chatGPT(5.5paid):
The paper, “Cestode infection is linked to transcriptional shifts in neuropeptide signalling and caste-specific ageing pathways in a social insect” by Blasi et al. (BMC Genomics, 2026), asks how the tapeworm Anomotaenia brevis produces a remarkable phenotype in its ant host Temnothorax nylanderi: infected workers become less active and can live roughly as long as queens. The authors test whether the parasite is effectively pushing worker physiology toward a “queen-like” molecular state, and whether it might do so by manipulating neuropeptide signalling.
Summary
The central biological observation is that T. nylanderi queens and workers have the same genome but radically different physiology, behaviour and lifespan. Queens are long-lived and reproductive, while workers are shorter-lived and behaviourally active. Infected workers acquire several queen-like traits, including greatly extended lifespan and reduced activity. The authors therefore propose that the cestode may exploit regulatory systems that already generate normal caste plasticity.
They generated tissue-specific RNA-seq data from brain and fat body, comparing three groups: healthy queens, infected workers and uninfected workers. There were seven biological replicates per category/tissue, giving 42 ant RNA-seq libraries. They also analysed existing A. brevis transcriptomic and proteomic data to investigate parasite neuropeptides and possible molecular mimicry.
1. Infected workers become partially queen-like, especially in the fat body
The clearest result is in the fat body, the insect tissue analogous in some respects to vertebrate liver/adipose tissue and central to metabolism, immunity and endocrine regulation.
In the PCA shown on page 5, infected workers fall between normal workers and queens along the major expression axis. This is much less obvious in the brain. The fat-body data therefore support a partial shift from a worker transcriptional programme toward a queen-like programme.
The differential-expression analysis strengthens this interpretation. In the fat body:
- queens and infected workers shared 111 upregulated genes;
- queens and uninfected workers shared only 37;
- the queen–infected-worker overlap was nevertheless statistically depleted relative to random expectation, but substantially less depleted than the queen–uninfected-worker overlap.
The authors therefore appropriately describe the state as partially queen-like, rather than claiming infected workers actually become molecular queens.
The shared genes include candidates associated with metabolism, cellular maintenance, immunity and ageing, including:
- SAMTOR, an upstream sensor in mTORC1 signalling;
- EGFR, connected with insulin/IGF signalling;
- transferrin, associated with oxidative protection;
- trehalose transport and α-glucosidase genes;
- chaperonins and stress-response genes.
These results point toward modulation of the conserved IIS–TOR–juvenile hormone network involved in insect reproduction and longevity.
2. Brain responses are quite different
The brain does not simply become queen-like.
Only 59 genes were jointly upregulated in queens and infected workers, compared with 709 shared between infected and uninfected workers. PCA separation was also weaker and heterogeneous.
Shared queen/infected-worker brain genes included stress-maintenance candidates such as:
- Apolipoprotein D
- takeout-like
- TERT, telomerase reverse transcriptase
as well as ribosomal, translational and oxidative-stress-associated genes.
The paper therefore suggests that parasite-associated lifespan extension may involve relatively queen-like peripheral/metabolic physiology, whereas changes responsible for altered behaviour may follow a different neural programme.
That distinction is one of the most interesting conclusions of the study.
3. Infection strongly suppresses brain neuropeptide signalling
The authors identified 34 host neuropeptides and 23 candidate receptors, corresponding to 18 receptor loci after grouping paralogues.
They find surprisingly broad suppression of neuropeptide transcription in infected worker brains. Important examples include:
- tachykinin (TK)
- short neuropeptide F (sNPF)
- allatostatin A
- orcokinin
- CAPA
- DH1 and DH2 diuretic hormones
and several receptors were also reduced.
This provides plausible connections to the parasite-induced behaviour.
For example, tachykinin has roles in locomotion, aggression and social behaviour, while sNPF is involved in feeding, locomotion and foraging. Suppression of these pathways therefore fits reasonably well with the characteristic inactivity of infected workers.
The effect is not simply “make the brain queen-like”: the infected-worker brain frequently forms its own distinct transcriptional state.
4. The fat-body neuropeptide system is more queen-like
Neuropeptide and receptor expression in the fat body behaves differently. Major clusters of genes show infected workers moving toward queen-like expression patterns.
An especially interesting example is neuroparsin (NP), which changes in opposite directions between tissues: infection lowers it in the brain but increases it in the fat body.
The authors argue that this could connect caste plasticity, juvenile-hormone regulation and parasite manipulation.
5. No evidence for straightforward parasite neuropeptide mimicry
A particularly clear negative result concerns the mechanism.
