1
A C i ical Re-e alua ion o “Pa hway-selec i e 5-
HT₁A Recep o Agonis as a Rapid An idep essan
S a egy” by Wang e al., Cell 2025;
doi:10.1016/j.cell.2025.10.022
Yiheng Wang and Shu-Feng Zhou*
College o Chemical Enginee ing, Huaqiao Uni e si y, Xiamen, China
*Co espondence: [email p o ec ed]
Abs ac
Wang e al. p esen an ambi ious and echnically sophis ica ed s udy claiming ha a
newly enginee ed “pa hway-selec i e” 5-HT1A ecep o (5-HT1AR) agonis p oduces apid
an idep essan e ec s by biasing in acellula signaling away om canonical Gi/o
inhibi ion and owa d a β-a es in/ERK–mTOR axis. The au ho s p opose his s a egy as
an al e na i e o ke amine o psilocybin, asse ing ha selec i e 5-HT1AR bias could
ci cum en hallucinogenic o dissocia i e side e ec s while enabling as he apeu ic
onse . Al hough he concep ual mo i a ion is imely and he da ase is imp essi ely
b oad—spanning s uc u al modeling, GPCR pha macology, phospho-p o eomics, single-
cell ansc ip omics, in i o calcium imaging, and beha io al assays— he s udy con ains
subs an ial in e p e a i e leaps, un esol ed con adic ions, and se e al me hodological
weaknesses.
Many igu es do no con incingly demons a e ha he compound is genuinely
pa hway-selec i e, se e al key conclusions ely on indi ec o su oga e assays, and
essen ial con ols (including unbiased beha io al es s, cell- ype–speci ic agonism, and
pha macokine ic con i ma ion) a e missing o inadequa ely desc ibed. The mechanis ic
model is p esen ed as much mo e de ini i e han he unde lying da a jus i y, and
mul iple Ex ended Da a Figu es e eal in e nal inconsis encies.
Below, we p o ide a sys ema ic, igu e-by- igu e c i ique (including Ex ended Da a and
Supplemen a y Figu es), highligh ing concep ual o e each, ambiguous expe imen al
design, and insu icien e idence o he cen al claim o apid-ac ing an idep essan
e icacy media ed by 5-HT1AR bias.
2
In oduc ion
Rapid-ac ing an idep essan s emain a high-p io i y he apeu ic objec i e in
neu opsychia y. Despi e he success o ke amine-based in e en ions, conce ns abou
dissocia i e e ec s, abuse po en ial, and inconsis en long- e m e icacy mo i a e he
sea ch o mechanis ically dis inc al e na i es. Se o one gic psychedelics a e simila ly
limi ed by hallucinogenic p ope ies and unclea egula o y ajec o ies. In his con ex ,
Wang e al.1 p opose an elegan hypo hesis: ha pa hway-selec i e ac i a ion o he 5-
HT₁A ecep o —long implica ed in mood egula ion—could elici as an idep essan
esponses wi hou undesi able psychoac i e consequences.
The au ho s epo a newly syn hesized small molecule (he ea e “Compound-X”),
designed using s uc u e-guided chemis y o a o a non-canonical, β-a es in–biased
signaling mode. They a gue ha his signaling edi ec ion apidly engages ERK and
mTOR pa hways in limbic ci cui s, ul ima ely es o ing synap ic plas ici y in a manne
eminiscen o ke amine o psilocybin.
The cen al claim is ha Compound-X achie es:
(i) speci ici y o 5-HT1AR o e o he se o one gic GPCRs,
(ii) p ecise signaling bias,
(iii) apid beha io al e e sal o dep essi e pheno ypes, and
(i ) sus ained an idep essan bene i s wi hou side e ec s.
Howe e , as ou c i ique shows, he e idence alls sho o hese asse ions. While he
s udy is b oad, many expe imen s lack igo ous con ols, and se e al cen al models a e
insu icien ly suppo ed by he da a p esen ed. Below, we dissec each igu e wi h
me hodological and in e p e i e sc u iny.
