AKAP9 regulates activation-induced retention of T lymphocytes at sites of inflammation

The mechanisms driving T cell homing to lymph nodes and migration to tissue are well described but little is known about factors that affect T cell egress from tissues. Here, we generate mice with a T cell-specific deletion of the scaffold protein A kinase anchoring protein 9 (AKAP9) and use models of inflammatory disease to demonstrate that AKAP9 is dispensable for T cell priming and migration into tissues and lymph nodes, but is required for T cell retention in tissues. AKAP9 deficiency results in increased T cell egress to draining lymph nodes, which is associated with impaired T cell re-activation in tissues and protection from organ damage. AKAP9-deficient T cells exhibit reduced microtubule-dependent recycling of TCRs back to the cell surface and this affects antigen-dependent activation, primarily by non-classical antigen-presenting cells. Thus, AKAP9-dependent TCR trafficking drives efficient T cell re-activation and extends their retention at sites of inflammation with implications for disease pathogenesis.

WT refers to the wild-type allele. In mice expressing the targeted allele, AKAP9 lacZ/Neo the lacZ/Neo cassette was excised using mice expressing flp recombinase to generate AKAP9 cko mice. AKAP9 cko mice were then bred with CD4-Cre mice to obtain AKAP9 cko/CD4 mice by excision of Exon 8. (b) Genotyping of DNA from tail biopsies of AKAP9 cko/CD4 mice and control mice. Primer pair 1 (PR1) binds after Exon 7 and before Exon 8 thereby spanning the loxP sites, yielding a 759bp fragment in mice with the mutant allele and a 579bp band for the WT allele. Primer pair 2 (PR2) binds after Exon 7 and within the loxP site yielding a PCR product of 286bp in mice with a floxed allele. (c-d) Verification of protein knock-down in purified T H 1 cells using an AKAP9 antibody that recognizes the N-terminus of AKAP9. Western blot analysis revealed that the expected 450kDa AKAP9 protein observed in wild-type samples was absent in AKAP9 cko/CD4 samples (c). Immunofluorescence of T H 1 cells stained with AKAP9 antibody and counterstained with DAPI. Staining at the centrosome, observed in AKAP9 wt cells is largely absent in AKAP9 cko/CD4 cells (d). # p < 0.05.   Top blot is CD3 loading control and IP, developed from a membrane strip cut below the 25kDa band as used to control for successful IP before blotting for ZAP70 and EB1. Lower blots left: exposures used for ZAP70 (upper) and EB1 (lower). Lanes 1 and 2 are loading controls, 2 and 3 are IPs, 4 and 5 are IgG control and 6 and 7 are IP but half of what was loaded in lane 4 and 5. Lower blots rights: loading control blots for ZAP70 (upper) and EB1 (lower). ZAP70 was developed from a membrane strip cut at the 55kDa and at the 100kDa, EB1 was developed from a membrane strip cut at the 55kDa band and at the 25kDa band.

Generation of T cell specific AKAP9 knock-out mouse
Mice with conditional deletion of AKAP9 in CD4+ and CD8+ T cells (AKAP9 cko/CD4 ) were generated on a C57Bl/6 strain (Supplementary Figure 1a).

CD4 counts and surface molecule expression.
We found no differences in spleen size (Supplementary Figure 2a), total CD4+ T cell counts in the blood, inguinal lymph node and spleen (Supplementary Figure 2b), and distribution of L-selectin and CD44 on splenic CD4+ T cells (Supplementary Figure 2c). These data suggest that maturation and expansion of CD4+ T cells is not dependent on AKAP9. Furthermore, surface levels of T cell receptor molecules, costimulatory molecules and CCR7 were similar in AKAP9 cko/CD4 and AKAP9 wt mice (Supplementary Figure   2d).

AKAP9 deficient CD4+ T cells exhibit normal recruitment, homing, adhesion and migration.
To investigate T-cell accumulation at the site of inflammation, we examined recruitment to the air pouch 24 hours after installation of PBS or TNF in AKAP9 cko/CD4 and AKAP9 wt mice. Both groups showed similar recruitment to their pouch (Supplementary Figure 3a). To examine homing to lymph nodes, naïve T-cells of AKAP9 cko/CD4 and AKAP9 wt mice were differentially labeled ex vivo and co-transferred into wild-type recipient animals. Lymph nodes and spleens were analyzed for transferred cells 3 Figure 3f). These data indicate that adhesion and migration both on the isolated but also on stimulated endothelial cells is not affected by AKAP9.

Co-localization of TCR with TfR and LAMP-1
To assess co-localization of TfR and LAMP-1 with the T cell receptor in a standardized manner, we used pixel-by-pixel comparison and determined the pearson's R in a computerized manner of cells in Figure   5d using the Fiji Coloc2 plugin. We observed increased co-localization of the TCR with LAMP-1 in AKAP9 wt , and increased co-localization with TfR in AKAP9 CD4/ko cells (Supplementary Figure 5). It is noteworthy that biological meaningful co-localization is indicated by values >0.5. As both endosomal compartments are expected co-localize with the TCR, we expected values >0.5 in all groups.