DDW Neuroscience Showcase

DDW hosted showcase co-sponsored by ApconiX and Axosim.

During this showcase, leading industry figures will outline the areas of opportunity and challenges in the Neuroscience drug Discovery and development sector and the following questions will be covered; where is the opportunity in the neuroscience market, how to uncover the best research tools to boost the development of new treatments, how to recognize and tackle barriers in neuroscience research, what are the technologies that can help create efficiencies and get to market faster and what does the future hold for neuroscience.

Back to video resources

Transcript of the Video

00:00:08,364 –> 00:00:08,840

[Music]

00:00:08,840 –> 00:00:18,240

Hello and welcome to the DDW Neuroscience showcase. I’m Megan Thomas multimedia editor at DDW and

00:00:18,240 –> 00:00:24,320

your host for today’s event. During today’s showcase, leading industry figures will outline

00:00:24,320 –> 00:00:29,960

the areas of opportunity and challenges in the Neuroscience drug Discovery and development sector.

00:00:30,680 –> 00:00:37,360

In this showcase hosted by DDW and co-sponsored by ApconiX and Axosim, the following questions

00:00:37,360 –> 00:00:42,960

will be covered; where is the opportunity in the neuroscience market, how to uncover the

00:00:42,960 –> 00:00:48,440

best research tools to boost the development

of new treatments, how to recognize and tackle

00:00:48,440 –> 00:00:53,600

barriers in neuroscience research, what are the technologies that can help create efficiencies

00:00:53,600 –> 00:01:00,080

and get to market faster and what does the future hold for neuroscience. Throughout the

00:01:00,080 –> 00:01:05,440

event you will have the opportunity to ask the experts questions in the chat panel. If

00:01:05,440 –> 00:01:11,080

we do not get to your questions we will follow up after the event. We will now begin with an

00:01:11,080 –> 00:01:15,960

expert panel discussion looking at the use of

new drug discovery and development technology

00:01:15,960 –> 00:01:21,440

concepts and tools in neuroscience, as well as what the future looks like for neuroscience therapies.

00:01:21,440 –> 00:01:26,440

The panel consists of Professor Ruth Roberts Co-founder and Director of Safety Science at

00:01:26,440 –> 00:01:33,880

ApconiX, Dr Mark Trahern Director and chairman of Talisman Therapeutics and Janet Sasso information

00:01:33,880 –> 00:01:42,160

scientist at CAS. I’ll now hand over to DDW editor Reece Armstrong and the panelists. Thanks everyone

00:01:42,160 –> 00:01:46,760

for joining. So for this panel discussion we’ll

be covering topics such as the challenges facing

00:01:46,760 –> 00:01:52,040

neuroscience drug developers and opportunities for the future. I’d like to start with a

00:01:52,040 –> 00:01:57,400

question for you Ruth. We know the likelihood of success in neuroscience is particularly low, so

00:01:57,400 –> 00:02:02,280

how are clinical trials in neuroscience can only being impacted by adverse events such as seizures

00:02:02,280 –> 00:02:09,360

and other central nervous system related events? That’s a great question Reece and adverse events

00:02:09,360 –> 00:02:15,800

in clinical trials and neuroscience are the main reason why clinical trials are failing. So the main

00:02:15,800 –> 00:02:23,080

issue here is, we fail to detect CNS type issues in the preclinical animal studies, then we go into the

00:02:23,080 –> 00:02:31,360

clinic and drugs fail for a variety of CNS related reasons. But one of the main ones is seizure uh

00:02:31,360 –> 00:02:37,320

there’s also other issues such as sedation which of course is the inverse of seizure but generally

00:02:37,320 –> 00:02:43,560

speaking our preclinical models don’t do a very good job on predicting and detecting seizure

00:02:43,560 –> 00:02:50,560

that may happen in the clinic. Great, thank you Ruth, that ties in nicely with the next question. Mark,

00:02:50,560 –> 00:02:54,600

one of the ways developers are aiming to overcome these limitations in the field is through the use

00:02:54,600 –> 00:03:00,160

of physiologically relevant models for study. So can you talk a little bit about how induced flury

00:03:00,160 –> 00:03:07,240

poent stem cell lines are helping to advance drug discovery within neuroscience? Sure, yes, so IPSC

00:03:07,240 –> 00:03:12,800

derived neurons and other sort of neural tissue, such as you know G cells, you know

00:03:12,800 –> 00:03:17,800

they’ve been developed over a number of years now. Really to sort of say, well in some cases you can

00:03:17,800 –> 00:03:24,200

recapitulate human disease in animals in some cases you can’t and and so what we and others

00:03:24,200 –> 00:03:30,280

companies like Talisman have been focused on, is taking IPSC cells from patients with particular

00:03:30,280 –> 00:03:36,520

disease condition and then trying to recapitulate that pathology in the dish. Then you have a

00:03:36,520 –> 00:03:41,760

totally humanized system which gets around some of the differences between ourselves and other mammalian

00:03:41,760 –> 00:03:47,200

species and and we believe and others believe that you can look at certain pathologies. So, for example at

00:03:47,200 –> 00:03:53,240

Talisman, we were very interested in the interneuronal transmission of Tau as a protein and how you can

00:03:53,240 –> 00:03:58,640

mimic that in the human condition, in the disease and then try and block that or manipulate that

00:03:58,640 –> 00:04:04,400

pharmacologically. So the idea is we’re using human diseases from specific groups of patients

00:04:04,400 –> 00:04:09,600

to recapitulate that in the dish, to answer questions and obviously very relevant to this

00:04:09,600 –> 00:04:14,520

webinar, enable and move forward drug discovery protocols. Whether that’s screening for small

00:04:14,520 –> 00:04:21,560

molecules or other therapeutic modalities. Amazing, thanks Mark and Janet, just moving on to you. In

00:04:21,560 –> 00:04:27,080

more of the sort of technological approaches that we can take, how is big data helping drug discovery

00:04:27,080 –> 00:04:33,120

in the Neuroscience sector? Great question Reece. Once we adjust the challenges

00:04:33,120 –> 00:04:39,720

we see with big data, such as the volume and data harmonization and standardization of that

00:04:39,720 –> 00:04:46,360

data, we can use advanced data processes such as AI. To help us derive meaningful drug discovery

00:04:46,360 –> 00:04:52,120

insights to help speed that process and enhance drug discoveries. One of the best advantages of

00:04:52,120 –> 00:04:58,560

having so much available drug development data is its completeness. It gives us access to that whole

00:04:58,560 –> 00:05:04,680

picture including all the experimental conditions not just that numeric ending data. So predictive

00:05:04,680 –> 00:05:10,280

analytics can help to suggest target protein selection and drug molecules of interest to help

00:05:10,280 –> 00:05:16,400

drug discovery scientists narrow their list of candidate molecules and AI is also extremely

00:05:16,400 –> 00:05:22,440

helpful in the prioritization of the multitude of the resulting drug synthesis options. So it’s so

00:05:22,440 –> 00:05:28,520

many factors scientists must consider including the ease of the synthesis of the compounds, the

00:05:28,520 –> 00:05:34,400

cost of starting material, yield safety and tox, along with the selectivity of those compounds and

00:05:34,400 –> 00:05:40,160

AI models enable that prioritization of those paths, that are more likely to be successful. So

00:05:40,160 –> 00:05:45,360

that saves a tremendous amount of time, as you can imagine, while enhancing both the safety and

00:05:45,360 –> 00:05:52,320

sustainability. This year we saw an ALS therapeutic drug candidate fully discovered and developed from

00:05:52,320 –> 00:05:58,880

target identification to efficacy assessments using AI inter clinical trials here in the US.

00:05:58,880 –> 00:06:04,480

The start of that development to completion

of patient enrollment took only two years, compared

00:06:06,936 –> 00:06:12,776

to that usual decade long conventional process. So that is how a little bit on how big data can

00:06:12,776 –> 00:06:16,640

help the new drug discovery. Thanks Janet. Ruth, I just want to pick up on a point you were making

00:06:16,640 –> 00:06:21,760

earlier about the animal models. Do you currently see animal models as a barrier for neuroscience

00:06:21,760 –> 00:06:29,800

drug discovery? That’s a really interesting way to

think of it Reece. At the moment we develop new targeted

00:06:29,800 –> 00:06:38,080

therapies for neuroscience in animal models because that’s all we have. We may also use human

00:06:38,080 –> 00:06:47,640

tissues, in vitro or salines of course, but we do find that a lot of the animal models then fail to

00:06:47,640 –> 00:06:56,520

translate into the human clinical trials both in terms of efficacy and in terms of safety. So they

00:06:56,520 –> 00:07:03,600

are a barrier in a sense that they don’t necessarily translate that well but for some diseases,

00:07:03,600 –> 00:07:08,840

the animal models are the best we have. So I think the focus of all our attention including the work

00:07:08,840 –> 00:07:14,680

that Mark and Janet have outlined, is in improving our preclinical models both for efficacy and

00:07:14,680 –> 00:07:23,040

safety, by focusing on big data and by focusing on human cells and human models preclinically. Thanks

00:07:23,040 –> 00:07:28,720

Ruth, that’s a really nice way to tie it in. Mark just moving on to you again, how essential are

00:07:28,720 –> 00:07:35,360

IPSC lines for advancing genomic and proteonic studies for Neuroscience? Sure I mean you can

00:07:35,360 –> 00:07:40,120

use them for any sort of therapeutic modality but obviously you can also use them for gene

00:07:40,120 –> 00:07:46,360

therapy and other sort of techniques where you manipulate the gene at different levels. So I

00:07:46,360 –> 00:07:52,520

think they’re fairly agnostic to the therapeutic approach being used. Ideally what they do, is

00:07:52,520 –> 00:07:58,160

mimic the disease as accurately as possible and then try and predict what may well then happen in

00:07:58,160 –> 00:08:02,640

human clinical trials down the track and as we know, there’s a particularly high attrition rate

00:08:02,640 –> 00:08:07,680

in neuroscience and we sort of alluded to that earlier, than perhaps other therapeutic areas.