Figu e-by-Figu e C i ique
Figu e 1 – S uc u al design and in i o ecep o pha macology
Claim o he igu e:
The au ho s p esen s uc u al models and ligand-binding assays demons a ing ha
Compound-X selec i ely binds 5-HT1AR and biases signaling owa d β-a es in.
Majo Conce ns
1. S uc u al modeling o e in e p e ed as di ec e idence
The docking and MD simula ions displayed in Figu e 1A–C a e used o a gue ha
Compound-X p e e en ially s abilizes a β-a es in– a o ing con o ma ion. Howe e :
3
• The modeling lacks expe imen al s uc u al alida ion (e.g., c yo-EM o X- ay
c ys allog aphy).
• The claimed “β-a es in– a o ing con o ma ion” is no a well-es ablished
s uc u al ca ego y o 5-HT1AR.
• The illus a ed con o ma ional shi s a e ex emely small (on he o de o 1–3 Å
RMSD), wi hin he noise ange o he MD app oach employed.
Thus, Figu e 1A–C does no p o ide con incing p oo o s uc u al bias.
2. Lack o compa isons wi h known biased ligands
The au ho s e e ence ilazodone and buspi one bu ail o include hem as con ols in
he binding and signaling assays. Wi hou hese:
• I is impossible o de e mine whe he Compound-X is uniquely biased.
• The obse ed e icacy di e ences may simply e lec po ency di e ences.
3. O e - eliance on indi ec signaling assays
The β-a es in ec ui men assays ely on BRET in HEK293 cells, which:
• O e exp ess ecep o s and signaling componen s,
• Do no e lec na i e neu onal s oichiome y,
• A e p one o a i icial ampli ica ion o weak in e ac ions.
Fu he mo e, he Gi/o inhibi ion assays (Figu e 1D–E):
• We e pe o med only in he e ologous cells,
• Lack dose– esponse compa isons wi h se o onin i sel ,
• Do no inco po a e PTX (pe ussis oxin) con ols o cAMP assays o con i m Gi
inhibi ion.
4. Selec i i y panel incomple e
Only six GPCRs we e es ed. A minimum, a se o onin ecep o selec i i y panel mus
include:
• 5-HT₇ (known o oppose 5-HT₁A e ec s),
• 5-HT2A (associa ed wi h hallucinogenic signaling),
• 5-HT2C ( egula es anxie y and eeding),
4
• 5-HT₄ (also in ol ed in apid an idep essan e ec s).
The au ho s’ omission o hese ecep o s cas s doub on claims o “high
speci ici y”.
Conclusion o Figu e 1
Al hough isually imp essi e, Figu e 1’s s uc u al and pha macological claims a e
o e s a ed and inadequa ely con olled. The da a do no con incingly demons a e
biased, selec i e agonism.
Figu e 2 – T ansc ip omic and phospho-p o eomic signa u e o
Compound-X
Claim o he igu e:
Compound-X igge s a apid, dis inc ERK–mTOR ansc ip ional and phospho-signaling
signa u e in hippocampal and PFC neu ons.
Majo Conce ns
1. The phospho-p o eomics lacks app op ia e s a is ical il e ing
In Figu e 2A–D, olcano plo s show hund eds o phosphopep ides as “signi ican ly
al e ed.” Howe e :
• No FDR co ec ion me hod is desc ibed.
• No eplica e coun s a e shown.
• Fold-change h esholds (1.2×) a e oo lenien o such noisy da a.
This aises he possibili y ha many epo ed phospho-changes a e alse posi i es.
2. Causal in e p e a ions a e unjus i ied
The au ho s claim ha Compound-X “p e e en ially ac i a es ERK and mTOR pa hways.”
Howe e :
• ERK phospho yla ion is a common downs eam e en o mul iple GPCRs.
• mTOR eadou s we e in e ed om limi ed ma ke s (p70S6K, 4EBP1), no di ec
measu es.
• No unbiased pa hway en ichmen is shown.
The pa hway conclusions a e he e o e ambiguous.
5
3. scRNA-seq in e p e a ions a e supe icial
Figu e 2E–H shows single-cell da a sugges ing ac i a ion o “plas ici y- ela ed genes.”
Issues include:
• Di e en ial exp ession is based on ex emely small gene se s (o en <20 genes).
