JRT "increased dendritic spine density by 46% and was nearly 100 times more potent than ketamine as an antidepressant"

#1

Researchers at the University of California, Davis, have developed a groundbreaking molecule that delivers the powerful brain-repairing benefits of LSD — without any hallucinations.

By simply swapping two atoms in the LSD structure (what lead researcher David E. Olson playfully called a molecular “tire rotation”), the team created JRT — a non-hallucinogenic “neuroplastogen.”

This innovative compound keeps LSD’s remarkable ability to stimulate neuronal growth and repair damaged synapses, while eliminating the psychedelic effects that make traditional LSD too risky for many patients.

The results are impressive: in preclinical studies, JRT increased dendritic spine density by 46% and was nearly 100 times more potent than ketamine as an antidepressant. Crucially, it avoided the gene expression patterns and behaviors linked to psychosis.

This could be a game-changer for psychiatric care — especially for people with schizophrenia and other conditions involving synaptic loss and brain atrophy, who were previously excluded from psychedelic-based therapies.

9 Likes
#2

This is game changing! How can we obtain it?

1 Like
#3
1 Like
#4

The hallucinations were the best part! :sweat_smile:

3 Likes
#5

We don’t know that until this molecule becomes available.

1 Like
#6

The main mapping is:

  • 5-HT2A: same gross Gi placement, different interface partition. ICL3 gains Galpha use (24.4 → 25.0), distal receptor C-tail loses Galpha (47.0 → 38.4), while that same C-tail gains Gbeta (19.4 → 24.0) and Ggamma (4.0 → 10.0). TM7/H8 drops slightly, but it is secondary to the ICL3 vs distal C-tail reassignment. The detailed writeup starts at line 43 (line 43).
  • 5-HT2C: even cleaner C-tail reassignment. The holo frame barely moves, but receptor-Galpha drops hard (206.6 → 163.8) while receptor-Gbeta/gamma rises. Galpha loses the distal C-tail (22.4 → 16.4), plus smaller losses at ECL2, ICL2, and the TM7-H8 linker; Gbeta and Ggamma pick up that same distal C-tail. The section is at line 109 (line 109).
  • 5-HT2B: not the same story. This one looks like a slight whole-core tightening, with R-A/B/C all moving inward by about 1-1.25 A, plus local rebalancing across TM5, ICL2, and ICL3, not a distal C-tail handoff. That section is at line 175 (line 175).
  • 5-HT1A: this no longer looks like “same docked state, different bookkeeping.” Galpha and Gbeta shift by about 9-10 A after receptor alignment, and both receptor-partner contact maps drop sharply. I kept this as a global holo-state divergence call because the trimmed 7E2Y construct does not support the same clean segment annotation as 5-HT2A/2B/2C. That section is at line 207 (line 207).

So the short geometric answer is: for 5-HT2A, “different partition of the same holo-state interface” specifically means ICL3 takes slightly more of the Galpha burden, while the distal receptor C-tail stops being as Galpha-dominated and becomes more Gbeta/gamma-dominated, with only minor TM7/H8 attenuation. For 5-HT2C, the same logic is even stronger and more C-tail-centric.

==

  • 450: 0.8 -> 0.2
  • 453: 0.8 -> 0.0
  • 456: 0.6 -> 0.0
  • distal C-tail gaining Gbeta1:
    • 409: 0.2 -> 1.0
    • 410: 0.0 -> 0.8
  • distal C-tail gaining Ggamma2:
    • 425: 0.0 -> 0.6
    • 427: 0.0 -> 0.6
    • 428: 0.0 -> 0.8

So geometrically, JRT on 5-HT2A looks like this:

  • ICL3 becomes a slightly more prominent Galpha contact surface
  • the distal C-tail becomes a less prominent Galpha surface
  • that same distal C-tail becomes a more prominent Gbeta/gamma surface
  • TM7/H8 participation is mildly reduced, but it is not the dominant story

This is not random noise. It is a specific reassignment of contact load from Galpha toward Gbeta/gamma, with ICL3 partially compensating on the Galpha side.

5-HT2C: the same theme, but even more explicitly a C-tail reassignment

5-HT2C shows the same general logic as 5-HT2A, but the segment that gets reassigned is even more obvious.

The whole-complex geometry still barely moves:

  • R-A: 36.093 -> 36.561 A (+0.468 A)
  • R-B: 47.890 -> 47.699 A (-0.191 A)
  • R-C: 54.158 -> 54.258 A (+0.100 A)

So again this is not a big rigid-body rearrangement.

But the contact redistribution is large:

  • receptor-Galpha: 206.6 -> 163.8 (-42.8)
  • receptor-Gbeta: 115.2 -> 136.8 (+21.6)
  • receptor-Ggamma: 6.0 -> 10.4 (+4.4)

The segment-level reallocation for receptor → Galpha_q_chimera is:

  • C-tail: 22.4 -> 16.4 (-6.0)
  • ECL2: 15.4 -> 14.0 (-1.4)
  • ICL2: 8.8 -> 7.6 (-1.2)
  • TM7-H8 linker: 2.0 -> 1.4 (-0.6)

The compensating gains sit mostly on the receptor C-tail but now on the Gbeta/gamma side:

5-HT2C receptor → Gbeta1

  • C-tail: 35.6 -> 39.2 (+3.6)

5-HT2C receptor → Ggamma2

  • C-tail: 4.0 -> 6.4 (+2.4)

The residue-level picture is extremely explicit:

