Finding. In three children with juvenile systemic sclerosis, blood monocytes start with high NF-κB inflammatory signaling and, over two years after autologous stem cell transplantation, lose systemic-sclerosis-linked genes such as SERPINE1 while mitochondrial-function and oxidative phosphorylation genes rise. The transplant is billed as an immune reset. The single-cell time series says it is also a metabolic one.
Why this paper matters
Juvenile systemic sclerosis is rare, fibrotic, multi-organ, and brutal when it ignores standard immunosuppression. Autologous stem cell transplantation is already an emerging rescue for the worst adult and pediatric cases. What it does to the myeloid cells that feed fibrosis has been a black box.
Elrod, Sanyal, Torok and colleagues open that box with CITE-seq on PBMCs from three patients, sampled before transplant and at 6, 12, and 24 months, against healthy controls. They focus on monocytes because those cells traffic inflammation into skin and viscera. The mito-relevant sentence is not a side hit. Mitochondrial and OXPHOS genes increase with time after ASCT as NF-κB recedes from a disease-high baseline. That pairing is the reason this paper belongs on a mitochondria desk: inflamed myeloid cells are often OXPHOS-poor, and a therapy that works clinically may be restoring that axis rather than only deleting autoreactive lymphocytes.
What they actually measured
Cellular indexing of transcriptomes and epitopes (CITE-seq) on PBMCs, then a deliberately simple longitudinal model: pseudobulk monocyte expression on a log scale, regressed on months since transplant. That model is how they claim "widespread changes" rather than a volcano-plot fishing trip.
The named gene-level example is SERPINE1, a systemic-sclerosis-linked transcript that falls. Pathway-level, NF-κB-associated inflammatory signaling is high at baseline versus healthy monocytes and declines after ASCT. Genes annotated to mitochondrial function and oxidative phosphorylation move the other way: up, progressively. Monocyte subset composition also changes, and the authors say those shifts may contribute to the expression trends. That is an honest caveat inside the result, not a reviewer afterthought.
What they did not measure, at least in the abstract: oxygen consumption, membrane potential, mitochondrial mass, or any causal link from the OXPHOS rise to skin or lung scores. n is three. Conditioning, infection, and drugs ride along with time.
How to read the score
Mid-high for a pediatric immunometabolism time series; not a mechanism paper on the organelle. The mitochondrial claim is a coordinated gene-set trajectory, not a respiratory phenotype. Confidence is medium: the design (paired longitudinal CITE-seq, healthy controls, time regression) is clean, the n is tiny, and composition versus cell-intrinsic metabolism is unresolved. Score it as a signal to follow in larger ASCT series, not as proof that transplant fixes monocyte mitochondria.
What to do with it
If you study myeloid immunometabolism or fibrotic disease, take the NF-κB-down / OXPHOS-up pairing as a hypothesized companion biomarker of ASCT response. When the objects post, extract the time coefficients for OXPHOS genes and SERPINE1 and test whether subset composition explains them. Do not write that stem-cell transplant is a mitochondrial therapy. Do write that in this disease the monocyte that stops shouting NF-κB starts reading out a more oxidative program, and that those two arrows should be plotted together.