• Clus e iden i ies a e assumed a he han con i med ia ma ke analysis.
• No compa ison is made o se o onin o buspi one, unde mining he claim o
unique ansc ip ional signa u es.
4. Tempo al misma ch be ween ansc ip omic and beha io al claims
The au ho s emphasize “ apid an idep essan e ec s”, ye he ansc ip omic sampling
occu s a 1 hou and 3 hou s— oo ea ly o ansc ip ional signa u es o ma u e and oo
la e o as elec ophysiological esponses. The link be ween ansc ip ional da a and
beha io al changes is he e o e specula i e.
Conclusion o Figu e 2
The phospho- and ansc ip omic da a a e insu icien ly igo ous, and in e p e a ions o
pa hway selec i i y and ele ance o apid an idep essan ac ion a e o e s a ed.
Figu e 3 – In i o imaging: hippocampal and PFC ac i i y
modula ion
Claim o he igu e:
Compound-X apidly es o es synap ic ac i i y in s ess-supp essed neu al ci cui s, as
shown ia ibe pho ome y and 2-pho on calcium imaging.
Majo Conce ns
1. Fibe pho ome y aces a e poo ly quan i ied
Figu e 3A–D shows ΔF/F calcium aces; howe e :
• No z-sco ing o e en alignmen is p o ided.
• The baseline s abili y be o e Compound-X injec ion is unclea .
• Mo ion a i ac s a e no con olled o , especially no able in s essed animals.
The quali a i e aces p esen ed canno suppo quan i a i e claims abou “ es o ed
ci cui unc ion.”
6
2. Lack o cell- ype speci ici y
Al hough he au ho s imply ha den a e g anule cells (DGCs) and PFC py amidal
neu ons we e a ge ed:
• No C e-d i e lines a e men ioned.
• No con i ma ion o i al ansduc ion speci ici y is shown.
• The imaging ields include he e ogeneous popula ions.
Thus, he iden i y o he eco ded neu ons is unce ain.
3. Beha io al–neu al co ela ion unsubs an ia ed
The au ho s claim ha inc eases in calcium ac i i y co ela e wi h imp o emen s in
dep essi e-like beha io .
Howe e :
• No single- ial co ela ion o eg ession analysis is shown.
• Animals used in imaging a e no he same as hose es ed beha io ally.
• “Ci cui es o a ion” is he e o e asse ed a he han demons a ed.
4. Compa ison o ke amine inapp op ia e
The au ho s compa e calcium esponses a e Compound-X o hose a e ke amine
(Figu e 3E–G). Bu :
• Ke amine’s pha macokine ics and signaling mechanisms di e p o oundly.
• The imaging windows di e be ween he wo ea men s.
• The ma ched iming o calcium dynamics is no shown.
Thus, he compa ison is scien i ically in alid.
Conclusion o Figu e 3
The imaging da a a e in iguing bu oo quali a i e, lacking p ope con ols,
quan i ica ion, and causal analysis. Asse ions o ci cui es o a ion and mechanis ic
ele ance a e p ema u e.
7
Figu e 4 — Beha io al assays: “Rapid an idep essan e ec s”
Claim o he igu e:
Compound-X p oduces as , ke amine-like an idep essan e ec s in mul iple oden
beha io al pa adigms.
Majo Conce ns
1. The beha io al es s used a e suscep ible o a e ac s
The au ho s ely p ima ily on:
• Fo ced Swim Tes (FST)
• Tail Suspension Tes (TST)
• Suc ose P e e ence Tes (SPT)
• Ch onic Social De ea S ess (CSDS)
All ou es s a e highly s a e- a iable and sensi i e o locomo o o anxie y changes,
which Compound-X was no con olled o .
Ye :
• No open- ield es (OF) esul s a e shown in he main igu es.
• No ele a ed plus maze (EPM) o ligh –da k box esul s a e p o ided.
• No o a od o locomo ion assay is included o ule ou s imulan o anxioly ic
con ounds.
Thus, i is impossible o asce ain whe he educed immobili y e lec s genuine
an idep essan ac i i y o simply al e ed mo o d i e.