00:08:07,680 –> 00:08:13,240

I think by focusing on the models that are more likely to predict clinical efficacy and human

00:08:13,240 –> 00:08:17,680

clinical trials, then we can hopefully reduce that attrition rate and actually increase the

00:08:17,680 –> 00:08:22,960

productivity of our industry, and as Janet

was saying, get drugs to patients quicker, faster,

00:08:22,960 –> 00:08:27,920

with less risk and one of the things we’ve

sort of published on a couple of years ago now

00:08:27,920 –> 00:08:34,000

in nature, was using a particular measure, we call it predictive validity and using predictive

00:08:34,000 –> 00:08:40,280

validity to say how can we learn from historic data and then use that to model sort of future

00:08:40,280 –> 00:08:45,040

likelihood of success and we think that models with a higher degree of predictive validity are

00:08:45,040 –> 00:08:50,120

more likely to produce beneficial

clinical outcomes than ones that have a lower

00:08:50,120 –> 00:08:55,840

level and we sort of modeled this using

basian decision Theory. But you know, I won’t

00:08:55,840 –> 00:08:59,800

go into too much detail but it sort of overlaps a bit with some of the things I think that

00:08:59,800 –> 00:09:05,840

Janet was alluding to and Ruth mentioned earlier. Yes absolutely, it’s a great challenge for drug

00:09:05,840 –> 00:09:11,640

developers to speed of those processes and the access of historical data is so important

00:09:11,640 –> 00:09:16,640

Janet just moving on to you, how do you see the role of biomarkers evolving in personalizing

00:09:16,640 –> 00:09:23,560

treatments for neurological disorders? Yes great question. Biomarkers are quickly

00:09:23,560 –> 00:09:29,400

evolving to transform personalized medicine in the neurological conditions. They enable earlier

00:09:29,400 –> 00:09:34,960

and more accurate diagnosis. They can guide our individual treatment choices and offering

00:09:34,960 –> 00:09:42,000

those real-time monitoring of the disease and therapeutic responses. As the field advances, we

00:09:42,000 –> 00:09:47,840

will probably see the integration of multimodal biomarkers enhanced by technologies like

00:09:47,840 –> 00:09:54,080

non-invasive liquid biopsies, but also AI and this will help drive improvements in the effectiveness

00:09:54,080 –> 00:09:59,520

and the precision of the treatments, improving outcomes again for patients. Just

100

00:09:59,520 –> 00:10:05,680

last year we saw Parkinson’s disease, we saw the validation of the biomarker alpha synuclein for the

00:10:05,680 –> 00:10:11,440

first time through the seeding amplification assay and it can detect the pathology and

102

00:10:11,440 –> 00:10:16,680

spinal fluid, not only for people diagnosed with Parkinson’s disease but also individuals who are

103

00:10:16,680 –> 00:10:22,040

not yet showing those clinical symptoms, but are at the high risk for the disease. So as

104

00:10:22,040 –> 00:10:28,240

we optimize these assays for the future, we may be able to to detect alpha synuclein through blood draws,

105

00:10:28,240 –> 00:10:35,080

the skin biopsy or possibly nasal swabs. So we’re going to improve access and the early diagnosis

106

00:10:35,080 –> 00:10:41,200

and the advance that may soon be used to develop better diagnosis but more importantly that rapid

107

00:10:41,200 –> 00:10:46,040

accelerate the search for the treatment of the diseases, and even speeding up that clinical

108

00:10:46,040 –> 00:10:52,160

trial process that Ruth talked about. Great thank you Janet. So for this next portion, I’d just like

109

00:10:52,160 –> 00:10:57,360

to open the questions to everyone now and I think we’ll start with bit more a general question

110

00:10:57,360 –> 00:11:01,680

What are some of the biggest challenges

currently facing neuroscience drug developers and

00:11:01,680 –> 00:11:08,200

obviously, we can pick up points already mentioned but Janet, do you want to stop us off. Yes, so I’ll

112

00:11:08,200 –> 00:11:13,120

take it back to the big picture here. One of the biggest challenges I see with my work here at

113

00:11:13,120 –> 00:11:18,400

CAS which is a non-for-profit division of the American Chemical Society and we specialize in

114

00:11:18,400 –> 00:11:23,840

scientific knowledge management. It’s just how drug developers again are tackling the

115

00:11:23,840 –> 00:11:29,480

large amounts of data available to them. We all know that big data for drug development

116

00:11:29,480 –> 00:11:35,440

is multifaceted and from different sources, so drug development involves complex processes

117

00:11:35,440 –> 00:11:41,360

that generate data from genomics, proteomics imaging data, clinical trials again and then

118

00:11:41,360 –> 00:11:47,200

even electronical medical records. Of course, all of these different sources come in different

119

00:11:47,200 –> 00:11:54,200

data formats so harmonization and standardization of that data is paramount to be able to interpret

120

00:11:54,200 –> 00:12:01,200

and gain those insights. So using AI including machine learning, complex algorithms and

00:12:01,200 –> 00:12:08,280

large language models, that can help get that specialized data kind of formatted and in

122

00:12:08,280 –> 00:12:15,240

the right format for you to gain expertise. So big data holds great promise for improving and

123

00:12:15,240 –> 00:12:20,760

reducing the time of drug development but first we have to tackle each one of those challenges one

124

00:12:20,760 –> 00:12:27,320

by one. Yes and if I can build on what Janet said, I I think we are taking this back to big

125

00:12:27,320 –> 00:12:34,320

picture and you explain that very well Janet, about big data and how we can use it to assist in our decisions.

126

00:12:34,320 –> 00:12:40,640

But looking at the last 30, 40 or 50 years of neuroscience and drug discovery and development,

127

00:12:40,640 –> 00:12:46,840

I think our major challenge has been translation, which is an over dependence on animal models to

128

00:12:46,840 –> 00:12:54,320

predict both efficacy and to predict safety. So taking a step back from that, we’ve been working

129

00:12:54,320 –> 00:13:01,440

on an invitro seizure liability assay which uses human ion channels and human induced pluripotent

130

00:13:01,440 –> 00:13:09,320

stem cell neurons, to detect and predict seizure. Both to detect it in lead optimization and then to

00:13:09,320 –> 00:13:15,360

design that out but then also to provide answers to people who have seen seizures in preclinical

132

00:13:15,360 –> 00:13:21,360

studies in animals and then also in the clinic. So really, we’re not trying to replace any

133

00:13:21,360 –> 00:13:26,600

particular animal study per se, we’re trying to redesign the paradigm where we say ‘how do we

134

00:13:26,600 –> 00:13:32,280

best understand what’s happening in humans?’ I’m very excited about the work that Mark and

135

00:13:32,280 –> 00:13:38,000

his team are doing because it also enables us to look human induced pluripotent stem cell

136

00:13:38,000 –> 00:13:45,280

neurones from patients with epilepsy, for example, to help us understand the whole model and perhaps

137

00:13:45,280 –> 00:13:51,560

even redesign our iSLA assay for discovery of new anti-epileptics. I think there’s so much

138

00:13:51,560 –> 00:13:56,600

exciting work happening at the moment.

I think so and to add to that, I think both very good

139

00:13:56,600 –> 00:14:01,200

points that Ruth and Janet have just mentioned there and I mean, I think you know it comes back

140

00:14:01,200 –> 00:14:05,600

to this attrition point, in neuroscience a lot of things don’t translate for the reasons we’ve just

00:14:05,600 –> 00:14:10,800

discussed and we need better assays, better models and we would say those models would have greater

142

00:14:10,800 –> 00:14:17,080

predictive validity and therefore reduce the the likelihood of failure. I think you know

143

00:14:17,080 –> 00:14:21,040

again, I think Janet’s point about big data or whatever you want to call it, you know you often

144

00:14:21,040 –> 00:14:26,640

get very complicated readouts from these models right, it can be very complicated imaging. It can be

145

00:14:26,640 –> 00:14:31,680

electrophysiology as we’re talking about epilepsy which very complicated traces, you know and the

146

00:14:31,680 –> 00:14:37,920

ability to analyze those quickly and efficiently,

so you can you know much quicker and faster and

147

00:14:37,920 –> 00:14:42,960

more accurately then the human mind can do, I think is a huge advance and I think one of the things

148

00:14:42,960 –> 00:14:48,360

for sort of chronic neurogenerative diseases is, they happen slowly they often take 20 years, 50

149

00:14:48,360 –> 00:14:54,720

years, 60 years, 80 years. How can you compress that pathology into a dish or even into an animal model

150

00:14:54,720 –> 00:14:59,280

or whatever it is, so that you can actually produce that in a timely fashion and I think that’s one of

00:14:59,280 –> 00:15:05,520

the sort of unique features of neurodegeneration, is that the disease may take a long long time to

152

00:15:05,520 –> 00:15:11,480

evolve in or develop, I would say, in the

human. If you can compress that into, say an invitro

153

00:15:11,480 –> 00:15:17,840

model, where you can see that, in say a month or so, that allows you to then look at a chronic

154

00:15:17,840 –> 00:15:23,320

condition in a relatively short time scale,

which is compatible with you know accelerating

155

00:15:23,320 –> 00:15:28,760

the whole drug discovery process for example. Great, thanks everyone, really interesting points there.

156

00:15:28,760 –> 00:15:33,520

I think we’ve got time just for one last

question. What are you most excited about for a

157

00:15:33,520 –> 00:15:38,800

future of neuroscience? Janet, do you want

to start this one off again? Yes definitely, this

158

00:15:38,800 –> 00:15:44,800

is a great question. I am most excited to watch the evolution of research over time like we have

159

00:15:44,800 –> 00:15:50,200

just talked about right now. With my work, I get to see the bird’s eye view advantage to see

160

00:15:50,200 –> 00:15:55,840

the field broadly and, 100%, I’m most excited about the discoveries and the cures that are

00:15:55,840 –> 00:16:00,363

to come. From data collection to drug development, including target and

162

00:16:00,363 –> 00:16:07,880

lead compound selection, to advanced assay techniques that we spoke on and even lab automation, I can’t wait to see how these

163

00:16:07,880 –> 00:16:13,560

processes evolve and help speed drug discovery. I’m truly excited about that big picture

164

00:16:13,560 –> 00:16:19,320

and seeing more disease modifying drugs make it into the hands of patients. Yes, I thinking along

165

00:16:19,320 –> 00:16:24,640

the same lines. I think some of the work you’ve described at CAS Janet, where we can use big

166

00:16:24,640 –> 00:16:30,560

data to look for new drug targets and there’s

the really interesting and exciting developments

167

00:16:30,560 –> 00:16:35,960

around targeted protein degraders, where we can target previously un-targetable targets,

168

00:16:35,960 –> 00:16:41,880

if that’s not saying the same thing three times. That, I think is super exciting and then one of

169

00:16:41,880 –> 00:16:48,400

the projects that we’re just about to release, which is a little left field, is we’ve been able to

170

00:16:48,400 –> 00:16:55,440

take the 16 track recording from MEA from our new human neurons and make music from it. Now I know

00:16:55,440 –> 00:16:59,920

the science of that is super interesting but it’s

a really great way for public engagement because

172

00:16:59,920 –> 00:17:04,800

the public are super curious about neuroscience and the more we can engage them, the more we can

173

00:17:04,800 –> 00:17:10,680

assist our field with Patient Group engagements, with charitable funding. So targeted protein

174

00:17:10,680 –> 00:17:17,280

degraders and singing neurons are the things I’m thinking about right now. I think that’s great

175

00:17:17,280 –> 00:17:20,760

and just to sort of add to that, I think it’s also trying to get those therapies into the

176

00:17:20,760 –> 00:17:27,080

right patient, at the right time. You know it’s a big challenge and obviously we are all fantastically

177

00:17:27,080 –> 00:17:32,360

neuro diverse and the big problems over many years of, you know being involved with clinical trials in

178

00:17:32,360 –> 00:17:37,440

neurosciences, we tend to put the drugs in the wrong patients and find they don’t work. If we

179

00:17:37,440 –> 00:17:42,840

can actually get the same quality of data and the same, you know we would call predictive

180

00:17:42,840 –> 00:17:49,080

validity of data in new models, that means we can then do faster and more efficient clinical trials

00:17:49,080 –> 00:17:53,840

which are going into patients that are more likely to respond and therefore, you know the

182

00:17:53,840 –> 00:17:59,000

whole costly clinical trials process, by putting some investment up front, can be much more

183

00:17:59,000 –> 00:18:03,160

efficient and actually you’re not giving drugs in clinical trials, to people even though they’re

184

00:18:03,160 –> 00:18:06,960

volunteers, you’re not putting drugs into patients where they’re less likely to work. So I think

185

00:18:06,960 –> 00:18:13,240

there’s a very strong ethical component as well. Amazing, thanks Mark. Unfortunately that does

186

00:18:13,240 –> 00:18:17,960

bring us to the end of the panel discussion. I’d like to thank Ruth, Janet and Mark once more for

187

00:18:17,960 –> 00:18:25,800

joining us and thank you everyone for listening.

