C57BL/6J, B6.SJL-Ptprca PepCb (CD45.1), and CAG-EGFP (Okabe et al., 1997) were bred and maintained under specific pathogen-free conditions in the Mayo Clinic animal facility. All strains were sourced from Jackson Laboratories, including aged C57BL/6J mice, in this study 16–20 months old. To reduce sex-related variability (Klein and Flanagan 2016), aged mice in the study were all female. Aged mice were monitored and health status visually assessed, and mice were removed from the study if in declining health. Mouse maintenance and experimental procedures were carried out with approval from the Institutional Animal Care and Use Committee at the Mayo Clinic.

CD8⁺ naïve T cells were enriched from CAG-EGFP mouse spleens using EasySep Mouse Naïve CD8⁺ T cell isolation Kit (Stem Cell 19858) per the manufacturer’s instructions. CD8⁺ naïve T cells were confirmed to be >98% pure by flow cytometry. ∼10⁶ cells in 100 μL of PBS were injected retro-orbitally into B57BL/6 mice 24 h before imaging. Inguinal lymph nodes for imaging were dissected and embedded in 4% (w/v) NuSieve GTG low-melting-temperature agarose (Lonza) in PBS at 37°C. The solidified agarose block was sectioned into 300-μm-thick lymph node slices using a VT1200S Microtome (Leica) in a bath of ice-cold PBS, with speed at 0.20 mm s⁻¹ and amplitude at 0.6 mm (Lancaster and Ehrlich 2017). Slices were collected in PBS + 10% FBS and kept on ice until ready for imaging.

Lymph node slices were transferred and secured in an imaging chamber (Harvard Apparatus) on the microscope stage. Perfusion medium, consisting of phenol red-free RPMI (Gibco) supplemented with 2 g L⁻¹ sodium bicarbonate, 5 mM HEPES, and 1.25 mM calcium chloride, was circulated through the imaging chamber using a micro-perfusion high-flow pump (Bioptechs). The perfusion medium was aerated with 95% O₂/5% CO₂ and maintained at 37°C with a heated microscope stage and inline perfusion heater (Harvard Apparatus). Images were acquired every 15 s, through a depth of 40 μm, at 5-μm intervals for durations of 15–20 min, using an Investigator microscope (Bruker) with a 20× water immersion objective (Olympus, NA 1.0) and PrairieView software (v.5.5, Bruker). The sample was illuminated with a Chameleon Discovery NX ultrafast pulsed laser (Coherent) tuned to 920 nm. Emitted light was passed through 460/50 and 525/50 band-pass filters (Chroma) to separate GaAsP detectors for detection of second-harmonic generation (SHG, blue) and EGFP (green) fluorescence, respectively.

Migratory paths for T cells were tracked, and mean cell velocity and path straightness for each cell calculated using Imaris (v9.8, Bitplane). SHG signal was rendered as collagen via Surfaces. Distance Transformation function in ImarisXT was used to calculate distances between T cells and collagen Surfaces for every time point, and T cell Tracks were filtered as being proximal to collagen if the minimum Track distance was less than 12 μm at any time point.

Inguinal, axillary, brachial, cervical, and mesenteric lymph nodes from mice were harvested and immediately submerged in 1% paraformaldehyde in PBS overnight and fixed overnight in 20% sucrose (w/v) in PBS (Choi et al., 2020). Fixed lymph nodes were embedded in Optimal Cutting Temperature (OCT) compound (Ted Pella) within TissueTek Cryomolds (VWR) and snap frozen using liquid nitrogen or isopentane with dry ice. Frozen tissue blocks were then sectioned on a HM525NX Cryostat (Thermo Scientific) at 16-μm thickness at −15°C onto poly-L-lysine-coated microscope slides (VWR).

To stain, defrosted slides were fixed for 20 min in ice-cold acetone then washed in PBS and PBS-T (PBS + 0.1% Tween-20). Sections were blocked for 15 min in TNB blocking buffer [0.1 M Tris-HCl, pH 7.5; 0.15 M NaCl; and 0.5% (w/v) blocking reagent (Perkin Elmer FP1020)]. Sections were stained overnight at 4°C with the following primary antibodies diluted in TNB blocking buffer: rabbit polyclonal α-mouse alpha-smooth muscle actin (1:200, Abcam 32575), rabbit polyclonal α-mouse collagen IV (1:200, Abcam 6586), rabbit polyclonal α-PDGFRβ (1:200, Abcam 6586), rat α-mouse/human B220 (1:400, RA3-6B2, Biolegend 103201), Syrian hamster α-mouse gp38/podoplanin (1:200, 8.1.1, Biolegend 127401). Sections were washed with PBS-T before incubation for 1 h at RT with secondary antibodies diluted (1:200) in TNB blocking buffer: AF488 goat α-rat IgG (1:400 dilution; Biolegend Poly4054), AF488 goat α-Syrian hamster IgG (Jackson Immunoresearch 107-545-142), AF594 donkey α-rabbit IgG (Invitrogen A32754), Cy3 donkey α-rabbit IgG (Jackson Immunoresearch 711-165-152). Nuclei were stained with 3 μM DAPI and coverslips sealed with Prolong Gold antifade reagent (Sigma Aldrich). Sections were imaged using a LSM 800 confocal microscope (Zeiss) equipped with Plan-Neofluar 20× objective lens (Zeiss, NA 0.4). Quantitative comparison of αSMA fluorescence density was conducted using Imaris on 20× magnification images. Threshold for αSMA^hi was set based on aged lymph node samples, and used to render Surfaces. Total volume of Surfaces was calculated as a percent of total field of view.