The authors searched the cestode genome/transcriptome for candidate neuropeptides and examined previously identified proteins released into ant haemolymph.
They identified six candidate cestode neuropeptide families, but:
- none was predicted from the proteomic evidence to be secreted into host haemolymph;
- parasite peptides did not show convincing sequence similarity to host neuropeptides;
- previously detected secreted parasite proteins likewise did not resemble host neuropeptides.
They therefore reject the simplest hypothesis that A. brevis manipulates the ant by secreting molecular copies of ant neuropeptides.
The parasite instead appears more likely to alter the host’s own regulatory pathways.
What is novel about the paper?
I would separate the novelty into four levels.
1. Tissue-specific demonstration of the “pseudo-queen” state
Earlier work had reported some transcriptional resemblance between infected workers and queens using pooled abdomen/gaster samples.
The important advance here is showing that this resemblance is strongly tissue dependent.
The queen-like state is pronounced in the fat body, but much weaker in the brain.
That substantially changes the biological interpretation: infection is not inducing one global queen programme. Instead it seems to produce a mosaic phenotype in which different organs are reprogrammed in different directions.
2. Linking parasite-induced longevity to normal caste ageing mechanisms
The work specifically connects parasite infection to components of the normal ant longevity network:
IIS ↔ TOR ↔ juvenile hormone
The implication is conceptually interesting: rather than inventing a completely new longevity programme, the parasite may exploit an existing developmental/plasticity axis that ants already use to generate extraordinarily long-lived queens.
That is likely the most significant evolutionary message of the paper.
3. Identification of a broad neuropeptide suppression phenotype
The coordinated reduction of multiple behaviourally relevant neuropeptides in infected worker brains is a relatively specific mechanistic finding.
Rather than implicating a single “manipulation peptide,” the data suggest that infection produces a system-level reduction in worker-associated neuroendocrine signalling.
The repeated involvement of tachykinin is particularly notable because earlier work had also implicated it.
4. Evidence against simple molecular mimicry
Negative results can be important here.
A plausible model was:
cestode peptide → resembles ant neuropeptide → binds host receptor → alters behaviour.
The authors specifically test this idea and find little support.
That redirects attention toward mechanisms such as parasite-secreted proteins altering:
- host neuropeptide production,
- receptor expression,
- endocrine signalling,
- metabolism,
- transcription or chromatin regulation.
Critique
The paper is interesting and the basic result is convincing, but I think its mechanistic claims should be interpreted substantially more cautiously than the title and parts of the discussion might suggest.
1. The study establishes association, not manipulation mechanism
This is by far the largest limitation.
Everything central to the paper is based on gene-expression correlations.
The study shows:
infection → altered transcription
and it shows that some of those alterations resemble queen transcription.
It does not establish:
parasite factor → pathway X → neuropeptide Y → behavioural change/longevity
For example, reduced tachykinin expression could be:
- directly induced by the parasite;
- secondary to reduced activity;
- secondary to altered feeding;
- a consequence of infection-associated metabolic changes;
- a compensatory host response;
- or part of the mechanism producing behavioural manipulation.
RNA-seq alone cannot discriminate among these possibilities.
The authors acknowledge this limitation, but some phrases such as parasites “exploit” or “recruit” pathways are stronger than the experimental evidence strictly warrants.
2. Age is an important confounder
This is potentially substantial.
Infected workers live much longer than uninfected workers and therefore the infected animals sampled may on average have been older.
Age itself changes:
- TOR/IIS signalling,
- stress pathways,
- detoxification,
- protein homeostasis,
- immune expression,
- mitochondrial/metabolic gene expression.
The authors explicitly acknowledge this and refer to unpublished age-controlled experiments.
But unpublished results cannot really resolve the confounding in the experiment reported here.
A stronger design would have compared:
young infected vs young uninfected
and
old infected vs old uninfected
while separately analysing queens.
That would allow infection × age interactions to be estimated directly.
3. Queens are not perfectly matched controls
The queens came from seven additional colonies, whereas infected and uninfected workers came from matched colonies.
That introduces potential:
- colony-genotype effects,
- environmental effects,
- colony-state effects.
Because social insect gene expression can be colony dependent, the queen-versus-worker comparison is consequently not as clean as the within-colony infected-versus-uninfected worker comparison.
This does not invalidate the results, but it weakens the interpretation of modest queen/infected-worker overlaps.
4. “Queen-like” may be slightly overstated
The numbers themselves deserve emphasis.
In the fat body, queens and infected workers shared 111 upregulated genes, but queens possessed 922 uniquely overexpressed genes.
