Zhang et al. · Cellular and Molecular Immunology · 2025

Antisense to human CD39 dysregulates immune metabolism in inflammatory bowel disease

Zhang L, Cagle C, Nguyen DH, Gomes GS, Gromova B, Csizmadia E, Karimitar A, Lee GR, Chen G, Kokkotou E, Grossberg L, Jiang S, Cheifetz AS, Kota SK, Longhi MS

Read the paper The product used

The study

What was asked, and what was found

Zhang et al., Cellular and Molecular Immunology, 2025 asked what the natural antisense transcript of CD39 does to the cells that carry it. CD39 sits at the head of the pathway that turns extracellular ATP into adenosine, and it is low in inflammatory bowel disease, where regulatory T cells and Th17 effectors fall out of balance. The same laboratory had already shown that an endogenous antisense RNA holds CD39 down and that it is raised in patient cells. This study asked what silencing that antisense does to metabolism, and it found that the answer differs by cell type because the antisense is expressed as two different splice variants in the two cell types.

The work used self-delivering AUMlnc sdASOs at 10 μM for 72 hours, with no transfection reagent, against each variant separately and always beside a scrambled control. In the Jurkat line the block moved 4129 genes and lit up metabolic gene sets. In regulatory T cells from healthy donors and from patients, blocking the variant those cells express cut glucose uptake, extracellular acidification, glycolytic capacity and proliferation, raised free phosphate release, and improved suppression of effector cell proliferation in patient cells. Under an inflammatory challenge of IL-6 and IL-1 beta, which normally drives patient regulatory T cells towards an effector state, the treated cells held their FOXP3 and their RORC and IL-17 fractions returned to baseline. In Th17 cells, blocking the other variant lowered oxygen consumption, maximum respiration and spare respiratory capacity and improved the mitochondrial membrane potential reading. A CD39 silencing arm run alongside showed these changes were not a downstream consequence of CD39 rising.

The same two oligonucleotides were then given to NOD/scid/gamma mice carrying human CD4 lymphocytes, as a single intraperitoneal injection at 5.4 mg/kg at the moment of an intrarectal challenge, with animals harvested 72 hours later. Both lowered the disease activity index, preserved colon length, lowered the histology score and reduced T cell infiltration into the colon, with five animals per arm. The metabolic split held in the animals too: the regulatory T cell oligonucleotide lowered GLUT1 on gut CD4 lymphocytes, and the Th17 oligonucleotide raised the mitochondrial ratio in the intraepithelial and lamina propria compartments.

Key findings

AUMlnc sdASOs against the regulatory T cell variant of the CD39 antisense transcript, at 10 μM with no transfection reagent, cut glucose uptake, extracellular acidification, glycolytic capacity and proliferation in regulatory T cells from both healthy donors and patients.(Results, Silencing of CD39-AS RNA controls Treg glucose metabolism, Fig. 2C)
The same silencing also limited extracellular acidification, glycolytic capacity and proliferation in regulatory T cells, in both healthy donors and patients.(Results, Silencing of CD39-AS RNA controls Treg glucose metabolism, Fig. 2D to 2G)
Free phosphate in the culture supernatant rose, which the authors read as higher ectoenzymatic activity, and suppression of effector cell proliferation by patient regulatory T cells improved.(Results, Silencing of CD39-AS RNA controls Treg glucose metabolism, Fig. 2H and 2I)
Under an inflammatory challenge that normally pushes patient regulatory T cells towards an effector phenotype, silencing the antisense held them: FOXP3 rose and the RORC and IL-17 positive fractions returned to baseline.(Results, Silencing of CD39-AS RNA controls Treg glucose metabolism, Fig. 3A to 3C)
A second oligonucleotide against the Th17 variant lowered oxygen consumption, maximum respiration and spare respiratory capacity, and raised the mitochondrial membrane potential reading, so the two variants gave two different metabolic effects in two different cell types.(Results, Blockade of CD39-AS modulates oxidative responses in Th17 cells, Fig. 4B to 4E)
The mitochondrial membrane potential reading rose in Th17 cells given the second oligonucleotide, which the authors read as less mitochondrial depolarisation.(Results, Blockade of CD39-AS modulates oxidative responses in Th17 cells, Fig. 4F to 4H)
One intraperitoneal injection at 5.4 mg/kg in mice carrying human CD4 cells lowered the disease activity index, preserved colon length, lowered the histology score and reduced T cell infiltration into the colon.(Results, Silencing of CD39-AS modulates metabolic function and clinical outcome in experimental colitis, Fig. 5B to 5D)
The metabolic readouts followed the same split in the animals as in the dish: the regulatory T cell oligonucleotide lowered GLUT1 on gut CD4 cells, and the Th17 oligonucleotide raised the mitochondrial ratio.(Results, Silencing of CD39-AS modulates metabolic function and clinical outcome in experimental colitis, Fig. 5E and 5F)
The gene changes were not a knock-on from CD39 rising, because they persisted when CD39 was silenced at the same time, which is the control that makes the antisense itself the actor.(Results, Silencing of CD39-AS RNA controls Treg glucose metabolism)
In the Jurkat line, blocking the antisense moved 4129 genes, with tryptophan metabolism, xenobiotic metabolism and TGF beta signalling among the enriched sets, which is what pointed the study at metabolism in the first place.(Results, Silencing of CD39-AS RNA controls Treg glucose metabolism, Fig. 1B and 1C)