Our pleasure thank you, thank you. Thanks very much

188

00:18:25,800 –> 00:18:32,200

for that fascinating panel discussion. Today’s showcase is co-sponsored by ApconiX and Axosim

189

00:18:32,200 –> 00:18:38,840

and we will now hear from ApconiX’s Ruth Roberts for a product presentation. Thank you for the kind

190

00:18:38,840 –> 00:18:45,640

introduction and today I’m going to be talking to you about seizure and epilepsy with a focus on new

00:18:45,640 –> 00:18:52,760

opportunities for detection and treatment. First of all, let’s frame the problem we’re addressing

192

00:18:52,760 –> 00:18:58,840

and that is that safety is the main reason why drugs fail in discovery and development. As we

193

00:18:58,840 –> 00:19:07,320

see in this graph over on the left hand side, in preclinical phases 82% of drugs that fail do so

194

00:19:07,320 –> 00:19:14,080

because of safety reasons and this failure due to safety continues through phase one and into phase

195

00:19:14,080 –> 00:19:22,280

two. It’s an issue we must address. Looking at this data in more detail, one thing that stands out is

196

00:19:22,280 –> 00:19:28,720

that when we look at preclinical, in orange red, and clinical failures, in a beige brown colour, here

197

00:19:29,400 –> 00:19:37,640

we can see that damage to the CNS or CNS toxicity is the main reason why drugs fail in the clinic

198

00:19:37,640 –> 00:19:43,520

and looking at those data in more detail, seizure is a big issue and one of the main reasons for

199

00:19:43,520 –> 00:19:51,480

these failures. So current approaches to screening for seizure don’t work. So we can see here an

200

00:19:51,480 –> 00:19:57,960

overview, generalized overview of drug discovery and development. So in discovery lead we look at

00:19:57,960 –> 00:20:03,640

lead generation and optimization and then we go into the GLP toxicology phase, then we enter first

202

00:20:03,640 –> 00:20:10,040

time in humans in clinical trials. During the GLP toxicology phase, we undergo standard regulatory

203

00:20:10,040 –> 00:20:16,320

toxicology testing usually in two species for small molecules and during that phase is where

204

00:20:16,320 –> 00:20:22,680

we see seizures in animals preclinically and that obviously stops the program because we don’t know

205

00:20:22,680 –> 00:20:29,520

how that translates to humans. Even more drastic, is when we see seizures in the clinic as we saw in

206

00:20:29,520 –> 00:20:35,720

the stats in the previous slide, so our current approach for screening for seizures doesn’t work.

207

00:20:35,720 –> 00:20:42,160

Seizures are hard to detect, methodologies are low throughput, expensive and have poor translation to

208

00:20:42,160 –> 00:20:50,600

humans. Back in 2009 Alison Easter et al proposed that we could look at seizures in rat brain slices

209

00:20:50,600 –> 00:20:57,400

and in zebra fish and indeed those screens have been introduced but still the problem persists

210

00:20:57,400 –> 00:21:02,760

and we have expensive follow-up confirmation

in rodent behavior and EEG tests. So the question

00:21:02,760 –> 00:21:09,360

is ‘can we improve preclinical seizure screening?’ So what I’m going to tell you about today in more

212

00:21:09,360 –> 00:21:16,120

detail is our invitro seizure liability assay or

iSLA, where we’ve developed a human ion channel

213

00:21:16,120 –> 00:21:22,720

panel and human neurons in micro electrode array, to replace the rat brain slices and zebra fish for

214

00:21:22,720 –> 00:21:30,040

higher throughput and improved human translation, to address this issue of screening for seizure.

215

00:21:30,040 –> 00:21:33,680

So that’s what we’re going to hear about today and let me just fill you in on a little bit of

216

00:21:33,680 –> 00:21:40,800

the detail that supports the use of iSLA. So first

of all we looked at micro electrode array (MEA) for

217

00:21:40,800 –> 00:21:48,280

seizure detection in human neurons. So we put human neuronal cells into micro electrode array shown

218

00:21:48,280 –> 00:21:55,560

here and that each well has got 16 electrodes at the base. That then detects the signals from the

219

00:21:55,560 –> 00:22:01,200

cells in the wells and you can see here initially at day 5 the cells are not connected and they’re

220

00:22:01,200 –> 00:22:07,880

firing in their own way. But by day 23 the cells have coordinated their firing response and we see

00:22:07,880 –> 00:22:16,240

a nice integrated human neuronal network in the plate. So we can look at human neurons, we can also

222

00:22:16,240 –> 00:22:22,160

look at ion channels associated with seizure. So we started with hundreds of ion channels based

223

00:22:22,160 –> 00:22:28,880

on human mutations associated with epilepsy and also pharmacology and narrowed it down to table

224

00:22:28,880 –> 00:22:34,440

one here, where we see the ion channels, we have shortlisted the 15 seizure related ion channels

225

00:22:34,440 –> 00:22:41,720

that are in iSLA that are part of the panel. We also selected compounds to test and these are

226

00:22:41,720 –> 00:22:47,960

everyone’s favourite seizurogenic compounds, either in humans or in preclinical animal species and we

227

00:22:47,960 –> 00:22:54,920

tested these a range of concentrations based on the human toxic dose and/or therapeutic dose and

228

00:22:54,920 –> 00:23:02,560

the rat toxic dose where we saw seizures. To take all of those data today I wouldn’t have time to

229

00:23:02,560 –> 00:23:08,240

go through them all but I’d refer you to this paper, where we present the results that show that we can

230

00:23:08,240 –> 00:23:14,640

detect seizurogenic compounds in iSLA which is an integrated seizure liability assay using human

00:23:14,640 –> 00:23:20,040

neurons and human ion channels and it’s very very effective. We have some interesting case

232

00:23:20,040 –> 00:23:27,640

studies to share as well and meetings following

on from this. So we were very excited when iSLA, the

233

00:23:27,640 –> 00:23:34,880

human seizure liability assay was quoted in a recent paper from FDA CDER, shown here on the right hand

234

00:23:34,880 –> 00:23:41,560

side, where Avira et al, referenced our work saying that this battery of invitro ion channel functional

235

00:23:41,560 –> 00:23:48,000

assays can be used to predict invivo human CMAX levels for small molecule drug associated

236

00:23:48,000 –> 00:23:54,120

with seizure, giving a clear context of use of COU for assay. So that was very exciting, we

237

00:23:54,120 –> 00:24:01,840

were very happy to see that. So iSLA is been very effective in what we developed it for, which is

238

00:24:01,840 –> 00:24:08,480

to screen pro seizurogenic compounds, during lead optimization, to optimize lead series or indeed

239

00:24:08,480 –> 00:24:14,680

for problem solving when seizures are seen in preclinical species or even in the clinic. But

240

00:24:14,680 –> 00:24:21,000

the development of iSLA presents us with additional opportunities and what I’m going to talk about in

00:24:21,000 –> 00:24:27,200

the last part of this talk is, new therapeutics for epilepsy and to address this, I want to acknowledge

242

00:24:27,200 –> 00:24:33,840

Rajiv Mohanraja who is a Consultant neurologist in Manchester and part of the Manchester Epilepsy

243

00:24:33,840 –> 00:24:40,080

Research Network. I spoke to him about our ion channel panels and he said this was a fabulous

244

00:24:40,080 –> 00:24:48,400

opportunity to develop new therapeutics for epilepsy. So there’s a large proportion of patients

245

00:24:48,400 –> 00:24:54,440

that have drug resistant epilepsy that do not have control of their seizures with currently available

246

00:24:54,440 –> 00:25:01,120

medications, leading to all these terrible consequences such as depression, premature death

247

00:25:01,120 –> 00:25:08,600

and cognitive impairment. So uncontrolled seizures a terrible impact on quality of life. If we look at

248

00:25:08,600 –> 00:25:14,920

the mechanisms of seizure, you can see here that it involves sodium channel signaling, potassium

249

00:25:14,920 –> 00:25:21,080

channel signaling and then calcium channel signaling, through glutamate and gaba, to either

250

00:25:21,080 –> 00:25:28,600

excitatory or inhibitory neurons in post-synaptic space. So as you can see from this simple diagram,

00:25:28,600 –> 00:25:35,160

seizures and epilepsy are dependent upon ion channel signaling just like seizures induced

252

00:25:35,160 –> 00:25:41,920

in inadvertently in preclinical drug discovery and development. So the question now, the opportunity we

253

00:25:41,920 –> 00:25:50,560

have now, is ‘can we use our ion channel panel and our MEA to develop new therapeutics for epilepsy?’

254

00:25:50,560 –> 00:25:58,040

So the opportunity for new therapeutic approaches, when I talk to Rajiv, the reason we haven’t seen

255

00:25:58,040 –> 00:26:04,040

many new therapeutics coming through in epilepsy, is that there is a continued focus on animal

256

00:26:04,040 –> 00:26:10,960

models for developing new anti-epileptics. There are new treatments coming through for epilepsy

257

00:26:10,960 –> 00:26:15,920

but they seem to be targeting the same mechanism of action, therefore have the same failure rates

258

00:26:15,920 –> 00:26:22,480

and the same refractory population as existing therapies. So our opportunity is to adapt the

259

00:26:22,480 –> 00:26:28,480

ApconiX MEA ion channel panel to look for new anti-epileptics based on human ion channels

260

00:26:28,480 –> 00:26:35,840

and human neurons. What we’re currently doing is to induce seizurogenic phenotype in human neurons at MEA

00:26:35,840 –> 00:26:40,800

and evaluate if current therapeutics can intervene and that would be a proof of concept for the

262

00:26:40,800 –> 00:26:47,280

essay and then how would we induce a seizurogenic phenotype? What’s the best way to do that and which

263

00:26:47,280 –> 00:26:53,040

therapeutics would be a gold standard to test? Personally, as a preclinical safety specialist and

264

00:26:53,040 –> 00:26:58,240

toxicologist, I’m a little bit out of my depth with this but I can see a fabulous opportunity here.

