2023-10-18

Jean Hebert

video: https://www.youtube.com/watch?v=mcyc8miGLpI

other video: https://www.youtube.com/watch?v=3s31WctpHSM&list=PLH78wfbGI1x0Bpmq40HEVtYGj_CsEfHSk&index=120&pp=iAQB

Progressive brain tissue replacement and engineering replacement tissue to reverse damage to the central nervous system for longevity

Introduction

Our next speaker is Dr. Jean Hébert. He's one of our extramural researchers. He's a professor at the Albert Einstein College of Medicine in New York. His presentation today is Progressive Brain Tissue Replacement and How We Beat Aging. You can read his presentation summary and bio on the events website. And now let's welcome him to our stage.

Hi, everyone. uh thanks so much for having me um i'd like to acknowledge the SENS Research Foundation in several ways one for their support of our projects aimed at reversing brain aging but also for the excellent uh students that they've sent our way uh you know recently we've had uh Sanjana, who's presenting a poster tomorrow. And so really excellent students. So thanks for having that program and letting us participate in it.

Brain tissue replacement approach for brain rejuvenation and longevity

So, you know, there are a couple of things that we work on in the lab. One is a microglial replacement project. And I think I've talked about that in the past. So I'm going to focus a little bit more on a brain tissue replacement approach and why I think that that's really important for defeating not just brain aging, but overall aging. And I'll give you my reasons why, and then I'll be happy to discuss and take questions for, you know, if anybody else has any ideas or especially critiques. So definitely open to critiques and then being able to make this a better project.

Biological damage from aging

So the reason that we're thinking at the tissue level of replacing brain tissue is that really the damage that accumulates with age occurs largely outside of cells. So a lot of the focus on cellular processes with the epigenome, with mitochondria, with genetic reprogramming, et cetera, you know, might make the cells perform better and the organs overall do better. But the bulk of aging that's occurring outside of cells is complex. It's stochastic and it not enzymatically driven and it cannot be enzymatically fixed. There are no genes encoded in our genomes that can address this damage.

Some examples are shown here on this slide just for proteins. There's also carbohydrates and DNA that accumulate complex damage. And so there's no way for at least the extracellular protein damage that the cell can repair that. And without that environment, the cell behaves as an old cell, even if it's young.

In heterochronic transplants, where you take young cells, put them in an old brain, they behave like old cells and vice versa. If you take old cells out of the brain and leave them in culture, neurons, for example, they can outlive the age of the organism. Or if you put them in a young environment, then they'll do better as well. So I think it's really important to think of if we want to beat aging, to do that at a tissue level, tissue or greater level.

So there are other people working on replacing, you know, other parts of the body, either as a whole or progressively with different parts as you need them. But in any case, those are not what my lab works on. But in conjunction with those, I think there's a lot of promise here for actually beating aging. So is brain tissue replacement possible without a discontinuity of self or function, right? Because that's important. The whole point is to preserve who we are and, you know, continue enjoying life? So I think the answer is yes. And I'll give you the reasons for that.

We're focused initially on the neocortex, which is the outer part of the brain. It's the biggest part. It's arguably the most important. It encodes our long-term memories, cognition, highest cognitive functions, consciousness. So, you know, if we can't fix that, then what's the purpose of working on anything else really?

Principles

The two principles that support that progressive neocortical tissue replacement is possible is, first, plasticity. Memories and functions can seamlessly move from one part of the neocortex to another. And I given this example before but it only one of many examples where language something that very dear to us as a function can move from one area to another if that initial area is slowly destroyed over time from a pinpoint out in the case of a benign brain tumor, for example. So these individuals, you know, eventually have the tumor removed, they have a big hole in their brain, but they never lost the ability to speak. And no one ever noticed any difference. They didn't notice a difference. So they're, you know, they're perfectly fine. And then when they go look for where speech is encoded, it's somewhere else in the neocortex, not always the same place, but it has moved. So this is a level of plasticity that tells us that we can replace tissue in the neocortex without losing continuity of self.