Inguinal, axillary, brachial, superficial cervical, and mesenteric lymph nodes from mice were dissected into plastic cell culture dishes containing phosphate buffered saline (PBS). Excess fat was trimmed away using surgical scissors under a dissection stereomicroscope. Lymph nodes were enzymatically digested into single cell suspension as follows (Fletcher et al., 2011; Knop et al., 2020): lymph nodes were cut in half using surgical scissors and incubated at 37°C, 5%CO₂ for 30 min in separate wells of a twelve-well plate with 1 ml each of an enzyme digest cocktail, which consisted of complete RPMI 1640 medium [RPMI 1640 (Corning) and 10% fetal bovine serum (Gibco), supplemented with 100 U/mL penicillin-streptomycin (Thermo Fisher), non-essential amino acid solution (Sigma Aldrich), 1 mM sodium pyruvate, GlutaMAX (Thermo Fisher), and β-mercaptoethanol (Sigma Aldrich)] with 0.2 mg/ml of Dispase II (Sigma D4693), 0.4 mg/ml Collagenase P (Roche 11213857001), and 0.010 mg/ml of DNAse I (Roche 04536282001)]. Another 1 ml of the enzyme digest cocktail was added to each well and the well contents gently agitated by aspiration using a glass pipette before a second 30 min incubation. The tissue was completely dissociated by more aggressive aspiration using a glass pipette, and the digest was quenched by addition of 0.5 ml of complete RPMI 1640 medium containing 10 mM EDTA and transferred through 100-μm mesh filters.

Lymph node stroma were enriched in the young and aged samples by depletion of CD45⁺ hematopoietic cells using magnetic bead-based cell separation as follows (Fletcher et al., 2011): samples were centrifuged and resuspended in 90 μL of stromal wash buffer (2% FBS in PBS with 5 mM EDTA) per 10⁷ cells. 7 μL of CD45 Microbeads (Miltenyi Biotech 130-052-301) were added per 10⁷ cells and incubated on ice for 20 min. The samples were washed and resuspended in stromal wash buffer. CD45⁺ cells were collected by passing the samples through LD Columns (Miltenyi Biotech 130-042-901) supported by a QuadroMAX Separator (Miltenyi Biotech 130-090-976). Column flow-through was collected as enriched CD45⁻ lymph node stroma. Stroma was confirmed to be >98% CD45⁻ by flow cytometry.

Up to 10⁷ cells were incubated for 20 min on ice in 100 μL of PBS + 2% FBS with fluorochrome-conjugated antibodies directed against the following mouse targets: CD3-PECy5 (145-2C11, Biolegend 100309), CD19-PECy5 (6D5, Biolegend 115509), CD31-APCCy7 (390, Biolegend 102439), CD35-AlexaFluor700 (7E9, Biolegend 123431), CD45-PECy5 (30-F11, Biolegend 103109), B220-BV711 (RA3-6B2, Biolegend 103255), gp38-FITC (8.1.1, Biolegend 127415), ICAM-1-BV421 (YN1/1.7.4, Biolegend 116141), LTBR-PECy7 (5G11, Biolegend 134409), MAdCAM-1-APC (MECA-367, Biolegend 120711), SCA-1-BV605 (D7, Biolegend 108133), VCAM-1-PE (429 (MVCAM.A), Biolegend 105713). Antibodies were diluted (1:200) from stock concentrations of 0.5 mg ml⁻¹. Cells were washed and resuspended in 10 μg ml⁻¹ propidium iodide (PI) to determine viability. Samples were analyzed on an LSR Fortessa or Symphony flow cytometer (BD Biosciences) and data were analyzed using FlowJo (v.10, TreeStar).