265

00:26:58,760 –> 00:27:04,480

So we’re currently taking this out into the wide world for discussion and I’m leading a session

266

00:27:04,480 –> 00:27:09,840

at the Society for Neuroscience in Chicago, where we put a session together to discuss

267

00:27:09,840 –> 00:27:15,120

the idea of new opportunities for detection and treatment of seizure and epilepsy and we have

268

00:27:15,120 –> 00:27:19,560

three clinicians including Rajiv speaking in the session. Where we’re going to really

269

00:27:19,560 –> 00:27:25,120

explore the concept of whether this assay can be used for looking for new anti-epileptics. So

270

00:27:25,120 –> 00:27:30,640

this is going to be really interesting, getting global International input and expertise into

00:27:30,640 –> 00:27:36,880

this idea and help us shape the next steps and

the next experiments that we’ll do at ApconiX. So

272

00:27:36,880 –> 00:27:42,800

to conclude, seizures remain an issue in both preclinical and clinical development for new

273

00:27:42,800 –> 00:27:51,640

drugs. iSLA uses human pluripotent stem cell neurons in microelectrode array, plus a human

274

00:27:51,640 –> 00:27:58,320

ion channel panel to detect seizures, both for lead optimization and also for understanding

275

00:27:58,320 –> 00:28:04,720

the mechanism of seizures when they’re seen preclinically and clinically. As well as this use,

276

00:28:04,720 –> 00:28:10,600

we’re currently working on adapting iSLA to search for new anti-epileptics and we will be exploring

277

00:28:10,600 –> 00:28:15,680

that concept at the Society of Neuroscience with international experts. Finally, we’re

278

00:28:15,680 –> 00:28:22,440

really excited with our new project called tubular cells, where we’ve taken the 16 channel electrical

279

00:28:22,440 –> 00:28:28,760

output from our MEA cells, shown bottom left and used it to make music with the input of

280

00:28:28,760 –> 00:28:34,480

a very talented Loop artist musician Michael Sebastian and Michael’s used the output of the

00:28:34,480 –> 00:28:41,080

16 channel MEA to make a very beautiful piece of music, which you can hear on our website by

282

00:28:41,080 –> 00:28:47,320

scanning the QR code, to hear all about the tubular cell’s project and hear the music. With that I’ll

283

00:28:47,320 –> 00:28:53,640

stop and thank you for your attention. Thank you Ruth. Before we move on to the next presentation

284

00:28:53,640 –> 00:28:58,800

of the summit, I’d just like to remind everyone that this event will be available to access on

285

00:28:58,800 –> 00:29:07,960

demand on the DDW website www.ddw-online.com. Next we will hear from Dr Sheldon Preskorn ,

286

00:29:07,960 –> 00:29:13,160

whose presentation highlights the accelerating evolution of psychiatric treatment development,

287

00:29:13,160 –> 00:29:19,520

from one dependent on chance observations to more sophisticated and less risky molecular biology

288

00:29:19,520 –> 00:29:27,840

approaches. Over to you Sheldon. Well thank you for introducing me and I’m Sheldon Preskorn. I’ll

289

00:29:27,840 –> 00:29:36,800

be talking about neuroscience showcase, from chance discovery to molecular biology in psychiatric and

290

00:29:36,800 –> 00:29:44,320

neurological drug development. I’m going to be going through quite a set of slides so I’ll go

00:29:44,320 –> 00:29:51,480

relatively quickly because you can read the slides faster than I can speak. I’m highlighting a

292

00:29:51,480 –> 00:29:58,040

few things I did an undergraduate thesis in, what today would be called neurobiology,

293

00:29:59,280 –> 00:30:08,160

in the departments of psychology, chemistry and biology. I tended medical school, did an anatomical

294

00:30:08,160 –> 00:30:15,400

pathology fellowship, completed my rotating internship in psychiatry residency and then

295

00:30:15,400 –> 00:30:21,480

received a National Institute of Mental Health research scientist development award. That award

296

00:30:21,480 –> 00:30:29,600

was used to be in the department of clinical pharmacology at Kansas University School of

297

00:30:29,600 –> 00:30:38,040

Medicine, where I did mainly basic wet lab research and then moved into drug development from phase

298

00:30:38,040 –> 00:30:45,040

one onward. So my career has sort of spanned all aspects of drug development research from

299

00:30:45,040 –> 00:30:54,320

preclinical pharmacology, which was those early years, into phase one and through phase three.

300

00:30:54,320 –> 00:31:00,000

I actually have to update this slide, I’ve

actually worked on seventeen successful new drug

00:31:00,000 –> 00:31:09,880

applications. This is a nice quote from Emal Craitlin and basically what he is saying is,

302

00:31:09,880 –> 00:31:15,480

that to understand psychiatric illnesses and neurological illnesses, we have to

303

00:31:15,480 –> 00:31:22,120

have a connection between knowledge of physical changes in the brain and the mental symptoms

304

00:31:22,120 –> 00:31:28,240

associated with them. He also indicated that one could learn from the effects of drugs,

305

00:31:28,240 –> 00:31:37,360

something about the true nature of the symptom

so you could work either way. Now this shows you

306

00:31:37,360 –> 00:31:45,680

the hierarchial level of diagnosis in all aspects of medicine, this is not restricted to CNS drug

307

00:31:45,680 –> 00:31:53,880

development or medicine but all. It begins with the simplest one on base and that is symptom.

308

00:31:53,880 –> 00:32:00,520

Then you can move to syndromic diagnosis, such as, migraine, headaches or major depressive

309

00:32:00,520 –> 00:32:08,640

disorder and then you can move further upward as knowledge increases to pathophysiology and

310

00:32:08,640 –> 00:32:15,760

the height is actually ideological diagnosis. Psychiatry has been, until recently, stuck at

00:32:15,760 –> 00:32:21,040

the level of syndromic diagnosis and we’ll explain why that’s a problem and why that’s

312

00:32:21,040 –> 00:32:31,360

being addressed. So syndromic diagnoses are based on a cluster of signs and symptoms and that is

313

00:32:31,360 –> 00:32:37,680

imprecise, that’s the reason why we want to go to pathophysiology and pathology because it’s

314

00:32:37,680 –> 00:32:46,600

more specific. But having started at the syndromic level and having not known much about the brain,

315

00:32:46,600 –> 00:32:52,920

we were dependent in the early years on chance observation, such as the therapeutic

316

00:32:52,920 –> 00:33:00,600

benefit of chlorpromazine. Then once we had a molecule, we could modify it and that

317

00:33:00,600 –> 00:33:08,240

would be chemistry and thioridazine. Then we could move to understanding the biology that the drug

318

00:33:08,240 –> 00:33:14,840

was working on and that led to halaperidone. Then finally molecular biology and that

319

00:33:14,840 –> 00:33:23,400

leads to drugs such as aripiprazole. So again, another way of looking at this, we started with chance. We

320

00:33:23,400 –> 00:33:29,000

actually use the drugs that were discovered by chance to understand classic neurotransmitters.

00:33:29,000 –> 00:33:35,600

Now we’ve moved on to neuropeptides, growth factors and the other ones that you see listed on the

322

00:33:35,600 –> 00:33:45,120

slide. So this gives you sort of a milestone in psychiatric drug discovery. It shows what

323

00:33:45,120 –> 00:33:52,560

happened in the 1940s which were all chance discovery drugs, moving into the 1950s

324

00:33:52,560 –> 00:33:58,200

still chance and that’s the MAOts inhibitors and the tricyclic anti-depressants

325

00:33:58,200 –> 00:34:06,120

and in the 1960s, we took advantage of what Emal Criplin said and actually used the drugs to begin

326

00:34:06,120 –> 00:34:13,440

to understand mechanisms of action of those drugs developed animal models and so on. In

327

00:34:13,440 –> 00:34:22,640

the 1970s, we were able to develop more highly potent anti psychotic medications and we could

328

00:34:22,640 –> 00:34:29,760

synthesize serotonin selective reuptake inhibitors. In the 80s, they were marketed and in

329

00:34:29,760 –> 00:34:39,240

the 90s newer antis psychotics. So this just shows, you in the 1960s we were able to and and

330

00:34:39,240 –> 00:34:47,920

certainly then into the 1970s, understand various mechanisms of drug action and those would be

00:34:47,920 –> 00:34:56,360

on the bottom left uptake pumps, neurotransmitter uptake pumps or transporters and receptors.

332

00:34:56,360 –> 00:35:05,200

You can see that Imipramine, which was actually a analog of chlorpromazine, hits a lot of targets

333

00:35:05,200 –> 00:35:13,080

at almost the same concentration. On the x axis is the binding affinity of the drug and what you

334

00:35:13,080 –> 00:35:20,760

see is that with Imipramine, you had many targets that were affected at exactly the same concentration.

335

00:35:20,760 –> 00:35:28,040

Only some of those might actually be relevant to their anti-depressant properties and the candidate

336

00:35:28,040 –> 00:35:35,520

target, that was chosen, was the serotonin transporter and you can see here that one

337

00:35:35,520 –> 00:35:44,800

was able to modify the structures to only in terms of the receptor shown on this slide. The serotonin

338

00:35:44,800 –> 00:35:51,920

transporter and to have wide separation between that binding affinity and all other targets. That

339

00:35:51,920 –> 00:35:58,000

means that you limited the pharmacology of the drug only to the desired target that you

340

00:35:58,000 –> 00:36:06,480

needed to treat major depressive disorder. So returning to the levels of diagnosis recently,

00:36:06,480 –> 00:36:12,760

well perhaps over the last 10 to 15 years,

certainly slowly at the beginning and growing now

342

00:36:12,760 –> 00:36:20,680

in speed, we’ve moved to pathophysiology and that is basically understanding the circuits underlines

343

00:36:20,680 –> 00:36:27,560

specific psychiatric conditions. So this just shows you five movement disorders at the top,

344

00:36:28,160 –> 00:36:36,240

then memory disorders, OCD addictive disorders and depression, just as examples the circuits

345

00:36:36,240 –> 00:36:44,960

and the brain structures involved and molecular targets within those structures. So with this

346

00:36:44,960 –> 00:36:53,800

approach, six novel CNS drugs were approved, actually in the 2000s to 2010 and this shows

347

00:36:53,800 –> 00:37:02,680

you the indication, it shows you the drug and

so for sleep, Ramelteon and Suvorexant and going down,

348

00:37:02,680 –> 00:37:08,440

smoking sensation Varenecline. I’m going to focus on that one a little bit more, but we were

349

00:37:08,440 –> 00:37:17,080

able to have specific biological processes which is the indication and specific drugs developed

350

00:37:17,080 –> 00:37:25,360

for those biological processes. Now these were all based on their effectiveness on a single symptom

00:37:25,360 –> 00:37:32,280

that is whether it was sleep, appetite etc and

the symptoms or behaviors could be expressed in

352

00:37:32,280 –> 00:37:39,200

a clinically meaningful binary manner, either words yes or no, the circuitry underlined the behavior

353

00:37:39,200 –> 00:37:46,000

was understood and the circuits were relatively simple that allowed for understanding target

354

00:37:46,000 –> 00:37:55,040

engagement and there was highly predictive power between animal models and man. Here gives you

355

00:37:55,040 –> 00:38:01,440

just an expanded view of what’s circuits mean. This is a little complicated but I’m going to

356

00:38:01,440 –> 00:38:09,320

focus basically on two specific areas in the brain and the first starting from your left,

357

00:38:09,320 –> 00:38:16,880

is the nucleus accumbens and this is the reward center in the brain and then moving over to the

358

00:38:16,880 –> 00:38:25,240

further right, you see the VTA, that’s the ventral tegmentum area and that contains dopamine neurons.