The other sort of established principle that supports progressive brain tissue replacement as a means of beating aging is that young neocortical precursors, when you transplant them into the adult environment, don't seem to care that they're in an adult environment and follow their innate developmental program and do remarkably well, even in the adult brain, to wire correctly, even to distant targets in the brain. So this is something that's been reproduced by many labs. We've done some work with this as well, both with mouse-into-mouse graphs, but also human-into-mouse graphs. So here's a human example where we have the human graft in green here. And it projects its axons in the correct orientation down to this major tract here, the corpus callosum, which you can see is bright green. And that's because of all the projections coming from this new brain tissue going along these tracks. And if we look further at this more distant targets in the brain, we see that those neurons are innervating those areas. And I'll give you a little bit more detail on that in a bit. But this is just to say that you know these cells these precursors of the brain have this established potential to generate the connections potentially that we want in the brain. And so you know I mentioned other groups have done this. These are examples of high profile paper in this case where they all use human precursor cells, neocortical precursor cells that they transplanted into adult rodents. And all showing all these pretty amazing properties. And many of these are from different labs, so highly reproducible.

So, you know, the tumor example I gave with the plasticity, if done right, means that we can remove old tissue. And the example I gave you and these other examples from other labs show that you can add new, potentially new tissue. Certainly, the cells seem to do very well in the adult if they're immature precursor cells. So we should be able to add new tissue that can encode information that's useful to the host and allow function to transfer, allow self-identity to transfer without us even noticing. But we're not there yet.

What's missing?

Despite those very nice examples where connectivity of transplanted cells is remarkably good, it's not good enough. There are things missing. One thing that's missing is these grafts typically don't contain the full complement of cell types that are normally found in the neocortex, and all the cell types are there for a reason. They're all doing something important for function, and without them, the tissue can't function normally. So that's one thing is they're missing cell types.

Another thing is the organization of the cells is very abnormal. So if you look at, you know, one of these acclaimed successes of transplanting organoids into the brain, you can see that the tissue is very different than what you have in a normal tissue where you have this very nice layered structure. That is not recapitulated. Maybe in micro locations it is, but that's not good enough to get normal function because the normal lamination is important to get normal wiring within the graft. So that's two things that are missing.

Our goal is to address these missing aspects of grafts so far to make grafts that will develop into a normal structure and will have the full complement of cells. And to do that, we don't want to reinvent the wheel. We're basing our approach on how our neocortex neocortex normally developed from a relatively simple fetal tissue at around gestational week six, where there's a limited number of precursor cell types in a very well-defined organization that we want to recapitulate in a prototissue that we will then use to graft. And then with anticipation that this tissue will develop the correct layers and as already shown for disorganized tissue will project normally outside of the graft as well.

Our platform

This is an example of a couple of key aspects that have been missing from all grafts so far. In green here there's a small number but a very important pioneer neurons called Cajal-Retius neurons. They're very important for setting up the architecture and the wiring of the neocortex, the layers that come afterwards. Also very important are a stable radial glial fiber attachment to an overlying layer here that's not shown, but I'll come back to that. So these fibers guide new neurons. So these radial glial fibers are part of the radial glial stem cells that are not only stem cells, but by virtue of these fibers act as scaffold for the new neurons to migrate and form these layers in the neocortex.

So here's sort of where we're at with this project. We've, this is again the prototype tissue that we, or the proto-tissue that we want to make. It has three layers. This PIA layer, this marginal zone, ventricular zone, doesn't matter too much for your purposes what they're called. But they have relatively few numbers of precursor cells in them and all these cells can be derived and have been derived by us or our collaborators from human IPS cells except for one we're still working on and then we can also purify all these cells again except for the one that we haven't managed to derive from IPS cells yet. Sorry about the noise if you can hear it.

It's not just the cells that are important here it's also the extracellular environment that are in to make sure that these cells kick off the right developmental program from the start. So we're doing proteomics to identify the extracellular components that surround these cells in each of the different layers. So this is an example of the transcriptomic, single cell transcriptomic analysis that we do to compare our IPS derived neocortical precursor cells. This is for the principal neuronal lineage and we compare that to real human fetal neocortical tissue and find that we get a very good match. And we do this for all the cell types, again, just as sort of a quality control.

