Unlike developmental systems that display progressive tissue growth and maturation1, the homeostatic state of adult tissues is robust, maintaining form and function2. Secondary lymphoid organs, are uniquely able to dramatically change size in response to immune challenge, adapting to the increased space requirements of millions of infiltrating and proliferating lymphocytes, while remaining structurally and functionally intact throughout3–5. As a mechanical system, lymph nodes continually resist and buffer the forces exerted by trafficking of lymphocytes entering and leaving the tissue in steady state6, and managing diurnal fluctuations in cell trafficking7. Tissue size is determined by lymphocyte numbers, highlighted by the small organ size in genetic models blocking lymphocyte development (Rag1 KO)8, and the ability of the tissue to expand 2-10 fold in size to accommodate lymphocyte proliferation through adaptive immune responses3–5,9.
Lymph nodes function at the interface of immunity and fluid homeostasis, constructing a physical 3-dimensional cellular meshwork linking fluid flow, immune surveillance and adaptive immunity10. The most populous stromal cell component are fibroblastic reticular cells (FRCs) which span the whole tissue generating a interconnected cellular network with small world properties11, forming robust clustered nodes with short path lengths, and surrounding bundles of extracellular matrix fibres12. It is widely assumed that extracellular matrix scaffolds are the predominant force-bearing structures determining tissue mechanics1,2. However the relative contributions of the cellular structures versus the underlying extracellular network to tissue mechanics have not been addressed in a highly cellularised system undergoing such extensive expansion2. As tissue size and cellularity increase in response to immunogenic challenge, it is known that lymph nodes become more deformable when challenged with compressive force3. During the remodelling process the FRC network maintains connectivity through the elongation and increased spacing between FRCs, increasing mesh size of the network3. It is also known that CLEC-2+ dendritic cells are required to prime the stromal architecture for tissue expansion3,4 but the downstream impacts on the mechanical properties of the cellular network and extracellular matrix scaffolds driving the adaptation in tissue mechanics are unknown. During tissue expansion, FRCs reduce their adhesion to the underlying extracellular matrix bundles and these matrix scaffolds become fragmented9. This makes lymph nodes an ideal model system to address the relative mechanical contributions of cellular and material structures to emergent tissue mechanics.
As fibroblasts are contractile, force-generating cells13, we hypothesise that the interconnected FRC network determines tissue mechanics. The FRC network, identified by podoplanin expression, spans the whole lymph node tissue but specifically supports CD3+ T cell function in the paracortex, providing trafficking routes from high endothelial venules to B cell follicles14,15,16,17 (Fig. 1A-B). We asked how the FRC network and associated extracellular matrix scaffolds participate in tissue mechanics and architecture in the immunological steady state. Following laser ablation (Fig. 1C-D, and supplementary Movie. S1A), we measured a mean recoil of 0.42mm/s-1 in the fibroblastic reticular network of the paracortex, using a fibroblast-specific membrane-eGFP mouse model (PDGFRa-mGFP-CreERT2 (Fig. 1C, Fig. S1, Fig. S2A), formally demonstrating that the reticular network is under mechanical tension in the tissue (Fig. 1E).
To investigate how cellular-scale components of the stromal architecture contribute to tissue mechanics, we examined the cytoskeletal and extracellular matrix structures of the reticular network. In steady state, FRCs tightly adhere to and enwrap collagen bundles forming a continuous meshwork18. Therefore, recoil following laser ablation is a measure of the combined mechanical properties of both the cell and the extracellular matrix structures in steady state (Fig. 1C-F). Maximum z-projections and orthogonal views of FRCs show F-actin cables aligned proximal to the collagen bundles and spanning beneath the T-cell-facing FRC plasma membrane (Fig. 1F, and supplementary Movie. S1B). These F-actin structures co-localised with phosphorylated myosin regulatory light chain (pMLC2)19 (Fig. 1F), indicating that FRCs are contractile and resisting strain in steady state20. Since resting lymph node size is constant, forces are balanced in the steady state tissue.