Furthermore, the statistical overlap between queens and infected workers had an odds ratio of 0.16, i.e. it remained depleted relative to the chosen background.
So infected workers are clearly not close replicas of queens.
A more precise biological description might therefore be:
infection moves a subset of worker metabolic and longevity pathways in a queen-associated direction.
That is somewhat different from saying infection activates a queen programme.
The paper generally recognises this distinction, although the broader narrative occasionally pushes the pseudo-queen interpretation rather hard.
5. Transcript abundance is particularly weak evidence for neuropeptide activity
This is an important technical point.
For neuropeptides:
mRNA abundance ≠ mature active peptide concentration.
Neuropeptide activity additionally depends on:
- translation;
- precursor processing;
- proteolytic cleavage;
- amidation and other post-translational modifications;
- vesicular storage;
- neuronal release;
- receptor localisation;
- receptor activation/desensitisation.
Therefore, the conclusion that “neuropeptide signalling is suppressed” is plausible but stronger than what RNA-seq alone establishes.
Ideally the next step would involve targeted peptidomics or mass spectrometry, receptor assays and perhaps neural localisation.
6. Bulk brain RNA-seq probably hides the most interesting biology
The authors themselves suggest this.
An insect brain contains many neuronal populations, and neuropeptide-producing cells may comprise a tiny fraction of the tissue.
A large expression change in ten neurons could therefore disappear in bulk RNA-seq, while changes in common glial or metabolic genes dominate the PCA.
The apparently weak queen-like brain signal could consequently reflect averaging rather than genuine absence of a specific neural mechanism.
Single-nucleus RNA-seq, spatial transcriptomics or in situ hybridisation would be much more informative.
7. Failure to detect sequence mimicry does not rule out receptor mimicry
The peptide-mimicry test is useful but relatively narrow.
A parasite peptide need not have >30% overall sequence identity with an ant neuropeptide to activate the same receptor.
For short peptides in particular:
- a small receptor-binding motif may be sufficient;
- structural similarity can occur without strong global sequence similarity;
- parasite molecules could act allosterically;
- non-peptide parasite proteins could alter receptor or peptide processing.
So the study rules against obvious sequence mimicry, rather than neuroendocrine mimicry in the broader functional sense.
This distinction is worth making more strongly.
Overall assessment
I think this is a good descriptive/mechanistic-hypothesis paper rather than a definitive mechanistic paper.
Its strongest contribution is not the identification of any single longevity gene or neuropeptide. It is the discovery that parasite-induced phenotypic transformation appears to be modular:
Fat body:
worker → partially queen-like metabolic/longevity state
Brain:
worker → a distinct infection-associated state with widespread suppression of behavioural neuropeptide signalling
That leads to an intriguing interpretation: the parasite may extend lifespan and suppress worker behaviour through partly separate mechanisms, rather than simply transforming an infected worker wholesale into a queen.
The next decisive experiment would be to manipulate candidate pathways—particularly tachykinin/sNPF signalling and IIS–TOR–JH components—in infected and uninfected workers and determine whether lifespan and behaviour can be independently reversed or reproduced. If that worked, it would turn the present transcriptional associations into a much stronger causal model.
Q2: does the infestation mean inhibition of TOR
No. In this paper, infestation does not look like simple inhibition of TOR/mTOR signalling.
In fact, the authors point in almost the opposite direction. In the fat body of infected workers they find increased expression of SAMTOR—the S-adenosylmethionine sensor upstream of mTORC1—and describe the shared queen/infected-worker transcriptional pattern as involving modulation of IIS, TOR and juvenile-hormone pathways. They also note that upregulation of IIS/TOR/JH components could be associated with the long-lived state.
There is an important distinction, though: SAMTOR upregulation does not mean TOR is activated. SAMTOR is itself an inhibitory regulator of mTORC1 under low-SAM/methionine conditions. Higher SAMTOR mRNA could therefore potentially favour reduced mTORC1 activity, but its functional effect depends on SAM availability and protein activity. The RNA-seq data do not measure TOR phosphorylation or mTORC1 kinase activity.
So the paper supports:
infection → alteration/remodelling of TOR-related signalling
rather than:
infection → demonstrated TOR inhibition.
To establish inhibition, they would ideally need measurements such as phospho-S6K, phospho-S6, phospho-4E-BP or equivalent insect TOR activity markers. Interestingly, this leaves open the possibility that the long lifespan of infected workers involves reduced effective TORC1 activity despite altered/upregulated transcription of some pathway components—but this paper does not establish that.
Well it looks like inhibition to me.