We've also established a platform, which we've published now, showing that we can get, you know, the major features that we want in testing these proto tissues in the mouse, in the adult mouse. We can create these very reproducible lesions that we can fill with our precursor cells and biogel. And then we find that all the cell types that we care about survive in this graft. And the neurons integrate, they send projections along normal axonal pathways like the corpus callosum, the major fiber tract that connects the two hemispheres, down to subcortical targets like the striatum. More importantly, they also seem to mature normally. Initially there's very little activity but over time these human neurons start to show action potentials as they normally would And moreover if we put the graft in the visual cortex over time the neurons will respond to light. So this, hopefully you can see this little orange arrowhead here is the onset of the light stimulus for these mice. And you can see that the neurons in the graft initially after four weeks are weakly responsive, but then become more robustly responsive over time.

We also, very importantly, have vascularization of the graft. So we include vascular endothelial cells in our grafts. Here they're labeled with GFP, so we can see them in green. They rapidly formed vessels. Those vessels fuse with invading host vessels, and you can see an example of this anastomosis here. And if we put a circulating dye in the animal that goes throughout the whole system, we can see that that dye goes through these green vessels as well, not shown here, but I may have shown that previously.

So we have this platform that's pretty good. We can also layer cells in this platform. Again, these are not the layers that we want yet. They're just a proof of concept that if we use a biogel that's virtually available, very popular one that supports neural survival, we can layer green GFP labeled cells and then overlay, I think in this case, It's TD tomato labeled cells, and this border is maintained, but blood vessels can still cross, as do neuronal projections, which is the properties that we would want.

So we have this platform now that we can test with, you know, tissues that are more and more resembling the normal fetal tissue. But for that, we have a little ways to go. We can isolate pretty much all the cells, as I mentioned, miss it minus one. We've made quite a bit of progress on identifying key components in the extracellular matrix to make the biogels that these cells are going to initially be embedded in. And then of course there going to be a fairly extensive phase of testing whether the tissue develops into these mature layers neural layers and whether it can encode functions.

Information preservation and re-encoding

We have outlined experiments that will prove that the neuronal function of the graphs, the electrophysiological activity, is encoding useful information to the host. And then if we can do that, then, you know, But we might need to do a few iterations before we can do that. So that would fall into the bigger project here, a bigger phase that comes afterwards, where we're not only adding new tissue, but we have to remove the old tissue. So we're going to want to mimic what I showed you as an example with the benign glioma that grows slowly from a pinpoint out and destroy it. destroys different functional areas. Of course, we don't want to use gliomas because when benign, there's a risk that they will become glioblastomas, and that's definitely not worth the risk.

Optogenetic silencing and re-encoding

So there are other ways that we can do this that we've designed on paper, but, you know, these have to be developed in reality where we use an optogenetic approach to start silencing areas from the pinpoint out. And as long as the individual is using those functions, they will get re-encoded elsewhere and they won't notice a difference. And then once that tissue is electrophysiologically dead, we can remove it. So that's shown here with the silencing of this tissue.

So this is our prototype tissue that we add in. The brain shrinks with age, so we can add some tissue without removing anything initially. But then we have to remove the old tissue and that's what's shown here with the method I just described. And then once that's removed, we can add the new prototype tissue. It will also grow in, and we repeat this, and in a few iterations, we have a completely young neocortex again, and this would be combined if we wanted to really beat brain aging, which we really do want to do, but don't have the resources to be working on these other parts of the brain yet. We would want to do this in combination with other parts of the brain, particularly the ones outlined here, so we need more people working on this, basically. And we already have a lot of collaborators and a lot of people in the lab who have worked on this basically. I don't know if Sanjana is in the audience but apologies you know I didn't update this acknowledgement slide but Sanjana has been a big help here as well.

Yeah and also to try to accelerate things here's a shameless plug for a company (BE Therapeutics) I've started where we're basically focusing on trying to develop this tissue initially for to treat patients, but it would be applicable to addressing aging and combined with replacement of the rest of the body.

End

You know, I think that we can in the foreseeable future actually beat aging, not just live, you know, healthier lives as we nevertheless age, but actually beat aging and then decide, you know, when we've had enough or keep going as long as we want. Okay. So let's see. I guess I can stop sharing and hopefully I left some time for questions. Okay. Dr. Robert, fantastic, exciting, and provocative as ever. Thank you very much. I am so delighted we started funding your work.

Q&A

Q: Unsurprisingly we have a whole bunch of questions so the first one comes from Dr. Sharma who's an immunologist and therefore he asks an immunological question he says the old brain tends to have senescent cells such as in the glia and also infiltration of T cells. How do you think their presence might affect the engraftment of the new tissue?