359

00:38:25,240 –> 00:38:30,640

Dopamine neurons project into the nucleus accumbens which is again the reward center in

360

00:38:30,640 –> 00:38:39,640

the brain and this area, this nucleus accumbens, appears to be a common pathway for all drugs

00:38:39,640 –> 00:38:47,720

of abuse. So you see opiates up there, you see alcohol, you see cannabinoids, you see psychos

362

00:38:47,720 –> 00:38:57,040

stimulants and nicotine. So nicotine affects the VTA to release dopamine into the nucleus

363

00:38:57,040 –> 00:39:04,360

accumbens and that is rewarding and that’s the reason why nicotine can become an abusable drug.

364

00:39:04,360 –> 00:39:10,960

Understanding that and understanding the receptor that causes the dopamine to be released, which is

365

00:39:10,960 –> 00:39:17,800

the alpha4 beta 2 nicotinic receptor, you could develop a drug specifically for that which was

366

00:39:17,800 –> 00:39:25,720

veronin. This simply shows you that by understanding the neurobiology of sleep. I realize

367

00:39:25,720 –> 00:39:32,640

this is a busy slide but this way you can actually go back and read it at your leisure if you wish

368

00:39:32,640 –> 00:39:41,160

and the reference is on the bottom left. But there are now seven different mechanistic classes of

369

00:39:41,160 –> 00:39:49,400

medicines to aid in sleep. So reading the slide over from the top, you have the medication class,

370

00:39:49,400 –> 00:39:57,520

you have the target transmitter, you have the nucleus within the brain that is involved

00:39:57,520 –> 00:40:05,600

in sleep physiology and then you have examples generic and brand names. So there are

372

00:40:05,600 –> 00:40:12,800

at least seven different areas in the brain that regulate sleep and these specific

373

00:40:12,800 –> 00:40:20,120

drugs, specific mechanism drugs, which are the antihistamines, the benzodiazapines, the so-called

374

00:40:20,120 –> 00:40:30,400

Z drugs, 5 ht2a antagonist, the melatonin receptors one and two and the DORAs. I’m going to focus

375

00:40:30,400 –> 00:40:39,640

on the DORAs and explain how they came to be. So around the year 2000, we did not know that

376

00:40:39,640 –> 00:40:48,200

orexin existed in the brain, but by studies of canine dogs with narcolepsy, one could actually

377

00:40:48,200 –> 00:40:55,560

determine that the animals that had narcolepsy, the litter mates that had narcolepsy, had an

378

00:40:55,560 –> 00:41:05,600

abnormality in one or two sites or both and that was two genes; hypocretin one

379

00:41:05,600 –> 00:41:14,880

or hypocretin 2, also now called orexin. So orexin and hypocretin are the same thing and so one could

380

00:41:14,880 –> 00:41:25,240

determine that orexin one receptors orexin 2 and the combination were important in terms of the

00:41:25,240 –> 00:41:31,920

narcolepsy in dogs. It turns out that in man, it’s not these receptors but it’s actually the loss

382

00:41:31,920 –> 00:41:41,840

of a orexin nuclei. So from the year 2000, we were able to map where the receptors were, then we could

383

00:41:41,840 –> 00:41:50,240

move back to where the nucleus was and find that it is in the lateral area of the hypothalamus. Once

384

00:41:50,240 –> 00:41:57,360

we understood the amino acid sequence for the receptor, we could develop drugs specifically to

385

00:41:57,360 –> 00:42:04,440

affect those receptors and that’s how we got the

DORA drugs. Now just showing where we’re going into

386

00:42:04,440 –> 00:42:11,000

the future and by the way, all this information is in the public domain from the company’s website.

387

00:42:11,000 –> 00:42:17,920

But this shows you pharma companies and mechanistic targets and so these are targets

388

00:42:17,920 –> 00:42:23,400

within those circuits and I’ll just go through

a couple of them. There’s insulin growth factor

389

00:42:23,400 –> 00:42:33,120

one, there’s the insulin growth factor binding protein three, there’s the KB7 family of voltage

390

00:42:33,120 –> 00:42:40,880

gated potassium channels and down under neumora, the kappa opioid receptors. So we can identify all

00:42:40,880 –> 00:42:49,000

of these by the way the kappa opioid receptors are involved with the nucleus accumbens. Again all this

392

00:42:49,000 –> 00:42:55,360

information is in the public domain but then we can have specific target illnesses. Fragile

393

00:42:55,360 –> 00:43:03,520

X, mood disorders etc, many of these are actually genetic disorders genetically determined. So you

394

00:43:03,520 –> 00:43:11,360

can see now that we have moved from syndromic diagnosis to pathophysiological and in the case

395

00:43:11,360 –> 00:43:19,560

of genetic disorders, actually to ideological and with this I’ve given you a short overview of

396

00:43:19,560 –> 00:43:28,120

CNS drug development, from chance observation to very specific molecular biology. Thank you very

397

00:43:28,120 –> 00:43:37,040

much. Thanks for that great presentation Sheldon. I’ve now got some audience questions to ask you.

398

00:43:37,040 –> 00:43:45,680

The first is, has genetics led to any success in Psychiatry? Yes, as I just said before,

399

00:43:45,680 –> 00:43:53,000

we can understand the genetics for example of fragile X and understand how that

400

00:43:53,000 –> 00:44:03,040

leads to the fragile X syndrome. We can also understand that mothers who have sons with

00:44:03,040 –> 00:44:10,720

fragile X, they themselves do not have fragile X because they have two x chromosomes but

402

00:44:10,720 –> 00:44:19,800

they do have subtle, in many instances, subtle behavioral and psychiatric

403

00:44:19,800 –> 00:44:27,280

symptoms, not anywhere to the extent of their sons who have fragile X. So genetics is explaining

404

00:44:27,280 –> 00:44:43,640

even how there is relationship between

a clear-cut syndrome, only in males or mostly in males, because males only have one X chromosome but also then understand how that is transmitted

405

00:44:43,640 –> 00:44:48,840

through the family. Great thank you for that answer. The next question I have here is, has

406

00:44:48,840 –> 00:44:57,160

antibody research led to successful neuroscience treatments? Yes, so for example, if one

407

00:44:57,160 –> 00:45:06,560

thinks about the nmda receptor syndrome

that leads to, what looks early on, can look like

408

00:45:06,560 –> 00:45:12,680

schizophrenia This is secondary to an autoimmune process, where an antibody has

409

00:45:12,680 –> 00:45:20,680

formed against the nmda receptor and that nmda receptor then leads to the psychiatric symptoms.

410

00:45:20,680 –> 00:45:28,080

Rather than being treated with drugs, small

411

00:45:28,080 –> 00:45:35,720

by using approaches for against the antibody that’s ideologically responsible for the

412

00:45:35,720 –> 00:45:43,000

condition, which is an nmda receptor in cephalitis. Great thank you and just a final question for

413

00:45:43,000 –> 00:45:49,200

you now Sheldon. Have there been any major recent developments in senile dementia in the Alzheimer’s

414

00:45:49,200 –> 00:45:59,560

type? Yes, so a recently published article in the New England Journal of Medicine has shown, when

415

00:45:59,560 –> 00:46:07,120

various proteins, that have been for a long time associated with Alzheimer’s disease, how far in

416

00:46:07,120 –> 00:46:17,360

advance those proteins are expressed before you develop the disease. So, amyloid beta 42 actually

417

00:46:17,360 –> 00:46:27,120

is expressed 17 years before you develop the disease process and a ratio of amyloid beta 42

418

00:46:27,120 –> 00:46:37,880

to the regular amyloid beta, is 14 years before and phosphorylated tau, the red, is 11

419

00:46:37,880 –> 00:46:44,960

years before and so on. So that we now understand the progression of when these proteins begin

420

00:46:44,960 –> 00:46:52,200

eventually leads to Alzheimer’s disease. With this

421

00:46:52,200 –> 00:46:58,640

kind of understanding of the molecular biology, we can intervene on any one of those specific

422

00:46:58,640 –> 00:47:05,000

targets. Fantastic thank you for answering those questions. I’d like to now remind our audience

423

00:47:05,000 –> 00:47:09,120

that if you’ve missed any presentations so far today, you’ll be able to access them on demand

424

00:47:09,120 –> 00:47:18,080

on the DDW website www.ddw-online.com. Before we move on to our final presentation, we will have

425

00:47:18,080 –> 00:47:25,240

another product presentation from our co-sponsor Axosim, including an audience Q&A. Thank you for

426

00:47:25,240 –> 00:47:30,280

having me. My name is Andy Croy and I’m a lead scientist at Axosim, leading the development of

427

00:47:30,280 –> 00:47:36,080

our brain organoid models and their applications to preclinical safety and efficacy testing. Today I’ll

428

00:47:36,080 –> 00:47:39,760

focus on the safety side of the equation and tell you about how we are predicting human

429

00:47:39,760 –> 00:47:46,720

clinical neurotoxicity using functional phenotypic screening on our cortical organoid platform. First,

430

00:47:46,720 –> 00:47:52,240

let’s look at the problem we’re trying to address. 90% of drugs fail in the clinic. Unfortunately

431

00:47:52,240 –> 00:47:57,080

this high failure rate is also accompanied by extremely high costs of failure. One of the

432

00:47:57,080 –> 00:48:02,240

main drivers of clinical failures is poor safety profile. Multiple studies have highlighted the

433

00:48:02,240 –> 00:48:06,720

difficulties in predicting CNS adverse events and the high rate of drug withdrawals due to

434

00:48:06,720 –> 00:48:13,960

neurotoxicity. This is where Axosim platforms come in. We have developed multiple invitro

435

00:48:13,960 –> 00:48:20,320

IPSC derived platforms that allow for more drugs to be tested in human relevant systems earlier.

436

00:48:20,320 –> 00:48:25,880

Getting effective treatments to patients faster while avoiding costly late stage failures. Before

437

00:48:25,880 –> 00:48:30,920

getting into the predictive neurotoxicity assay itself, I’d like to provide a brief introduction

438

00:48:30,920 –> 00:48:37,640

to the underlying organoid platform. There are a huge number of invitro and in vivo models that have

439

00:48:37,640 –> 00:48:44,080

distinct use cases during drug development. While the role of microphysiological systems MPS and

00:48:44,080 –> 00:48:48,800

organoid technologies has dramatically expanded over the years. I would argue that they’re still

441

00:48:48,800 –> 00:48:55,440

underutilized, in that they provide much needed tissue complexity, multiple cell types, 3D biology

442

00:48:55,440 –> 00:49:01,560

while maintaining the throughput required for

high throughput screening techniques. Our organoid

443

00:49:01,560 –> 00:49:07,360

technology is particularly powerful because it was developed as a product, not just a protocol.