For stromal subset sorting, pooled CD45-depleted cells from lymph nodes were incubated for 20 min on ice in 100 μL of PBS + 2% FBS with fluorochrome-conjugated antibodies directed against the following mouse targets: CD31-APC (MEC13.3, Biolegend 102509), CD35-APCCy7 (7E9, Biolegend 123417), CD45-APCCy7 (30-F11, eBioscience 47-0451-82), gp38-FITC (8.1.1, Biolegend 127415), and ICAM-1-BV421 (YN1/1.7.4, Biolegend 116141). Antibodies were diluted (1:200) from stock concentrations of 0.5 mg ml⁻¹. Cells were washed and resuspended in 10 μg ml⁻¹ PI to determine viability. Samples were sorted on an Aria FACS (BD Biosciences) based on the following criteria: fibroblastic reticular cells (FRCs; CD45⁻CD35⁻ICAM-1⁺gp38⁺CD31⁻), lymphatic endothelial cells (LECs; CD45⁻CD35⁻ICAM-1⁺gp38⁺CD31⁺), blood endothelial cells (BECs; CD45⁻CD35⁻ICAM-1⁺gp38⁻CD31⁺), and double-negative cells (DNs; CD45⁻CD35⁻ICAM-1⁺gp38⁻CD31⁻). Samples were confirmed to be >95% pure.

Spleens were dissected from 2-3-month-old CD45.1 mice and mechanically dissociated into a single cell suspension. Splenocytes were labelled with CellTrace Violet (Biolegend C34571), and CD4⁺ naïve T cells enriched using the EasySep Mouse Naïve CD4⁺ T cell isolation kit (StemCell), all per manufacturers’ instructions. CD4⁺ naïve T cells were confirmed to be >95% pure by flow cytometry. 2 × 10⁶ naïve T cells were adoptively transferred via intravenous injection into sub-lethally irradiated (600 cGy in split doses) 8-week-old and 17–20-month-old C57BL/6J recipient mice. After 6 days, spleens and lymph nodes were dissected and pooled from recipients. Up to 10⁷ cells were incubated for 20 min on ice in 100 μL of PBS + 2% FBS with fluorochrome-conjugated antibodies directed against the following mouse targets: CD3-APCCy7 (17A2, Biolegend 100222), CD4-BV510 (RM4-5, Biolegend 100553), CD8-BUV395 (53-6.7, BD Biosciences 563786), CD44-AF700 (IM7, Invitrogen 56-0441-82), CD45.1-PECy7 (A20, Biolegend 110730), CD45.2-AF700 (104, Biolegend 109822), and CD62L-BV605 (MEL-14, Biolegend 104437). Antibodies were diluted (1:200) from stock concentrations of 0.5 mg ml⁻¹. Cells were washed and resuspended in 10 μg ml⁻¹ PI to determine viability. For analysis of phosphorylated STAT5, cells were stained for surface markers and with fixable viability dye Zombie Red (Biolegend 77475) before fixation using FOXP3/Transcription Factor fixation kit (eBioscience), per manufacturer’s instructions, and intracellular staining with pSTAT5 (pY694)-FITC (BD Bioscience 612598) or isotype control (BD Phosflow mouse IgG1κ, BD Bioscience 557782). Antibodies were diluted (1:200) from stock concentrations of 0.5 mg ml⁻¹. Samples were analyzed on an LSR Fortessa or Symphony flow cytometer (BD Biosciences) and data were analyzed using FlowJo (v.10, TreeStar).

Lymph nodes were dissected from 2-3-month-old and 16–20 month-old C57BL/6J mice and enzymatically digested as above. Stroma were cultured as follows (Link et al., 2007; Fletcher et al., 2011): 10⁶ lymphocytes in complete RPMI medium were plated in 96-well flat-bottom plates and incubated overnight at 37°C, 5% CO₂. Non-adherent cells were gently removed by repeated washing with PBS, and the adherent stromal cells were again incubated in complete RPMI medium for 48 h.

Lymphocytes from 2-3-month-old CD45.1⁺ mice were mechanically disassociated into a single-cell suspension. 10⁶ lymphocytes were added to wells with stromal cells, along with fresh complete RPMI medium. In some wells, 0.1 ng ml⁻¹ recombinant murine IL-7 (Cell Signaling) and/or 10 μg ml⁻¹ α-IL-7 (M25, BioXCell BE0048), were added. The cells were incubated for 3 days before staining by incubation for 20 min on ice in 100 μL of PBS + 2% FBS with fluorochrome-conjugated antibodies directed against the following mouse targets: CD3-BUV395 (17A2, BD Biosciences 740268), CD4-APC (GK1.5, Biolegend 100412), CD8-Pacific Blue (53-6.7, Biolegend 100725), CD44-PE (IM7, Biolegend 103008), CD45.1-PECy7 (A20, Biolegend 110730), CD45.2-AF700 (104, Biolegend 109822), and CD62L-FITC (MEL-14, BD Pharmingen 553150). Cells were washed and resuspended in 10 μg ml⁻¹ PI to determine viability. Cell numbers were calculated with the addition of 2 × 10⁴ polystyrene beads (Polysciences). Samples were analyzed on an LSR Fortessa or Symphony flow cytometer (BD Biosciences) and data were analyzed using FlowJo (v.10, TreeStar).