A: Yeah I you know totally valid question. I mean, we have to go with what we know so far, which is even with naked cells transplanted in humans to treat Parkinson's, for example. So, you know, back in the 80s in Sweden, they started using dissociated fetal cells from midbrain to treat Parkinson's. And those you know in postmortem analyses showed no signs of the no pathogenic signs for at least a decade and a half And then in one case they had a patient that lived where they collected the brain 24 years after the transplant And those were starting to show pathology. And so, and we know these brains are inflamed, but those neurons were still functioning quite normally tell for a long time, you know, decades, and those were naked cells. So our, you know, expectation is that cells that are in a new tissue, so they're not just naked cells, will be first more insulated from any, you know, negative effects of the old brain environment, but also that, you know, this is over time, we're going to replace all the parts of the brain so that it will be completely young. So I think our tissue will certainly do well enough for long enough for that process to occur. That's the expectation. Thanks.

Q: I think you have a more optimistic read on those Parkinson's transplant studies than I do, but We couldn't have that argument the other day. I think he was actually asking a question that was more proximate, which is like not will the cells like survive for decades once they're implanted. But when you initially are trying to implant these, what do you think is the effect of the presence of, you know, in aging? There is already infiltration of T cells. There is already senescent cells, et cetera. if you're trying to do this in an aged brain, how do you suppose that's going to affect the initial engraftment in the first place?

A: Yeah, again, you know, we have not done enough in mice, but, you know, maybe more has been done in humans already with these, you know, examples of Parkinson's transplants. These are not young people that are getting these cells, and they seem to be surviving and doing quite well. This is a very recent example from Blue Rock Therapeutics where they have these wonderful dopaminergic precursors. They're in phase one, two trials, and they're getting positive effects, not just for safety but efficacy After I guess now it like 10 months So it still early stage But you know I think there is reason to be optimistic but I'm very happy to discuss potential challenges and caveats.

Q: Another thing I would add on just the immune side of things is, of course, that if one was working with IPS cells, that eliminates the allogeneic concern. So that actually was a question that I think addresses one of Yaffe's question right there, which is if you start with IPS cells, which is actually what you're doing in your animal experiments, for the most part, you avoid the allogeneicity problem. He also asks, when considering the aging brain, it's apparent that the entire organ may be affected, how can transplantation of a portion of brain tissue to enhance cognitive function? And I think this question sort of answers itself, but so let me ask it a bit more broadly than he's asked it. It's imperative that the organ may be affected. How are you going to handle areas outside of the neocortex where you're focusing?

A: Yeah, that's a very good question too. So, undoubtedly, the whole brain is affected, and undoubtedly, every part of the brain is important for function. And so we need to address all parts of the brain. Again, we're going to be, I think, guided, although again, we haven't started for the other parts of the brain, but I believe we're going to be guided by normal development. And so, for example, the visual thalamic would be replaced at the same time unilaterally as the primary visual cortex, for example, so that they wire together the same way they would during development. But yes, like, you know, you could replace the whole neocortex, but then you'll get a stroke in some subcortical structure and effect function or, you know, or just things will stop working. And so you don't get, you know, proper information flow from the outside world to your conscious cortex. So yes, yeah, those all have to be addressed. So this is a big, this is a monster project Yeah I don make it sound like it was simple. But the the takeaway too is that the result is a monster result too I mean I think you can beat it right. So that's, you know, you got to weigh the pros and the cons here. Yeah, yeah. Well, and of course, you had to start somewhere, literally. I mean, you had to start somewhere. And so like the neocortex is a very high value target, to say the least. So yeah, but absolutely a perfectly intelligent question. Like as you say, if you get a massive damage in the subcortical structure, it's not gonna matter if your cortex is functioning.

Q: I'm gonna keep asking questions while someone throws me off. So since that, Jason Zau asks, since the brain will continuously atrophy over time and each surgery will replace only a small percentage of the brain, what is your best estimate of how often brain tissue replacement surgeries need to be done in order to keep the brain healthy indefinitely?

