Recently a method for in vivo
Recently, a method for in vivo photoactivation of Fmoc-Gln(Trt)-OH receptor expressing PA-GFP in precise microanatomical compartments was described (Victora et al., 2010), which makes it possible to optically mark Tfh cells and track them 20 hr later (Shulman et al., 2013). Unexpectedly, it was reported that Tfh cells frequently migrated out of the follicle to invade neighboring GCs and proposed that this promoted affinity maturation by providing a diverse polyclonal source of CD4+ T cell help (Shulman et al., 2013). However, the temporospatial context of such promiscuous behavior was not defined. We have developed an alternative method for optical marking by two-photon photoconversion (TPP) of cells expressing the photoconvertible fluorescent protein Kaede (KD) (Chtanova et al., 2014). Our studies using TPP show striking differences in the migration and behavior of Tfh cells during three distinct phases: the primary response by naive CD4+ T cells; the memory phase following resolution of the GC response; and the secondary response by antigen-experienced cells. We demonstrate the migration of GC Tfh cells in the primary response was confined to the GC of origin and infrequently observed to cross into the follicular mantle (FM), a distinct region in the follicle surrounding the GC (Hardie et al., 1993). Follicular memory T cells were tracked to the outer follicle where they scanned CD169+ macrophages lining the subcapsular sinus (SCS) and became activated to divide upon antigen rechallenge. There was unrestricted movement of GC Tfh cells in the secondary response, and we show that they also enter and leave the follicle via the lymphatic flow in the SCS. Finally, we use TPP and single cell gene expression and functional analyses to show that the temporospatial cues guiding the positioning of Tfh cells during these phases of the immune response were provided in part by Epstein-Barr virus-induced G protein coupled receptor 2 (EBI2).
Discussion The origin and fate of Tfh cells has been intensely studied since their first description 14 years ago (Breitfeld et al., 2000, Schaerli et al., 2000). Although mice engineered to report BCL6 (Kitano et al., 2011, Liu et al., 2012) and interleukin-21 (IL-21) (Lüthje et al., 2012) expression have provided powerful tools to analyze Tfh cells, their usefulness has been limited by Tfh cell heterogeneity and plasticity (Cannons et al., 2013). In this regard, the development of methods for the optical marking and tracking of cells based on their microanatomical location have created further opportunities for more precise delineation of Tfh cell dynamics and the molecular cues that underpin their behavior. Here we have used optical marking by TPP to link Tfh cell location to their behavior, phenotype, and gene expression. Our studies show remarkable differences in the migration pattern and single cell gene-expression signatures between primary and secondary Tfh cells. In addition, we report a subpopulation of “follicular memory T cells” that reside in the follicle where they scan SCS macrophages to initiate the secondary immune response upon antigen re-exposure. This temporospatial dissection of Tfh cell dynamics offers multiple new insights into regulation of GC responses in naive and antigen-experienced animals. Imaging of primary Tfh cells at the peak of the GC response revealed clear spatial segregation in the FM and GC compartments. This confinement was confirmed by TPP and discontinuous cell tracking 24 hr later, which showed retention of the majority of photoconverted GC Tfh cells in the original GC and follicle. Furthermore, NMF analysis of single cell gene expression signatures of FM and GC Tfh cells support the notion that they represent molecularly distinct cell populations. Thus, we conclude that the primary GC is a closed structure designed to partition responding GC B cells and restrict their access to CD4+ T cell help. At face value, these data contrasts with the findings of Shulman et al. who concluded that the GC is an open structure designed to broaden the diversity of the available CD4+ T cell help (Shulman et al., 2013). However, the preliminary experiments in their paper only examined polychromatic responses in naive animals that demonstrated initial colonization by multiple clones of red, green, or cyan T cells with the same TCR specificity and not interfollicular exchange as claimed. Furthermore, their subsequent experiments involved prime-boost immunization protocols that involved repeated exposure to antigen. This is a critical point of difference as they do not show any equivalent photoactivation data from naive responses (Shulman et al., 2013). Hence, their data is more consistent with our memory responses. In fact, we not only observed the migration of Tfh cells out of the GC in the secondary response but also their transport in the lymphatic flow of the SCS. This passive transport mechanism whereby cells “surf” the lymph appears to be an efficient and rapid mechanism for dissemination of cells that bypasses the need to traverse multiple anatomical compartments across disparate chemokine gradients. Regardless, it will be interesting to determine what role factors enriched in lymph, such as S1P, play in driving secondary Tfh cells to enter and leave the follicle.