Product: AUMlnc sdASO
Targets: ENTPD1-AS1, the CD39 natural antisense transcript, variant v1 (ENTPD1-AS1-209) · human · long non-coding RNA, antisense transcript, the variant expressed in regulatory T cellsENTPD1-AS1, the CD39 natural antisense transcript, variant v2 (ENTPD1-AS1-201) · human · long non-coding RNA, antisense transcript, the variant expressed in Th17 cellsCD39-AS is described throughout as a long noncoding RNA · human · long non-coding RNA, the paper's own classification
Models: Jurkat human T cell line, used for the transcriptome experiment · cell linehuman regulatory T cells and Th17 cells from the peripheral blood of 60 inflammatory bowel disease patients and 30 matched healthy donors, and from the lamina propria of colon biopsies and resections · primary cellsNOD/scid/gamma mice reconstituted with human CD4 lymphocytes, trinitrobenzene sulfonic acid colitis · in vivo
Delivery: self-delivering, added to cultured cells at 10 μM for 72 hours with no transfection reagent; in the colitis model a single intraperitoneal injection at 5.4 mg/kg given at the same time as the intrarectal challenge
Readouts: bulk RNA sequencing, NanoString nCounter Metabolic Pathways Panel, RT-qPCR, glucose uptake luminescence assay, Seahorse extracellular acidification rate and oxygen consumption rate, malachite green free phosphate assay, cell proliferation and Treg suppression assays, flow cytometry, JC-1 mitochondrial membrane potential staining, live cell imaging for reactive oxygen species, disease activity index, colon length and histology score, immunohistochemistry
Measured: concentration and exposure in culture: 10 μM for 72 h (Jurkat cells, regulatory T cells and Th17 cells)dose in the humanized colitis model: 5.4 mg/kg, intraperitoneal, single injection (given at the time of intrarectal trinitrobenzene sulfonic acid, harvest 72 h later)genes changed in Jurkat cells by blocking the v1 variant: 4129 differentially expressed, 1368 down, 2761 up (bulk RNA sequencing, three replicates, 72 h at 10 μM)donors behind the glucose uptake result: 9 healthy subjects, 7 inflammatory bowel disease patients from blood, 6 from lamina propria (regulatory T cells, paired comparison against the scrambled control)animals per arm in the colitis experiment: 5 per group, three groups (scrambled control, v1 oligonucleotide, v2 oligonucleotide)human cell dose and engraftment threshold in the humanized mice: 2-3 × 10⁶ CD4 T cells, more than 10% human chimerism in 75% of the mice transferred (NOD/scid/gamma recipients, 3 weeks before sensitisation)
Figures: Fig. 2 What silencing the regulatory T cell variant does to glucose handling: transporter gene set, glucose uptake, extracellular acidification, glycolytic capacity, proliferation, free phosphate and suppression, in healthy and patient cells.Fig. 3 Regulatory T cells held at their phenotype under an IL-6 and IL-1 beta challenge when the antisense is silenced.Fig. 5 The humanized colitis experiment: the protocol, disease activity index, colon length, histology, GLUT1 on gut CD4 cells and the mitochondrial ratio after a single injection.
Citation: Zhang et al. Antisense to human CD39 dysregulates immune metabolism in inflammatory bowel disease. Cellular and Molecular Immunology 2025;22:730-742. doi:10.1038/s41423-025-01295-6

For research use only. Not for use in diagnostic or therapeutic procedures.