A: So I think we could replace the entire brain with like a half dozen to 10 surgeries, which sounds like a lot, but again, you know, that's not something that's not doable. And surgical techniques are getting better and better. I haven't been keeping up with surgical robots, but they're making a lot of progress a lot faster than I thought. So, and that reduces complications of surgery. So there is risks with surgeries, of course, but I think those will be diminishing and a little more controllable. So yeah, six to 10 surgeries over the course of 20 years. And then you basically have, you know, a young brain again. I don't want to say like a teenager brain again, because that'll scare a lot of people, but you know, you have a young brain. That will last for a decade um and then you don't have to do anything for you know a long time. Yeah I will also add from the senescence perspective that once you've done that you can slow down the further um progression of damage by doing all the other damage repair strategies that SENS has so once you have a a nicely pristine young brain in there it's going to be a lot easier to maintain it by doing things like removing amyloid and alpha-synucleon aggregates and direct molecular level repair of the extracellular matrix. But if you've nucleon aggregates and direct molecular level repair of the extracellular matrix. But if you've gone and replaced it wholesale as the way you're doing, that's a very, very systematic approach, to say the very least.

Q: If I'm still allowed to ask you questions, Laura Lynn asks, are there any concerns about the invasiveness of multiple brain surgeries?

A: Yeah. So I mean, I kind of addressed that a little bit there. And I think Maria's appearance here is telling us. You can answer that question and then, yeah, we have to wrap. It is. There are complications from brain surgeries. Not many deaths, because, you know, you can actually do away with a lot, a big part of your brain without dying. But there are complications and those, you know, we'd want to avoid those. And so, you know, surgical techniques that are minimizing that are being progressively improved upon. And at some point, I think you know the risk will be low enough in a short time i mean these are ai driven smart robots uh and i think progress there will be will be that that'll be ready by the time we're ready to go super you've got a another good question and a good comment and another question i'd like to ask you some time but i will get off and people can perhaps approach you privately if you're hanging around and thank you as always for a great presentation dr a there Thank you.

Summary (single paragraph)

Dr. Jean Hébert (Albert Einstein College of Medicine) presented a multi‑phase strategy for progressive neocortical tissue replacement as a route to arrest and reverse brain aging. He argued that most age‑related damage is extracellular and therefore inaccessible to intracellular repair mechanisms; consequently, replacing whole‑tissue units is required. Evidence of cortical plasticity (e.g., functional preservation after tumor resection) and the robust integration of immature neocortical precursors—both mouse‑and‑human‑derived—demonstrates that new tissue can adopt existing circuitry without disrupting self. Current grafts, however, lack the full complement of cell types and proper laminar organization, limiting functional recovery. Hébert’s lab therefore engineers a “proto‑tissue” that recapitulates the six‑week fetal neocortical scaffold (including pioneer Cajal‑Retzius cells, radial glial fibers, and appropriate extracellular matrix) using IPS‑derived precursors and defined biogels. In adult mouse hosts this construct supports vascularization, lamination, long‑range axonal projection, and activity‑dependent responses. The long‑term vision combines staged removal of senescent tissue (via optogenetic silencing and precise ablation) with implantation of successive proto‑tissues, potentially requiring 6–10 surgeries over two decades to maintain a youthful neocortex, and ultimately to be coupled with systemic anti‑aging interventions.

Reformatted Transcript

2.1 Introduction & Acknowledgements

Speaker: Dr. Jean Hébert – Extramural researcher, Albert Einstein College of Medicine. Topic: “Progressive Brain Tissue Replacement and How We Beat Aging.”

• Thanks to the SENS Research Foundation for funding and for providing graduate trainees (e.g., Sanjana).
• Overview: focus on brain‑tissue replacement (particularly the neocortex) as a strategy to combat brain and systemic aging.

2.2 Rationale for Tissue‑Level Intervention

1. Extracellular Damage Dominates Aging
2. Age‑related alterations (protein cross‑links, glycated carbohydrates, extracellular DNA damage) accumulate outside cells.

3. No genomic instructions exist to enzymatically reverse this damage.

4. Cell‑Centric Approaches Are Insufficient

5. Even when intracellular pathways (epigenetic, mitochondrial, genetic reprogramming) are optimized, an aged extracellular milieu forces cells into an “old” phenotype.

6. Heterochronic transplant studies: young cells placed in old brains behave like old cells, and vice‑versa.

7. Tissue‑Level Replacement as a Solution

8. By renewing the extracellular environment together with the cellular constituents, we can break the feedback loop that enforces cellular aging.