444

00:49:07,360 –> 00:49:12,040

Meaning these brain organoids are amenable to live shipping and are assay ready in seven

445

00:49:12,040 –> 00:49:17,520

days. This platform has been used by government, academic and industry groups around the world

446

00:49:17,520 –> 00:49:22,680

and can be adopted widely and quickly by anyone who needs it. We created these functional cortical

447

00:49:22,680 –> 00:49:29,040

organoids by seeding neuroprogenitor cells in ultra low attachment plates, either 96 well

448

00:49:29,040 –> 00:49:35,680

or 384 well formats, creating one organoid per well. These NPCs co- differentiate into cortical

449

00:49:35,680 –> 00:49:43,320

neurons and astrocytes in situ forming organoids approximately 600 microns in diameter. We see

00:49:43,320 –> 00:49:48,600

evidence of the mature biology in the cell types present in the platform. You can see multiple

451

00:49:48,600 –> 00:49:52,720

neuronal populations of varying maturity in the umap plot of the single cell data

452

00:49:52,720 –> 00:49:58,120

on the right we see primarily glutamatergic neurons in red but also the beginnings of a

453

00:49:58,120 –> 00:50:04,280

small gabic population in blue and more immature population of neurons in green. These neurons are

454

00:50:04,280 –> 00:50:09,160

supported by a large population of astrocytes and relatively few progenitor cells remain in the

455

00:50:09,160 –> 00:50:17,160

cultures. The hallmark of this platform is part of its name, spontaneous functional

456

00:50:17,160 –> 00:50:22,160

activity. This movie is showing a single organoid loaded with a fluorescent calcium

457

00:50:22,160 –> 00:50:27,000

reporter die, that enables visualization of this activity in high resolution.

458

00:50:27,000 –> 00:50:31,440

This activity can be quantified in terms of a fluoresence overtime trace, which is compatible

459

00:50:31,440 –> 00:50:36,320

with high throughput kinetic plate readers, like the flipper or hamamatsu. Production of

00:50:36,320 –> 00:50:42,000

organoids in 384 well plates, enables screening up to 20,000 wells in a single day, making this

461

00:50:42,000 –> 00:50:46,800

functional screening truly high throughput. While the robustness and scale of functional

462

00:50:46,800 –> 00:50:50,680

screening is important, the value of these measurements is their ability to answer

463

00:50:50,680 –> 00:50:56,840

deep biological questions. As you’ll see in the coming slides, changes in the number, size or shape

464

00:50:56,840 –> 00:51:02,200

of these waveforms is indicative of underlying biological changes, such as neurotoxicity, the

465

00:51:02,200 –> 00:51:08,520

focus of today, as well as synaptogenesis, excited toxicity, channel activity amongst

466

00:51:08,520 –> 00:51:16,240

others. Version 1.0 of the neurotoxicity prediction model was developed in collaboration with Takeda

467

00:51:16,240 –> 00:51:21,720

Pharmaceuticals. The task we gave this model was to predict whether a set of 84 reference compounds

468

00:51:21,720 –> 00:51:27,000

were neurotoxic or whether they were safe. These compounds were selected because they represent

469

00:51:27,000 –> 00:51:32,480

diverse mechanisms of action and are predicted to cross the blood brain barrier. They were classified

00:51:32,480 –> 00:51:37,480

as either safe or neurotoxic based on their reported adverse event rates and were separated

471

00:51:37,480 –> 00:51:42,720

into training and test sets to avoid overfitting during regression modeling. Perhaps most

472

00:51:42,720 –> 00:51:49,480

importantly, this set of 84 compounds is poorly predicted by pre-clinical animal safety tests.

473

00:51:49,480 –> 00:51:54,800

While animal testing provides good sensitivity for toxic compounds, three out of four toxic compounds

474

00:51:54,800 –> 00:52:00,600

are properly identified, those measurements provide poor specificity and yield a huge number

475

00:52:00,600 –> 00:52:06,080

of false positives. The fact that these standard non-clinical safety assessments falsely identify

476

00:52:06,080 –> 00:52:11,800

neurotoxic liability in safe compounds, means that we may be missing out on valuable and effective

477

00:52:11,800 –> 00:52:21,800

therapeutics. So how can we improve upon this using our functional cortical organoid? First

478

00:52:21,800 –> 00:52:26,320

we measure functional changes caused by each compound at concentrations we expect to see in

479

00:52:26,320 –> 00:52:32,760

vebo and we saw a variety of responses. On the left you can see a plate view of the compound

00:52:32,760 –> 00:52:37,640

responses two hours post treatment. I hope you can appreciate the consistency of response well

481

00:52:37,640 –> 00:52:45,040

to well in this 3D human IPSC drived organoid platform. This consistency means that as few as

482

00:52:45,040 –> 00:52:50,600

four replicates per condition can be used for compound profiling. Relative to normal activity

483

00:52:50,600 –> 00:52:56,520

in vehicle treated wells, DMSO at the top, known safe compounds like acetaminophen, rarely show

484

00:52:56,520 –> 00:53:02,600

CNS functional modulation. Green traces indicate test concentrations 1X of the cmax, while red

485

00:53:02,600 –> 00:53:09,240

is 10x of the cmax. In contrast, known toxic or neurodegenerative compounds, for example carboplatin,

486

00:53:09,240 –> 00:53:17,600

frequently show significant modulation at

or below the in vivo cmax. To give you a little

487

00:53:17,600 –> 00:53:23,080

deeper look into the data, here I’m showing the time and dose dependent response characteristics

488

00:53:23,080 –> 00:53:29,560

for safe, intermediate and toxic compounds, as classified by their clinical adverse event rates.

489

00:53:29,560 –> 00:53:34,720

Again in this quantitative data, you can see the consistency of pharmacological response, meaning we

00:53:34,720 –> 00:53:40,680

are highly confident in the potency values coming out of this assay. When we put these functional

491

00:53:40,680 –> 00:53:44,840

measurements in perspective with the TPC Max data we can clearly see the shrinking margin

492

00:53:44,840 –> 00:53:49,920

of safety for these compounds with poor clinical safety profiles. Since this work was published in

493

00:53:49,920 –> 00:53:55,720

2022, we wanted to assess the stability of these measurements and predictions over time. So we

494

00:53:55,720 –> 00:54:00,840

recently re screened all 84 compounds and saw strong agreement between the two studies,

495

00:54:00,840 –> 00:54:05,160

both in these representative measurements I’m showing here and in the overall model

496

00:54:05,160 –> 00:54:12,960

performance. So how do we put this all together

into an assay workflow? Again we start by measuring

497

00:54:12,960 –> 00:54:18,080

the potency of each test article, in our functional phenotypic assay. I’m showing eight representative

498

00:54:18,080 –> 00:54:24,080

test articles of varying potency here, in terms of

their peak count response. However, we can quantify

499

00:54:24,080 –> 00:54:30,240

several other features describing the size, shape and variability of the waveforms as well, providing

00:54:30,240 –> 00:54:35,440

even more predictors to improve model performance. We then compare the concentrations at which these

501

00:54:35,440 –> 00:54:40,600

functional changes occur, to the anticipated total plasma cmax of the test article, the

502

00:54:40,600 –> 00:54:45,960

green dotted line, I’ve shown you how neurotoxic compounds generally show functional modulation

503

00:54:45,960 –> 00:54:50,560

below the TPC Max while safe compounds require much higher concentrations to show functional

504

00:54:50,560 –> 00:54:56,320

changes. These margin of exposure measurements are fed into a logistic regression model

505

00:54:56,880 –> 00:55:01,840

that provides a toxicity prediction score between zero and one. Overall,

506

00:55:01,840 –> 00:55:05,840

this fully automated workflow provides consistent and unbiased neurotoxicity

507

00:55:05,840 –> 00:55:13,000

scoring. Looking at the prediction performance, we quickly see that this assay is excellent at

508

00:55:13,000 –> 00:55:19,400

identifying safe drugs as safe, as indicated by its excellent specificity. On the left, I’m

509

00:55:19,400 –> 00:55:24,840

showing our two classes of compounds; safe in green, seizurogenic or neurotoxic in red and

00:55:24,840 –> 00:55:29,800

their predicted neurotoxicity score on the y- axis. You can see that there are only two false

511

00:55:29,800 –> 00:55:34,520

positives in the data set. Additionally, over half of the toxic compounds were predicted

512

00:55:34,520 –> 00:55:40,760

as such. Relative to the standard preclinical safety package of two species animal testing,

513

00:55:40,760 –> 00:55:45,680

these human IPSC derived organoid predictions, in two independent studies, showed dramatically

514

00:55:45,680 –> 00:55:52,480

better specificity. Over 90% compared to only 30% and are excellent at identifying safe drugs as

515

00:55:52,480 –> 00:55:57,760

safe. Using this platform ahead of animal testing could avoid filtering out potentially valuable

516

00:55:57,760 –> 00:56:06,640

therapeutics. To recap; clinical failures due to

CNS toxicity are far too common, animal models do

517

00:56:06,640 –> 00:56:12,360

offer good predictive value but these predictions frequently suffer from port specificity or high

518

00:56:12,360 –> 00:56:17,720

false positive rates. Axosim functional cortical organoid model provides insight into clinical

519

00:56:17,720 –> 00:56:23,720

neurotoxicity and is a natural complimentary pre-screening assay prior to animal testing. Its

00:56:23,720 –> 00:56:28,960

excellent specificity correctly identifies safe drugs for further development, while filtering

521

00:56:28,960 –> 00:56:34,760

out over half of drugs that would be toxic in humans. Overall expanded use of IPSC

522

00:56:34,760 –> 00:56:40,000

organoid technologies like this, will accelerate the development and reduce the cost to deliver

523

00:56:40,000 –> 00:56:46,400

safe and effective treatments to patients. A final note, what I’ve shown today is just

524

00:56:46,400 –> 00:56:52,000

one of Axosim’s three human IPSC drive neuro platforms. Each interrogates distinct biology

525

00:56:52,000 –> 00:56:57,160

and has specific use cases. Today we examined a novel use case for the functional cortical

526

00:56:57,160 –> 00:57:02,800

organoid model on the left, the hallmark of which is coordinated network activity. Beyond toxicity

527

00:57:02,800 –> 00:57:07,360

testing, we’ve used this platform to study the mechanisms of seizure and epilepsy and model

528

00:57:07,360 –> 00:57:13,080

neurodevelopmental disorders like Rett syndrome and cdk5 deficiency disorder. We also have a

529

00:57:13,080 –> 00:57:17,840

related but distinct cortical brain organoid model licensed from the heart/tongue lab out of John’s

00:57:17,840 –> 00:57:23,640

Hopkins. The hallmark of which is the presence of an dendritic cell lineage that gives rise to

531

00:57:23,640 –> 00:57:28,560

myelination. These myelinating organoids are a critical tool to those studying or developing

532

00:57:28,560 –> 00:57:34,320

drugs for demyelinating diseases like multiple sclerosis. Finally, we have a peripheral nerve

533

00:57:34,320 –> 00:57:40,000

model that is grown in custom embedded electrode array plates, allowing for non-destructive efiz

534

00:57:40,000 –> 00:57:44,360

recordings of nerve conduction velocity, that we’ve shown to be altered by induced models of

535

00:57:44,360 –> 00:57:48,920

peripheral neuropathy or in neuropathic pain. Thank you again for your time and

536

00:57:48,920 –> 00:57:56,840

attention. Thank you for that presentation. I’ll now move on to the questions we have here for

537

00:57:56,840 –> 00:58:04,360

you. The first is, was this neurotoxicity prediction algorithm specifically designed on small molecule

538

00:58:04,360 –> 00:58:09,960

drugs, is the cortical organoid platform and functional screening compatible with other

