Preprint Club
A cross-institutional Journal Club Initiative

Keywords

• Viral infections

• CD8+ T cell immunity

• Microbiota

• Infections of newborns

Main Findings

In this preprint, the authors sought to understand why certain infants are predisposed to severe manifestations of respiratory viral infections such as influenza and respiratory syncytial virus (RSV) where other new-borns are not. Given that perturbations to the gut microbiota during infancy are associated with an increased risk of such severe infections, the authors hypothesized that the normal development of the gut microbiota during infancy is critical to fight respiratory viral infections during the infancy period.

To examine the role of the infant gut microbiota in generating immune responses against influenza, the authors first devised a clever model of dysbiosis in infant mice. Pregnant wildtype mouse dams were given an antibiotic (ABX) cocktail at embryonic day 15 (E15), and the new-born pups were subsequently cross-fostered to a non-ABX-treated dam (termed “dysbiotic” pups). Control pups were those who had not received ABX during the embryonic stage. The dysbiotic and control pups were subsequently infected with influenza-PR8 carrying the model OVA antigen at post-natal day 14 (PN14). The subsequent immune and physiological responses to infection were monitored. Interestingly, the dysbiotic pups lost significantly more weight during influenza challenge than control infant mice, highlighting the role of the microbiota in mediating responses against respiratory viral infection. This effect coincided with lower proportions of OVA-specific CD8+ T cells in the lungs of dysbiotic pups at days 7 and 14 post infection. Secondary rechallenge experiments with PR8-OVA or heterosubtypic influenza X31 confirmed an aberrant generation of T cell memory in dysbiotic pups; control pups had generated CD8+resident-memory (TRM) T cells 6 weeks post primary infection and were protected from weight-loss associated with secondary infection. These effects were abrogated in dysbiotic pups.

Next, adoptive transfer experiments of OT-I CD8+ T cells from dysbiotic or control pups revealed that dysbiosis during infancy leads to T cell-intrinsic defects in PR8-specific T cells. To examine the T cell-intrinsic defects that occur in influenza-specific T cells during infection, the authors performed multiomic analyses including scRNA-seq of influenza-specific CD8+ T cells from control or dysbiotic pups 7 days post infection. The authors’ findings confirmed that CD8+ T cells from dysbiotic pups retain a naive T cell phenotype and do not undergo differentiation into memory in contrast to CD8+ T cells from control new-borns. Gene-regulatory network analysis inferred NFIL3 as being a novel transcription factor important for memory T cell specification during infancy. Importantly, levels of Nfil3 mRNA and protein were significantly reduced in CD8+ T cells of dysbiotic pups relative to control animals. To confirm the role of NFIL3 in memory CD8+ T cell specification during infancy, the authors made use of an Nfil3 floxed mouse crossed to Cre recombinase under the control of the T cell specific Lck allele (Lck x Nfil3fl/fl). Nfil3fl/fl or control infant animals were then infected with influenza-PR8-OVA at PN14, and the lung antigen-specific CD8+ T cell compartment was analysed. Strikingly, Nfil3fl/fl infant CD8+T cells expressed a phenotype like CD8+ T cells from dysbiotic pups – a propensity to retain a naïve T cell phenotype. Chromatin immunoprecipitation (ChIP) experiments revealed that NFIL3 likely leads to epigenetic repression of the Tcf7and Lef1 loci. Together, these results implicate T cell-intrinsic NFIL3 as being regulated by elements of the infant microbiota, and as being an important regulator of T cell memory fate licensing.

The authors then wanted to confirm whether a similar mechanism of microbiota-mediated CD8+T cell NFIL3-expression occurs in human infants. The authors performed multiomic profiling of CD8+ T cells from lungs of human infants that have succumbed to upper respiratory viral infection. The authors were able to further stratify these infants into control or dysbiotic based on their faecal metagenomic profile. Strikingly, dysbiotic human infant lung CD8+ T cells expressed a similar reduction of Nfil3 as that observed in infant mice. Furthermore, the human dysbiotic infant lung CD8+ T cell profile similarly polarized away from a memory state and retained a naïve-like phenotype.

Lastly, in both human and mouse experiments, the authors were able to show that microbial dysbiosis is associated with a decrease in circulating levels of inosine, which the authors hypothesized may be important for inducing NFIL3-mediated CD8+ effector and memory T cell programs in the lung during neonatal influenza infection. Indeed, ex vivo inosine supplementation experiments confirmed that inosine induces increases in CD8+ T cell NFIL3 expression in both control and dysbiosis conditions. Furthermore, in vivo inosine supplementation of control and dysbiotic neonatal mice during influenza infection rescued the proportion of antigen-specific lung CD8+ T cells, and protected mice from the severe weight loss associated with infection.

Limitations

• The study has done a very thorough characterization of the effects of microbial dysbiosis on lung effector and memory T cell responses during influenza infection in both mice and human infants. However, it would be helpful to perform the in vivo inosine supplementation experiments in the context of a secondary viral rechallenge as the authors have done in Figures 1G-K.

• The study would be bolstered by including a germ-free control group supplemented with or without inosine for their in vivo inosine supplementation experiments. This control would be a bona fide proof that it is the microbiota, rather than host-derived inosine that is mediating the downstream protective effects against infection.

• The work may benefit from inclusion of cell numbers alongside the depicted proportions for their flow cytometry experiments if possible. This will give a more complete picture of the effects of T cell recruitment and cell death during lung infection.

Significance/Novelty

What is the novelty of the preprint for the specific field?

The preprint is novel in several regards. First, the work identifies a novel regulator of CD8+ T cell memory – NFIL3. Furthermore, the authors mechanistically show that NFIL3 acts to epigenetically repress the Tcf7 and Lef1 loci which have been well established to play a role during the T cell memory fate specification process. The authors also establish a novel mechanistic link between the infant microbiota and CD8+ T cell memory during neonatal viral infection. Namely, the authors find that microbiota-derived inosine is decreased in dysbiotic infants, and that inosine is important for lung CD8+ memory T cell licensing via NFAT3 during infection. Most importantly, the authors show these novel mechanistic links in both mouse and human neonates. In particular, the authors’ studies in human samples is quite remarkable.

How does the result of the preprint matter for general immunologists and/or patients?

The work is likewise quite impactful for immunologists and patients alike. Pregnancy is a critical period for foetal immune development in multiple regards. The authors’ study suggests that bacterial infections during pregnancy (e.g L. monocytogenes, S. typhimurium) requiring antibiotics likely negatively affect a new-born’s ability to fight common respiratory viral infections during the infancy period. The authors’ findings offer an intriguing strategy (in the form of inosine supplementation) that may potentially be used in pregnant women taking antibiotics or in new-borns born under such “dysbiosis”-predisposing conditions. It is possible that inosine supplementation during the embryonic or post-natal stages can help at-risk infants fight common respiratory viral infections, or, at least mitigate their deleterious effects.

Credit

Reviewed by Boyan Tsankov as part of a cross-institutional journal club between the Icahn School of Medicine at Mount Sinai, the University of Oxford, the Karolinska Institute and the University of Toronto.

The author declares no conflict of interests in relation to their involvement in the review.