2.3 Proof‑of‑Concept Evidence

2.3.1 Cortical Plasticity

• Clinical cases of benign brain tumor resection: patients retain language function despite removal of the original speech cortex.
• Demonstrates that memory and functional representations can migrate to adjacent cortical areas, preserving continuity of self.

2.3.2 Integration of Immature Precursors

• Mouse‑into‑mouse and human‑into‑mouse grafts of neocortical precursor cells show:
• Correct axonal orientation toward major tracts (e.g., corpus callosum).
• Long‑range projections to subcortical targets (e.g., striatum).

• Reproducibility across independent laboratories.

• These data confirm that immature precursors follow innate developmental programs even in an adult brain.

2.4 Limitations of Existing Grafts

2.5 Proposed “Proto‑Tissue” Engineering Approach

1. Developmental Blueprint

2. Emulate the fetal neocortex at ~GW 6: a simple, well‑defined set of precursor populations organized into three nascent zones (marginal zone, subventricular/ventricular zones, etc.).

3. Key Cellular Components

4. Cajal‑Retzius (CR) pioneer neurons – secrete Reelin, establishing cortical layers.
5. Radial glial fibers – scaffold for neuronal migration and serve as a stem cell pool.

6. IPS‑derived precursors for all major neuronal and glial lineages (except one still under development).

7. Extracellular Matrix (ECM) Specification

8. Proteomic profiling of fetal ECM to formulate biogels that provide the correct biochemical cues for each layer.

9. Quality Control via Single‑Cell Transcriptomics

10. Compare IPS‑derived cell signatures against bona‑fide fetal neocortical cells to ensure fidelity.

2.6 In‑Vivo Platform & Early Results

• Lesion‑fill Model: Reproducible cortical lesions in adult mice receive a mixture of proto‑tissue cells + biogel.
• Survival & Maturation: All targeted cell types survive; human neurons develop action potentials over weeks.
• Functional Integration:
• In visual cortex grafts, neurons become progressively responsive to light stimuli (weak at 4 weeks, robust later).
• Vascular endothelial cells (GFP‑labeled) form perfused vessels that anastomose with host vasculature; circulating dye confirms patency.
• Layered Graft Demonstration: Using a common biogel, distinct GFP‑ and tdTomato‑labeled cell populations maintain a visible interface while allowing axonal and vascular crossing.

2.7 Roadmap Toward Full‑Scale Replacement

1. Iterative Tissue Maturation – Refine ECM, add missing cell types, achieve proper lamination.
2. Functional Validation – Electrophysiology, behavioral assays, and encoding of host‑relevant information.
3. Selective Removal of Senescent Tissue
4. Optogenetic silencing of targeted cortical patches (pinpoint out) to force functional re‑encoding elsewhere.

5. Subsequent surgical excision of electrophysiologically silent tissue.

6. Cumulative Replacement Schedule

7. Approx. 6–10 staged surgeries over ~20 years to replace the entire neocortex, leveraging improving robotic neurosurgery and minimally invasive techniques.

8. Integration with Systemic Anti‑Aging – After achieving a youthful neocortex, combine with extracellular matrix repair, amyloid/α‑synuclein clearance, and other molecular rejuvenation strategies.

2.8 Q&A Highlights

Core Intuitions, Mechanistic Insights

• Extracellular Damage as the Primary Aging Driver – Shifts focus from intracellular rejuvenation to tissue renewal.
• Cortical Plasticity Guarantees Continuity of Self – Functional maps can relocate, allowing tissue removal without perceptible loss.
• Immature Precursors Are Environment‑Agile – Developmental programs dominate over adult milieu, enabling proper wiring after transplantation.
• Recreating the Fetal Scaffold – Inclusion of CR neurons and radial glia provides the necessary laminar cues absent in organoid grafts.
• Biogel Engineering Based on Fetal ECM Proteomics – Tailors the physical‑chemical niche to promote correct migration and layer formation.
• Optogenetic “Pinpoint‑Out” Silencing – Forces functional re‑encoding before surgical excision, minimizing behavioral deficits.
• Iterative, Modular Replacement – Treats the brain as a replaceable organ system rather than a monolithic entity, making the problem tractable.

Difficulties Encountered & Transcript Clarifications

Overall, the transcription was largely accurate, with only minor misspellings, occasional missing articles, and a few speech‑to‑text artefacts (e.g., “citizen cells,” “graph” for “graft”).