We next asked how the reticular network reacts to the forces exerted by increasing numbers of lymphocytes following immune challenge. It has been observed that the network of bundled matrix fibres becomes fragmented9, while the cellular component of the meshwork remains intact and connected3,4,11. Therefore, we next tested if the intact cellular network could compensate for the loss of extracellular matrix integrity to balance tissue tension. In response to immunisation with incomplete Freund’s adjuvant and ovalbumin (IFA/OVA), the tissue expands 2-3 fold over the first 5 days (Fig. 2A). Surprisingly, at day 3 post-immunisation, initial recoil velocity was decreased by 29% to 0.30µm/s-1 (Fig. 2B-D, and Supplementary Movie. S2A) despite a 1.5-fold increase in tissue mass (Fig 2A). In contrast, 2 days later (at day 5 post-immunisation), mean initial recoil velocity was 60% higher than in the steady state at 0.71µm/s-1 (Fig. 2B-D, and supplementary Movie. 2B). Since the extracellular matrix scaffold is fragmented at this phase of tissue expansion9, we hypothesised that the actin cytoskeleton must be the major contributor to the increased tissue tension. Indeed, recoil measurements following laser ablation of tissues pre-treated with ROCK inhibitor (Y27632) to inhibit cytoskeletal contractility19, demonstrated that tissue tension was reduced to basal levels at all time points tested (Fig. 2C, Fig. S2B-F). These results indicate that tissue tension is determined by differential resistive actomyosin forces (Fig.2C).
5 days post immunisation, the whole tissue is more deformable3, yet tension measured through the fibroblastic reticular network increases (Fig. 2B-C). Since elasticity of gels is known to scale as the inverse of mesh size21, the combination of the fragmentation of the matrix and increasing mesh size of the FRC network may explain increased tissue deformability through acute expansion. To investigate how the FRC network responds mechanically to the increasing tissue size (Fig. 2A) we compared the cytoskeletal and matrix structures of the FRCs at day 3 (lower tension) versus day 5 (higher tension). In reactive lymph nodes, 3 days after IFA/OVA immunisation, F-actin structures were located primarily proximal to the basement membrane (asterisk), but T-cell facing F-actin cables were absent (Fig. 2E, and Supplementary Movie S2C). However, 5 days post immunisation, prominent contractile F-actin cables structures span the length of FRC cell bodies even in the absence of underlying matrix bundles (arrowhead) (Fig. 2E, and supplementary Movie S2D). These data show that tissue tension varies in response to immunogenic challenge and that increased tissue tension occurs independently of extracellular matrix integrity. Instead, mechanical forces are generated by increased packing of lymphocytes in the FRC meshwork and resisted by actomyosin through the FRC network (Fig. 2).
Since the FRC network is not resisting increased packing of lymphocytes at day 3 (Fig. 2C), regulation of actomyosin contractility alone cannot explain how the FRC network remains connected as the tissue mass increases (Fig. 2A). We next sought to understand the mechanisms controlling the cell surface mechanics of FRCs. To increase mesh size without increasing FRC number3,4, FRCs must elongate or make protrusions, adapting their cell morphology to maintain network integrity3,11. Effective membrane tension (hereafter membrane tension), is determined by the in-plane tension of the lipid bilayer and the strength of membrane-to-cortex attachments22,23 and can regulate cell spreading, changes in morphology and cell fate22–24. Cell contact between antigen-presenting dendritic cells and FRCs through CLEC-2/podoplanin binding, peaking at day 3, is known to regulate lymph node deformability and tissue expansion3,4. Therefore, we asked whether the CLEC2/podoplanin interaction regulates FRC cell surface mechanics. We used optical tweezers to measure membrane tension of FRCs, (Fig. 3A, Fig. S3G, and supplementary Movie. S3A) and find that specific engagement of CLEC-2 to podoplanin reduces membrane tension (Fig. 3A-B) and downregulates phosphorylation of ezrin, radixin and moesin family proteins (pERM) (Fig. S3F), which tether the actin cytoskeleton to the plasma membrane. Signalling specificity was confirmed by exogenous expression of a PDPN mutant that cannot bind CLEC-2 (T34A) and a mutant which cannot signal through the cytoplasmic tail (S167A-S171A) (Fig. S3A-D). Both showed no change in membrane tension in the presence of CLEC-2 (Fig. S3A-D). As podoplanin interacts with CD4425, also known to bind ezrin26, we tested the relatively contributions of both CD44 and podoplanin to membrane tension in unstimulated FRCs and find that podoplanin is the key driver in steady state (Fig. 3C, Fig. S3E)