539

00:58:09,960 –> 00:58:15,680

types of therapeutics that may have a harder time penetrating the 3D tissue structure? Thank you for

00:58:15,680 –> 00:58:21,160

the question. You are correct, this specific work focused on small molecule drugs. However, this

541

00:58:21,160 –> 00:58:26,720

organoid platform can be used to study a variety of other treatment modalities. We now know that

542

00:58:26,720 –> 00:58:32,560

Asos and snas have good tissue penetration and can be used to target genes of interest. We’ve

543

00:58:32,560 –> 00:58:38,320

also tested several serotypes of AAV and see good expression of an m-cherry reporter especially with

544

00:58:38,320 –> 00:58:45,760

AAV9, at least 100 microns into this organoids. Finally, inferring from the difficulties of 3D

545

00:58:45,760 –> 00:58:50,920

immunocyto chemistry staining, antibodies take much longer to penetrate to the center of organoids but

546

00:58:50,920 –> 00:58:56,800

can still be delivered again about 100 micrometers into the shell. Thank you for that answer and now

547

00:58:56,800 –> 00:59:02,640

just a final question for you. Can these calcium flux measurements be multiplexed with other end

548

00:59:02,640 –> 00:59:09,080

points for deeper understanding of cell specific responses to drug treatments? Functional screening

549

00:59:09,080 –> 00:59:14,320

can indeed be combined with other endpoints. From a technical perspective, the fluorescent die can be

00:59:14,320 –> 00:59:21,160

rinsed out post flipper measurements enabling organoid to be used for other endpoints. For

551

00:59:21,160 –> 00:59:25,840

investigations where cell type specific responses are of interest, we can fix and stain organoids via

552

00:59:25,840 –> 00:59:32,040

ICC, the organoids can be pulled and submitted for transcriptomic analysis. We can always collect

553

00:59:32,040 –> 00:59:37,840

media supernat longitudinally and assess things like cytokine release to measure immune responses

554

00:59:37,840 –> 00:59:44,200

or lactate dehydrogenase for a quick cytotoxicity assessment. What I’ve shown today is just a single

555

00:59:44,200 –> 00:59:49,400

specific use case that I believe has the potential to dramatically improve preclinical safety

556

00:59:49,400 –> 00:59:55,680

investigations but this organoid technology can

be used for so much more. Those were great answers,

557

00:59:55,680 –> 01:00:02,360

thank you. You finally we will be hearing from Dr

Kiri Granger, CSO of Monument Therapeutics, about the

558

01:00:02,360 –> 01:00:07,160

company’s use of precision medicine to develop effective treatment options for areas of high

559

01:00:07,160 –> 01:00:14,160

unmet need in Psychiatry and neurology. Over to you Kiri. Hi everyone, I’m Kiri Granger. I just want to

01:00:14,160 –> 01:00:20,160

start by thanking the organizers for inviting me to present at today’s webinar. I’m Chief

561

01:00:20,160 –> 01:00:26,400

scientific officer here at Monument Therapeutics and we are a precision medicine biotech company.

562

01:00:26,400 –> 01:00:33,680

So we use digital biomarkers for patient selection

in the development of treatments for CNS disorders.

563

01:00:33,680 –> 01:00:38,680

Today I’m going to be talking to you about what the key challenges are in neuroscience

564

01:00:38,680 –> 01:00:45,440

drug development and what research tools we can be using to improve clinical trial outcomes and get

565

01:00:45,440 –> 01:00:52,640

effective treatments into the hands of patients, who need them the most. So focusing on clinical

566

01:00:52,640 –> 01:00:59,360

trials for CNS disorders. Now brain disorders remain among the greatest unmet medical needs

567

01:00:59,360 –> 01:01:05,480

due to reasons such as their complex biology, symptom variability and lack of biomarkers to

568

01:01:05,480 –> 01:01:13,080

guide personalized treatment approaches. So when it comes to clinical trials, a staggering 95%

569

01:01:13,080 –> 01:01:18,120

of neuroscience drugs fail during clinical development and there’s a number of reasons

01:01:18,120 –> 01:01:24,400

for this. There is the complexity of the brain, difficulty in translating preclinical findings to

571

01:01:24,400 –> 01:01:30,920

humans insensitive end points and crucially, patient heterogeneity. As patients with the same

572

01:01:30,920 –> 01:01:37,520

diagnosis can present a wide variability in their symptoms due to differences in their underlying

573

01:01:37,520 –> 01:01:44,320

biology, therefore they have different

responses to specific treatments. So approved drugs

574

01:01:44,320 –> 01:01:50,160

that we have available, for example for depression and schizophrenia, they only work for around half

575

01:01:50,160 –> 01:01:57,840

of patients or even only up to half of patients. So a lots of patients don’t take their prescribed

576

01:01:57,840 –> 01:02:04,160

medications either because of little benefit or the side effects that they experience, that ultimately

577

01:02:04,160 –> 01:02:12,120

leads to poor adherence and that can lead to symptom relapse. So the core challenges that we face in CNS

578

01:02:12,120 –> 01:02:17,800

development is that brain health just isn’t something that is routinely measured and our

579

01:02:17,800 –> 01:02:26,000

endpoints that we often use, sometimes have a heavy reliance on subjective self-report of symptoms and

01:02:26,000 –> 01:02:32,040

a variability in diagnoses that don’t account for underlying biology, leading to that lack of

581

01:02:32,040 –> 01:02:38,320

objective measures for both patient inclusion and also then ways to be able to measure treatment

582

01:02:38,320 –> 01:02:44,840

effects. So there’s a pressing need for biomarkers to help identify the right kind of patients for

583

01:02:44,840 –> 01:02:51,520

inclusion into clinical trials and also to

be able to effectively and accurately measure

584

01:02:51,520 –> 01:02:57,160

treatment effects So then when we reflect on the history of clinical trial failures, there

585

01:02:57,160 –> 01:03:03,800

are several learning opportunities that we can be reflecting upon; was it either the drug, the

586

01:03:03,800 –> 01:03:10,080

end points were they measurable or meaningful, was it the patient population or perhaps was it

587

01:03:10,080 –> 01:03:17,760

was it a combination of all three? So where does this leave us? Biomarkers ultimately save lives.

588

01:03:17,760 –> 01:03:24,480

So, thanks to advances in biomarker use, such

as cholesterol levels and blood pressure, we have

589

01:03:24,480 –> 01:03:30,840

seen significant dropping mortality from vascular diseases. That’s because biomarkers allow

01:03:30,840 –> 01:03:37,760

for early detection, improved risk stratification

and more personalized treatments. But unfortunately

591

01:03:37,760 –> 01:03:45,520

in contrast then mental health disorders, such as depression, these areas really lack biomarkers and

592

01:03:45,520 –> 01:03:52,600

other areas such as this and psychiatry as well. So we continue to see really concerning trends.

593

01:03:52,600 –> 01:03:59,520

On this slide now, we’re looking at death rate due to suicide in young individuals and we have

594

01:03:59,520 –> 01:04:06,480

just continued to witness an increase over the years, particularly in males. So this really

595

01:04:06,480 –> 01:04:15,080

underscores the need for us to have objective measures, objective biomarkers within psychiatry. So,

596

01:04:15,080 –> 01:04:22,040

just as cardiovascular biomarkers have transformed heart disease outcomes, then biomarkers are also

597

01:04:22,040 –> 01:04:27,920

really key to transforming how we approach brain disorders. So this is something that we really need

598

01:04:27,920 –> 01:04:35,680

to prioritize. When it comes to how we diagnose mental health conditions, then we are reliant on

599

01:04:35,680 –> 01:04:42,000

the DSM so the diagnostic and statistical manual of mental health disorders, which is of course

01:04:42,000 –> 01:04:49,800

useful and has been a useful tool in psychiatry but it’s primarily based on symptom clusters.

601

01:04:49,800 –> 01:04:56,920

Symptoms clusters are subjective and unfortunately they’re not reflective of the underlying biology of

602

01:04:56,920 –> 01:05:04,640

the disorders. So the DSM doesn’t account for the biological variability underpinning symptoms

603

01:05:04,640 –> 01:05:11,560

or the biological variability that goes on between patients. That is most likely to contribute to that

604

01:05:11,560 –> 01:05:18,320

high clinical trial failure rate that we just continue to witness in this space and that’s

605

01:05:18,320 –> 01:05:24,720

because two patients with the same dsm5 diagnosis might have vasly different underlying biological

606

01:05:24,720 –> 01:05:32,800

mechanisms, leading to that heterogeneity in response to drug treatment. So the DSM

607

01:05:32,800 –> 01:05:39,800

is subjective and ultimately in opposition to a biomarker driven approach, where biomarkers rely on

608

01:05:39,800 –> 01:05:47,920

objective measures that can identify biologically distinct subtypes, that exist within the DSM and to

609

01:05:47,920 –> 01:05:55,840

enable those more targeted treatment interventions and ultimately better outcomes for patients.

01:05:55,840 –> 01:06:00,600

Our mental state isn’t just something that can be measured when a person is in front

611

01:06:00,600 –> 01:06:05,920

of a clinician, in a clinician’s office, then in today’s world. It’s something that can be measured

612

01:06:05,920 –> 01:06:12,000

and monitored through real world interactions, everyday technologies that can really capture

613

01:06:12,000 –> 01:06:19,680

valuable objective data and that’s embedded within our daily environments and and activities. So I

614

01:06:19,680 –> 01:06:25,840

think this slide just touches on and highlights those diverse sources from which mental health

615

01:06:25,840 –> 01:06:32,080

related data can be collected and these sources range from physical locations such as hospitals

616

01:06:32,080 –> 01:06:40,200

and even schools, to devices that we use every single day. Our smartphones and wearable devices

617

01:06:40,200 –> 01:06:47,440

but also data that can be sourced from public places and personal interactions. So to take one of these

618

01:06:47,440 –> 01:06:56,080

examples, our smartphone generate such a wealth of data about our our brain health and that can

619

01:06:56,080 –> 01:07:03,680

be taken from the frequency of text messages that we send or the phone calls that we send or receive,

01:07:03,680 –> 01:07:10,480

our typing patterns, our sleep patterns and even screen time can provide insights into mood, anxiety

621

01:07:10,480 –> 01:07:16,960

levels and our cognitive function. Geolocation data and movement metrics are also increasingly

622

01:07:16,960 –> 01:07:23,640

being used, for example to infer states like depression and isolation. So by using objective

623

01:07:23,640 –> 01:07:29,480

tools, we can help to treat patients more as individuals and make treatments that little bit

624

01:07:29,480 –> 01:07:35,960

more proactive rather than just reactive, to help identify science of of mental health conditions

625

01:07:35,960 –> 01:07:43,720

sooner and be able to improve patient lives. Next I want to just be able to provide a use case

626

01:07:43,720 –> 01:07:50,160

example and provide an example to demonstrate how Monument are using digital biomarkers as

627

01:07:50,160 –> 01:07:55,960

objective tools to be able to improve clinical trial outcomes and this to is on some of the

628

01:07:55,960 –> 01:08:02,880

obstacles that I have laid out in in my previous slides. We recognize that most CNS drugs fail in

