After birth, the gut becomes colonized by billions of commensal bacteria in parallel with the appearance of numerous isolated lymphoid follicles (ILFs) along the length of the intestine. Now, a recent study published in Nature has pinned down the molecular mechanism for the genesis of ILFs in mice, showing that this process requires the detection of peptidoglycan from Gram-negative bacteria by the innate immune sensor NOD1 (nucleotide-binding oligomerization domain protein 1).

In the fetus, secondary lymphoid tissues develop in the sterile environment in a process involving lymphoid-tissue inducer (LTi) cells. In the first few weeks of life, hundreds of small clusters of LTi-like cells, known as cryptopatches, form between the crypts of the intestinal lamina propria. These cryptopatches are then thought to develop into ILFs as they become populated with B cells. To study what triggers the induction and maturation of ILFs, Eberl and colleagues generated transgenic mice that express enhanced green fluorescent protein under the control of the nuclear hormone receptor retinoic-acid-receptor-related orphan receptor-γt (RORγt), which is known to be expressed by LTi cells and cryptopatch cells. In these mice, ILFs could be detected throughout the intestine; immature ILFs, which contain few B cells, were more common in the proximal ileum whereas mature ILFs, which contain an organized B-cell follicle, were most prominent in the distal ileum and colon. This distribution correlates with the increasing bacterial density of the distal parts of the intestine. Consistent with a requirement for commensal bacteria in ILF induction and maturation, germ-free mice contained only cryptopatches and a few immature ILFs. The number of ILFs could be increased in the germ-free mice following reconstitution with various strains of commensal bacteria. In particular, Gram-negative bacteria (such as Bacteroides distasonis or Escherichia coli) but not Gram-positive bacteria (such as Lactobacillus acidophilus) could induce the formation of mature ILFs.

A role for bacteria in ILF formation suggests the involvement of pattern-recognition receptors in this process. Analysis of mouse strains that lack various Toll-like receptors, NOD receptors and signalling adaptors revealed that only NOD1-deficient animals showed a marked impairment in the formation of ILFs. This suggested that NOD1 detects peptidoglycan from Gram-negative bacteria and triggers ILF formation. Indeed, further experiments confirmed that NOD1 expression by stromal cells, such as epithelial cells, but not haematopoietic cells was necessary for the generation of ILFs. Moreover, ILF formation could be induced in germ-free mice that were fed a NOD1 ligand, but not a NOD2 ligand.

Given that ILFs fail to develop in mice that lack CC-chemokine receptor 6 (CCR6), the authors next tested for the involvement of its two ligands, CC-chemokine ligand 20 (CCL20) and β-defensin 3. Similarly to CCR6-deficient mice, mice that lacked β-defensin 3 or mice that were treated with a CCL20-specific blocking antibody had few ILFs. In addition, the finding that transcripts encoding β-defensin 3 and CCL20 were decreased in germ-free or NOD1-deficient mice suggests that NOD1 triggering induces β-defensin 3 and CCL20 expression, which facilitates the recruitment of CCR6-expressing B cells into cryptopatches and immature ILFs.

Finally, comparative analysis revealed profound alterations in the composition of the bacterial gut flora in mice that lacked mature ILFs (NOD1-deficient and β-defensin-3-deficient mice) compared with wild-type mice. This suggests a role for the cryptopatch–ILF system in intestinal homeostasis and the existence of reciprocal regulation between the intestinal bacterial flora and the immune system.