  • C-tail losing Galpha_q_chimera:
    • 450: 0.8 -> 0.2
    • 451: 0.8 -> 0.2
    • 452: 0.8 -> 0.2
    • 453: 0.8 -> 0.0
    • 454: 0.8 -> 0.0
    • 455: 0.8 -> 0.2
    • 456: 1.0 -> 0.2
    • 457: 1.0 -> 0.2
    • 458: 1.0 -> 0.2
  • C-tail gaining Gbeta1:
    • 452: 0.2 -> 0.8
    • 453: 0.2 -> 1.0
    • 454: 0.2 -> 1.0
    • 455: 0.2 -> 1.0
    • 456: 0.2 -> 0.8
    • 457: 0.2 -> 0.8
    • 458: 0.0 -> 0.8
  • C-tail gaining Ggamma2:
    • 458: 0.0 -> 0.8

So on 5-HT2C, the simplest geometric interpretation is:

  • JRT keeps the holo complex in essentially the same gross active-state location
  • but the distal receptor C-tail is much less Galpha-loaded
  • and that same C-tail becomes more Gbeta/gamma-loaded

That is an even cleaner example than 5-HT2A of a conserved active-state frame with changed intracellular interface ownership.

5-HT2B: not the same story

5-HT2B does not look like pure interface repartition. It looks more like a small whole-core tightening plus local interface edits.

The receptor->partner distances all move inward by about 1-1.25 A:

  • R-A: 39.190 -> 37.997 A (-1.193 A)
  • R-B: 53.181 -> 51.934 A (-1.247 A)
  • R-C: 59.917 -> 58.915 A (-1.002 A)

So JRT is pulling the miniGq core slightly closer to the receptor as a whole, rather than just redistributing contact load at a fixed radius.

The main receptor → Galpha_q_chimera segment changes are:

  • ICL3: 26.2 -> 25.4 (-0.8)
  • ICL2: 13.4 -> 14.0 (+0.6)
  • TM5: 8.6 -> 8.0 (-0.6)
  • ECL2: 2.6 -> 3.0 (+0.4)

Residue-level changes reinforce that:

  • TM5 loses at 229 and 239
  • TM5 gains at 227, 242, and 244
  • ICL3 gains at 263 and 266, while losing at 267, 271, 272, and 277
  • ICL2 gains at 127

So 5-HT2B is better described as:

  • a modest receptor/miniGq tightening
  • accompanied by local rebalancing across TM5, ICL2, and ICL3
  • not primarily a distal C-tail reassignment story

5-HT1A: a genuinely different Gi arrangement, not just reallocation

5-HT1A is the one that really does look like a different holo arrangement rather than a different partition of the same one.

After receptor alignment:

  • Galpha_i1 centroid shift vs LSD: 9.457 A
  • Gbeta1 centroid shift vs LSD: 10.005 A

And the contact losses are large on both receptor-partner interfaces:

  • receptor-Galpha: 247.6 -> 198.8 (-48.8)
  • receptor-Gbeta: 76.6 -> 31.8 (-44.8)

The receptor->Galpha distance is not dramatically different in isolation (40.744 -> 40.050 A), but that is misleading by itself because the aligned chain centroids and total receptor-partner contact patterns say the assembly has reorganized substantially.

In other words:

  • 5-HT2A and 5-HT2C look like “same general active-state footprint, different internal contact bookkeeping”
  • 5-HT1A looks more like “different receptor/Gi arrangement in the aligned receptor frame”

The residue-level 5-HT1A deltas support that broader interpretation:

  • gains at 254 and 271 to full Galpha occupancy
  • losses at 260, 261, 265, 247, 248, and 267
  • gain at 294

But the segment annotation for the trimmed 7E2Y receptor-only construct is not clean enough in this pass to claim the same precision as the 5-HT2A/2B/2C segment maps. For 5-HT1A, the stronger statement is the global one: JRT is not simply using the same Gi arrangement as LSD.

Bottom line on the residue-level pass

The residue-level pass makes the earlier statement much sharper.

For 5-HT2A, “different partition of the same holo-state interface” means:

  • the receptor-centered Gi pose is nearly the same
  • the intracellular interface is re-clocked rather than translated
  • ICL3 takes slightly more of the Galpha burden
  • the distal receptor C-tail takes less of the Galpha burden
  • that same distal C-tail takes more of the Gbeta/gamma burden
  • TM7/H8 participation is mildly reduced, but it is secondary to the ICL3 vs C-tail redistribution

For 5-HT2C, the same logic is even cleaner:

  • the holo-state frame is still basically preserved
  • the distal receptor C-tail is reallocated away from Galpha
  • and toward Gbeta/gamma

For 5-HT2B, JRT looks more like a slightly tightened whole complex with local TM5/ICL2/ICL3 reshaping.

For 5-HT1A, JRT looks like a genuinely different receptor/Gi arrangement rather than a subtle repartition of the LSD-like one.

That distinction matters because it is exactly the sort of thing a single scalar score can hide. Two ligands can both look “active-state compatible” in headline terms while still distributing force, contact occupancy, and intracellular segment usage differently across the receptor/G-protein interface.

#7

I am curious… from what specific baseline and depression measures is JRT "nearly 100 times more potent than ketamine as an antidepressant? " Not sure how those potency numbers add up? It does look like a very interesting molecule and seems like a great add on to the class of possible medications such as DLX-001, TN-001, SPT-348 etc. Here’s hoping it is safe and effective when it gets out of the lab…