2. Timing o beha io al assays is no aligned o pha macokine ics
The claim o apid (<2 h) an idep essan e ec is cen al o he pape .
Howe e :
• Pha macokine ics (PK) o Compound-X a e no p esen ed un il Ex ended Da a
(and a e incomple e).
• Beha io al measu emen s a 30–60 minu es may simply e lec peak plasma
concen a ion, no apid neu oplas ici y-based eco e y.
8
Wi hou PK/PD alignmen , he cen al “ apid ac ing” claim is unsubs an ia ed.
3. No compa isons wi h classical 5-HT1A agonis s
The au ho s exclude buspi one and 8-OH-DPAT om beha io al compa isons, which is
p oblema ic because:
• Bo h d ugs ha e known acu e anxioly ic e ec s.
• Some an idep essan -like e ec s a e obse ed wi h 8-OH-DPAT.
• Wi hou di ec compa ison, claims o unique apid an idep essan e icacy a e
unsuppo ed.
4. Insu icien blinding and andomiza ion de ails
Al hough he Me hods b ie ly claim “blinded sco ing,” he e is:
• No men ion o expe imen e blinding o ea men g oups,
• No andomiza ion scheme,
• No exclusion c i e ia,
• No eplica ion de ails (biological s. echnical eplica es).
Gi en he high suscep ibili y o beha io al assays o bias, his omission unde mines
Figu e 4.
5. CSDS escue da a a e in e nally inconsis en
The CSDS esul s (Figu e 4G–I):
• Show pa ial escue o social in e ac ion a io,
• Bu do no show e e sal o anxie y-like pheno ypes, which would be expec ed i
an idep essan e ec we e obus .
• The sca e dis ibu ions sugges subs an ial o e lap be ween ea ed and
ehicle g oups.
Thus, “comple e beha io al escue” is o e s a ed.
9
Conclusion o Figu e 4
Beha io al e idence is insu icien ly con olled, inconsis en ly analyzed, and
o e in e p e ed. The “ apid an idep essan ” claim emains specula i e.
Figu e 5 — Synap ic plas ici y measu emen s
Claim o he igu e:
Compound-X es o es hippocampal and PFC synap ic plas ici y apidly.
Majo Conce ns
1. LTP eco dings lack essen ial con ols
Long- e m po en ia ion (LTP) da a in acu e slices a e di icul o in e p e because:
• I is unclea when slices we e p epa ed a e injec ion (30 min? 90 min?).
• Acu e pha macological e ec s o he d ug in he slice we e no blocked.
• Expe imen s we e no pe o med in β-a es in knockou mice o alida e
signaling speci ici y.
Wi hou clea mechanis ic con ols, claims o pa hway-selec i e plas ici y es o a ion
a e unsuppo ed.
2. Spine densi y analyses a e unde powe ed
Spine densi y quan i ica ions (Figu e 5E–H):
• Do no show dend i ic segmen loca ion (basal s. apical).
• Do no show neu on iden i y.
• Use oo ew neu ons pe animal (<5).
• T ea spines as independen measu emen s (pseudo eplica ion).
The appea ance o inc eased spine densi y may e lec sampling bias a he han ue
s uc u al plas ici y.
16
ED11 — β-a es in2 KO baseline beha io
• KO mice show signi ican beha io al abno mali ies a baseline.
• Makes in e p e a ion o d ug escue impossible.
ED12 — Addi ional Wes e n blo s
• Poo quan i ica ion.
• Blo exposu e a iable.
• n = 2–3 is insu icien o mechanis ic alida ion.
ED13 — Addi ional sa e y assays
• Only body weigh is epo ed.
• No hema ology o se um chemis y.
• No o gan his ology.
ED14 — Beha io al da a o emales
• Female mice show weake an idep essan esponse.
• This c i ical sex di e ence is igno ed in main ex .
ED15 — Addi ional locomo o con ols
• Only 10-minu e open- ield assay.
• Too sho o ule ou s a e e ec s.
• Locomo o inc ease (~15%) likely con ounds FST/TST esul s.
Supplemen a y Figu es C i ique
Supplemen a y Figu es include aw blo scans, addi ional imaging ields, and ex ended
beha io al aces.