In the tissue, FRCs must each elongate and form protrusions to interdigitate with their network neighbours as they spread and expand the stromal cell network. These cell shape changes can be achieved by increasing the ratio of plasma membrane surface area to cell volume through exocytotic pathways or by unfolding membrane reservoirs27. FRCs contain active EHD2+ caveolae structures28 contributing to tension-sensitive plasma membrane reservoirs (Fig. 3D). We tested, using osmotic shock (Fig. S4A-B) whether regulation of membrane tension through the CLEC-2/podoplanin signalling axis impacted how rapidly FRC could utilise existing membrane reservoirs to respond to the external forces (Fig 3E-F). We found that CLEC-2 engagement to podoplanin (Fig. 3E, Fig. S4C, and supplementary Movie. S4-B) or knockdown of podoplanin (PDPN KD) (Fig. 3F, and supplementary Movie. S4C) permitted more rapid cell expansion in hypotonic conditions but did not alter total membrane availability (Fig. S4D, and supplementary Movie. S4D). In vivo, we observed individual, differentially labelled, neighbouring FRCs increasing cell-cell contact 3 days post immunisation (Fig. 3G), suggesting that additional plasma membrane is used for morphological adaptations and incorporated into cell extensions. We propose that changing cell surface mechanics of FRCs in combination with reduced actomyosin resistance (Fig. 2C) permits lymph node expansion whilst maintaining FRC network integrity.
However, at day 5, even larger number of lymphocytes puts the network under increased tension which is resisted by actomyosin contractility (Fig. 2C). We asked what the functional consequence of changing tissue tension had on tissue remodelling in response to immune challenge. Existing studies have found that expansion of FRC populations lag behind lymphocytes3–5, but it is not known how FRC proliferation is triggered or spatially regulated within the tissue. 5 days post IFA/OVA immunisation the number of proliferating FRCs (Ki67+) doubles compared to steady state (Fig. 4A-B), which correlates with increased tissue tension (Fig. 2B-D). Ki67+ proliferative FRCs were observed throughout the tissue, and we observed no specific proliferative niche surrounding blood vessels or beneath the bounding capsule (Fig. 4A), suggesting that the cue for entry into the cell cycle is not spatially restricted or limited to a subpopulation of FRCs. We hypothesised that increased mechanical tension may gate FRC entry into cell division. To test this, we blocked the increase in tissue tension unilaterally in vivo through pharmacological inhibition of ROCK19 (Fig. 4C-D) and found that proliferation of stromal cells, specifically the FRC population, was significantly attenuated 5 days post immunisation, while lymphocyte proliferation and the increases in tissue mass and cellularity were unaffected (Fig. 4E-I, Fig. S5A-D). This leads us to conclude that the stromal architecture is reactive to the physical space requirements of the lymphocyte populations. Indeed, previous studies have identified a remarkably robust ratio between fibroblastic stroma and T cells, maintained as the tissue expands29. A mechanical cue for stromal cell growth and proliferation would ensure that the steady state ratio of fibroblastic stroma to lymphocytes is perfectly restored, independently of the kinetics or scale of the immune reaction. Homeostatic tissue architecture is recovered as the tissue expands reinstating the supportive immune microenvironment for lymphocyte populations30.
Together our data demonstrate that the fibroblastic structure of the lymph node is the active mechanical component during tissue expansion. Using the dynamic cellular network rather than the more rigid ECM to respond to changing lymphocytes numbers in the tissue provides the lymph node with an elegant mechanical system that can proportionately respond to lymphocyte requirements. We show that the lymph node becomes mechanically permissive to expansion through CLEC-2/PDPN signalling priming the cell surface mechanics of FRCs. As the kinetics of expansion increase tissue tension, FRC’s initiate cellular division to restore tissue architecture. Other studies have shown that tissue scale properties emerge from cellular scale mechanics in the transition from developing to adult tissue31. We now directly address this concept in an immunological relevant adult mammalian tissue during homeostasis and immune challenge. We show that the interconnected cellular network deploys molecular signals controlling cellular mechanical properties to collectively determine tissue scale mechanics of lymph nodes.