629

01:08:02,880 –> 01:08:09,480

clinical trials because diagnostic criteria do not reflect biological reality. So all patients

01:08:09,480 –> 01:08:16,040

are enrolled into schizophrenia clinical trial for example, if they need a dsm5 diagnosis for

631

01:08:16,040 –> 01:08:22,680

a particular disorder but not actually based on their underlying biological dysfunctions. So what

632

01:08:22,680 –> 01:08:28,080

we end up within clinical trials, is this really hetrogeneous group of individuals that all have

633

01:08:28,080 –> 01:08:35,320

different underlying causes for their symptoms and subsequently only some patients, but not all, will

634

01:08:35,320 –> 01:08:41,480

show a positive response to treatment. So then what we do at Monument Therapeutics, is to use

635

01:08:41,480 –> 01:08:48,600

digital biomarkers to identify patients with abnormalities in specific brain processes. So

636

01:08:48,600 –> 01:08:55,080

therefore, we’re using these digital biomarkers as a way to identify biologically homogeneous

637

01:08:55,080 –> 01:09:01,120

subgroup groups of patients and at that point then, we use our novel drug formulations to be able to

638

01:09:01,120 –> 01:09:07,040

remediate that specific underlying biological mechanism that our digital biomarker has been

639

01:09:07,040 –> 01:09:14,840

able to identify. So the digital biomarkers that

we use at Monument Therapeutics to overcome some of

01:09:14,840 –> 01:09:20,160

these obstacles and provide a precision medicine approach to neuroscience drug development, is to

641

01:09:20,160 –> 01:09:27,760

use cognitive digital biomarkers. Now cognitive tests can measure brain function. So across a

642

01:09:27,760 –> 01:09:33,800

range of neuro psychiatric illnesses, then cognitive tests are able to measure demain

643

01:09:33,800 –> 01:09:40,240

specific cognitive impairments, they can measure specific brain region activity, so structure and

644

01:09:40,240 –> 01:09:45,960

neural networks through the work and validation that has been completed with neuroimaging

645

01:09:45,960 –> 01:09:52,320

and electrophysiology related tools. We also understand a lot about their sensitivity to

646

01:09:52,320 –> 01:09:59,440

neurotransmitter systems, such as chemical activity, the sector binding, as well as a lot of other wet

647

01:09:59,440 –> 01:10:05,080

biomarker work that’s being completed today and there’s a lot that we understand about digital

648

01:10:05,080 –> 01:10:10,520

cognitive assessments and their responsiveness to pharmacological manipulation, through their use in

649

01:10:10,520 –> 01:10:18,080

clinical research and also their associations

with genetic makeup and genomic variations. So

01:10:18,080 –> 01:10:23,240

the cognitive physical biomarkers that we use and have been validated to this extent

651

01:10:23,240 –> 01:10:28,480

is through the use of the the unab cognitive assessing and the cognitive testing battery,

652

01:10:28,480 –> 01:10:36,640

that we have an exclusive license to use. So to put this directly into the context of one of

653

01:10:36,640 –> 01:10:43,240

our drug development programs. Then here, I’m just going to talk you through our latent inhibition marker and

654

01:10:43,240 –> 01:10:49,600

how we’re using this for a precision medicine targeted approach in our clinical trial program. So

655

01:10:49,600 –> 01:10:55,560

our schizophrenia program is fating cognitive impairment in schizophrenia, for which there are no

656

01:10:55,560 –> 01:11:01,240

treatments that are currently available. The digital cognitive biomarker that we are using

657

01:11:01,240 –> 01:11:07,240

here is a cognitive task of something called latent inhibition. Now really simply, this is

658

01:11:07,240 –> 01:11:13,880

just a measure of our attentional gating. So this is something that measures our ability to focus

659

01:11:13,880 –> 01:11:20,080

on relevant information and block out irrelevant information so that we don’t get overwhelmed by

01:11:20,080 –> 01:11:25,520

everything that’s going on in our environment. However, this attentional gating process

661

01:11:25,520 –> 01:11:31,200

is something that is largely disrupted in some patients with schizophrenia, so they can’t use

662

01:11:31,200 –> 01:11:37,480

that filter and then unsurprisingly become very overwhelmed by everything that’s surrounding them

663

01:11:37,480 –> 01:11:43,840

and they can’t learn to switch between what’s relevant and what is irrelevant and this is

664

01:11:43,840 –> 01:11:51,120

really deemed to be a core deficit that underpins their symptom expression of cognitive impairments.

665

01:11:51,120 –> 01:11:57,280

In terms of its biological underpinnings, then there a huge body of of evidence to support that

666

01:11:57,280 –> 01:12:03,840

cognitive impairment as identified by this task, of latent inhibition represents a cholinergic state

667

01:12:03,840 –> 01:12:10,680

of cognitive impairment. So this data and validation makes the task a really promising

668

01:12:10,680 –> 01:12:17,880

biomarker for the investigation of cholinergic modulation, so nicotinic modulation on cognitive

669

01:12:17,880 –> 01:12:25,640

impairment in patients with schizophrenia. So ultimately we’re using latent inhibition

01:12:25,640 –> 01:12:31,280

as a way to identify the subpopulation of patients within our clinical trials, that would

671

01:12:31,280 –> 01:12:38,240

benefit from treatment with an alpha7 nicotinic receptor agonist. This could really enhance the

672

01:12:38,240 –> 01:12:43,960

prediction of treatment efficacy in this patient population, for which no current treatments are

673

01:12:43,960 –> 01:12:52,280

available right now. So we envisage using this biomarker and we will be using this biomarker

674

01:12:52,280 –> 01:12:59,800

as a way to stratify patient population, as a way to stratify the patient population and identify

675

01:12:59,800 –> 01:13:07,080

those patients that are most likely to respond to treatment based on their underlying biology

676

01:13:07,080 –> 01:13:14,480

Interestingly, as well so digital biomarkers, they aren’t just valuable for improving clinical

677

01:13:14,480 –> 01:13:21,680

trial success rates but also post approval so these tools can really help to overcome trial and

678

01:13:21,680 –> 01:13:30,320

error prescribing, as well as being able to provide a basis for value based reimbursement and if you’d

679

01:13:30,320 –> 01:13:36,480

like to see a demonstration of this cognitive biomarker in action or just to really get a feel

01:13:36,480 –> 01:13:43,520

for how these cognitive biomarkers look, then

I have embedded a link here within the slide but

681

01:13:43,520 –> 01:13:49,080

also very happy to hear from any individuals that that would like to discuss any of these topics

682

01:13:49,080 –> 01:13:57,560

further. I really hope this brief talk has been able to provide just a useful snapshot for how

683

01:13:57,560 –> 01:14:03,520

biomarkers can drive precision medicine and shape the future of neuroscience drug development

684

01:14:03,520 –> 01:14:10,280

and I would really love to take any questions now from the audience, thank you. Thank you for that

685

01:14:10,280 –> 01:14:17,320

presentation. I’ll now head over to some questions that we have in for you. The first one is, what

686

01:14:17,320 –> 01:14:25,200

role do cognitive biomarkers play in the shift towards precision medicine in neuroscience? So in

687

01:14:25,200 –> 01:14:30,880

essence, cognitive biomarkers help to stratify patient populations based on their biological

688

01:14:30,880 –> 01:14:38,360

characteristics rather than relying solely on

broad diagnostic categories defined and by the DFM

689

01:14:38,360 –> 01:14:44,080

and so cognitive biomarkers are instrumental in shifting neuroscience towards precision medicine,

01:14:44,080 –> 01:14:51,200

by linking cognitive function to its biological underpinnings and enabling targeted therapies.

691

01:14:51,200 –> 01:14:57,160

So I think this not only improves patient outcomes but it can also enhance the efficiency of clinical

692

01:14:57,160 –> 01:15:03,000

trials and drug development processes, by allowing for more precise recruitment and

693

01:15:03,000 –> 01:15:08,880

tailored intervention strategies. I also think that the integration of cognitive biomarkers

694

01:15:08,880 –> 01:15:14,800

into clinical practice allows for continue with monitoring treatment effects and progression of

695

01:15:14,800 –> 01:15:21,360

the disease, so this realtime feedback can really lead to dynamic treatment adjustments and ensuring

696

01:15:21,360 –> 01:15:26,960

that therapeutic strategies remain aligned with the patients involved biological and cognitive

697

01:15:26,960 –> 01:15:33,880

states. Thanks Kiri, great answer. Now just one more question for you, what challenges do you

698

01:15:33,880 –> 01:15:41,200

anticipate in implementing cognitive biomarkers in clinical practice and how might they be overcome?

699

01:15:41,200 –> 01:15:47,120

I think one of the key considerations is centered around clinician acceptance and integration into

01:15:47,120 –> 01:15:53,440

existing workflows. Clinicians are, of course, accustomed to traditional diagnostic methods. So

701

01:15:53,440 –> 01:15:59,000

to overcome these challenges, I think we need to focus on educating healthcare professionals about

702

01:15:59,000 –> 01:16:05,120

the value of cognitive biomarkers in enhancing patient care and this includes developing

703

01:16:05,120 –> 01:16:11,720

training programs and resources that demonstrate how to integrate these tools into routine practice

704

01:16:11,720 –> 01:16:16,960

effectively. I think as more evidence emerges showing their efficacy and improving patient

705

01:16:16,960 –> 01:16:23,320

outcomes then naturally we would expect acceptance and implementation to grow within the within the

706

01:16:23,320 –> 01:16:30,360

medical community. So ultimately by addressing educational gaps, fostering collaboration and

707

01:16:30,360 –> 01:16:36,720

emphasizing the clinical benefits through research, I think we can facilitate a smoother transition to

708

01:16:36,720 –> 01:16:45,080

incorporating cognitive biomarkers in research. Thanks so much for listening. That brings us to

709

01:16:45,080 –> 01:16:50,640

the end of the presentations today and to the

end of the DDW neuroscience showcase. I’d like

01:16:50,640 –> 01:16:54,840

to thank everyone for tuning in and a big thanks to our panelists and presenters today.

711

01:16:55,480 –> 01:16:59,720

If we did not get your questions in the chat panel, we will follow up after this

712

01:16:59,720 –> 01:17:14,480

event. If you would like to listen to the event again it will be available on demand at www.ddw- online.com. Thank you to the co-sponsors of today’s showcase, ApconiX and Axosim. You can now view

713

01:17:14,480 –> 01:17:22,440

drug discovery world content online or on your mobile or tablet. Thank you for joining. Goodbye.

Cookie	Duration	Description
cookielawinfo-checkbox-analytics	11 months	This cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Analytics".
cookielawinfo-checkbox-functional	11 months	The cookie is set by GDPR cookie consent to record the user consent for the cookies in the category "Functional".
cookielawinfo-checkbox-necessary	11 months	This cookie is set by GDPR Cookie Consent plugin. The cookies is used to store the user consent for the cookies in the category "Necessary".
cookielawinfo-checkbox-others	11 months	This cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Other.
cookielawinfo-checkbox-performance	11 months	This cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Performance".
viewed_cookie_policy	11 months	The cookie is set by the GDPR Cookie Consent plugin and is used to store whether or not user has consented to the use of cookies. It does not store any personal data.