separate proteins. If either one of these targets were to mutate, the immune system still has a second target as back-up and so would stop the virus infecting and replicating. The chances of both mutating at the same time in the correct place to allow immune escape, is incredibly small (read nearly 0), especially seeing as it seems Coronavirus mutates far slower than many other viruses.

We have seen this with HIV, initial drugs targeted a single HIV protein. These worked for a while, but then the HIV virus mutated that drug target site and then we saw that drug stop working. To prevent this happening, we now routinely use combination therapies - e.g. 2+ drugs which target different parts of HIV to prevent infection/virus reproduction and spread.

On point again Ivy.

The real question is what do you want from a vaccine? Do you want short term relief to stop symptoms of Covid-19 and allow the body enough time to build natural protection against this virus. Or do you want short term relief AND long term protection not just against this coronavirus but also any others which have already appeared and are still around (MERS, SARS) or any potential future coronavirus infection.

If you're wanting the former - then go for the quick and easy spike protein, antibody route.
If you want the latter - then go for something which targets both the spike and nucleocapsid protein.

First generation vaccines have tended to target the former as if this worked it would quickly prevent healthcare systems from being overloaded , but it also means that if another novel coronavirus came along we would be back to square one. Temporary fix, but doesn't address the long term problem.

Vaccines which target both S- and N-proteins have a far greater chance of providing long term immunity, not just against Covid-19 but also against other strains of coronavirus which may jump the species barrier in the future. Hence, why even if Covidity isn't first to market, they may be the tortoise in the race to provide long term immunity to the general population.

An incredibly example of good science communication. Not just with Covid but with the whole Scancell focus, explained clearly for the average person to understand. Lindy is an incredibly talented scientist.