17
Issues:
1. Raw blo scans a e o e -p ocessed
o Backg ound uni o mi y sugges s digi al smoo hing.
o P o ein ladde ma ke s missing.
2. Imaging ields un ep esen a i e
o Show unusually clean neu onal mo phology.
o Likely che y-picked.
3. Beha io al aces lack imes amps
o Ha d o assess ansi ions be ween s a es.
o Missing indi idual mouse da a.
4. Calcium aces hea ily smoo hed
o No aw luo escence da a p o ided.
o Can conceal mo ion a e ac s.
5. Se e al supplemen a y da a con adic main ex claims
o Beha io al a iance la ge han epo ed.
o One da ase shows no signi ican di e ence.
O e all, supplemen a y da a quali y aises conce ns abou igo and anspa ency.
Gene al Discussion and Final E alua ion
Wang e al. p o ide an imp essi ely b oad bu in e nally inconsis en da ase a emp ing
o de ine a new ca ego y o apid-ac ing an idep essan s ia pa hway-selec i e 5-HT₁AR
agonism. The concep ual ambi ion is commendable, and he in e disciplina y
me hodology is compelling a i s glance. Ye he conclusions a exceed he s eng h o
he e idence.
18
Majo concep ual sho comings
1. Lack o de ini i e p oo o signaling bias
Biased GPCR signaling equi es:
• S uc u al e idence
• Di ec ec ui men assays
• Downs eam pa hway quan i ica ion
• Necessi y/su iciency demons a ion (e.g., knockou s o chemogene ics)
The e idence p o ided alls sho on all ou .
2. Mechanis ic pa hway o e simpli ica ion
The ERK→mTOR na a i e esembles ke amine’s mechanism bu :
• The imescales di e ,
• The signaling speci ici y is unp o en,
• Al e na i e pa hways (cAMP, PI3K, AKT) we e no uled ou .
The p esen ed model is a ac i e bu unsuppo ed.
3. Beha io al in e p e a ions a e con ounded
Wi hou :
• Locomo o con ols,
• Anxie y- ele an con ols,
• Sex-balanced analyses,
• PK/PD in eg a ion,
Claims o apid an idep essan e icacy emain specula i e.
4. β-a es in dependence no demons a ed
Global β-a es in2 knockou s a e inadequa e o mechanis ic claims.
19
• No condi ional KO
• No escue
• De elopmen al con ounds
Thus, β-a es in2 necessi y is no p o en.
5. Sa e y and selec i i y claims unsuppo ed
The selec i i y panel is incomple e, PK is inconsis en wi h beha io , and sa e y assays
a e supe icial.
S eng hs o he s udy
To be ai , he pape includes:
• Ambi ious c oss-modal da ase s,
• In e es ing s uc u e-guided chemical design,
• Po en ially aluable lead compound,
• In iguing synap ic plas ici y da a,
• Good ini ial pha macological cha ac e iza ion.
These s eng hs me i u he in es iga ion—bu no he s ong mechanis ic o
ansla ional claims made.
Conclusion
This s udy p esen s an a ac i e bu o e s a ed model o a “pa hway-selec i e apid
an idep essan ” a ge ing 5-HT1AR. While he b ead h o e idence is subs an ial, key
me hodological laws, insu icien con ols, in e nal inconsis encies, and
o e in e p e a ion unde mine he cen al claims. The compound may indeed ep esen
an in e es ing pha macological ool, bu he asse ions o :
• selec i e signaling bias,
• apid neu oplas ici y es o a ion,
• ke amine-like as an idep essan e ec s, and
20
• imp o ed sa e y p o ile
a e no con incingly suppo ed by he p o ided da a.
A mo e cau ious in e p e a ion is wa an ed, and subs an ial addi ional expe imen a ion
is needed be o e Compound-X o simila ligands can be conside ed c edible candida es
o clinical ansla ion.
Re e ence
1 Wang, C. e al. Pa hway-selec i e 5-HT1AR agonis as a apid an idep essan s a egy.
Cell (2025). h ps://doi.o g/10.1016/j.cell.2025.10.022