RNA from up to 10⁶ cells was purified using the RNeasy Mini Kit (Qiagen, 74104) according to manufacturer’s instructions, and RNA concentration was then measured by NanoDrop One (Thermo Scientific). cDNA was synthesized using SuperScript^® IV First-Strand Synthesis SuperMix (Invitrogen, 18091050) per the manufacturer’s instructions, on a C1000 Touch thermocycler (BioRad). Gene expression levels of Ccl19, Ccl21, Il7, LIGHT, Lta, Ltb, and Ltbr were measured relative to Actb using LightCycler 480 SYBR Green I Master (Roche, 04707516001) on an LightCycler 480 system (Roche) following the manufacturer’s instructions, using the following primers: Actb: forward- CAT​TGC​TGA​CAG​GAT​GCA​GAA​GG, reverse- TGC​TGG​AAG​GTG​GAC​AGT​GAG​G; Ccl19a: forward- CTG​CCT​CAG​ATT​ATC​TGC​CAT, reverse-AGGTAGCGGAAGGCTTTCAC (Chai et al., 2013); Ccl21a: forward- AAG​GCA​GTG​ATG​GAG​GGG​GT, reverse- CTT​AGA​GTG​CTT​CCG​GGG​TG (Chai et al., 2013); Il7: forward- CTG​ATG​ATC​AGC​ATC​GAT​GAA​TTG​G, reverse-GCAGCACGATTTAGAAAAGCAGCTT (Martin et al., 2017); LIGHT: forward- CAG​GCC​CCT​ACA​GAC​AAC​AC, reverse- ACT​CGT​CTC​CCA​CAA​GGA​ACT (Seach et al., 2008); Lta: forward- GCT​TGG​CAC​CCC​TCC​TGT​C, reverse- GAT​GCC​ATG​GGT​CAA​GTG​CT (Seach et al., 2008); Ltb: forward- CCA​GCT​GCG​GAT​TCT​ACA​CCA, reverse- AGC​CCT​TGC​CCA​CTC​ATC​C (Seach et al., 2008); Ltbr: forward-TGCTCTTCACCACTGTCCTG, reverse- GAA​ATG​TGG​GTC​GGC​TCT​T (Grandoch et al., 2015).

All statistical analyses were performed using Prism (v. 9, GraphPad) with the corresponding test listed in the Figure Legends.

Given that T cells migrate within the lymph node to encounter antigen-presenting cells bearing their cognate antigen and to engage in survival crosstalk, we investigated whether the aged lymph node microenvironment could impact the ability of naïve T cells to efficiently navigate the lymph nodes. GFP-expressing naïve T cells, harvested from young (2–3 month-old) mice, were adoptively transferred into young and aged (16–20 month-old) C57BL/6J hosts 24 h prior to imaging. Lymph nodes were harvested and prepared as slices for immediate ex vivo imaging, a well-established technique for studying lymphocyte dynamics within an intact lymph node microenvironment (Bajénoff et al., 2006). Collagen is a birefringent material, and thus can be observed label-free by 2-photon microscopy via second harmonic generation (Bougherara et al., 2015). Indeed, young lymph nodes were encapsulated within a collagen capsule, with collagen also prominent in the medullary sinuses, but no collagen was observed within the T-cell zones. (Figure 1A, left panels). Aged lymph node slices, on the other hand, were much less consistent in their structure, with some regions bearing resemblance to the typical kidney bean-shaped form, but many other slices with an irregular outline. Such aged slices harbored varying degrees of encroachment by fibrosis into the T cell zones (Figure 1B, left panels). For aged slices with increased collagen signal and irregular shape, T cell zones were highly constricted, though aged slices with a more typical lymph node shape still retained identifiable T cell zones. GFP⁺ T cells were dispersed throughout expansive T cell zones that flank the medulla in young slices, and live imaging revealed that they migrate quickly and randomly (Figure 1A, right panels; Supplementary Movie S1). In some aged slices, collagen formed a mesh within the T cell zone, which severely restricted the motility of T cells (Figure 1B, middle panels; Supplementary Movie S2), but for slices in which collagen has not encroached too deeply, T cell velocity was high (Figure 1B, right panels; Supplementary Movie S3).

The mean T cell velocities and path straightesses were quantified between the young and aged slices. Naïve T cells traveling within aged lymph node slices had significantly reduced mean cell velocities than those on young slices (Figure 1C, left panel). The mean path straightness, a parameter which relates the overall length of the T cell path travels to total displacement, was also reduced for cells in aged slices (Figure 1C, right panel), indicating more restricted migration trajectories in the aged microenvironment. As the speed of T cell migration within aged slices appeared to depend on whether fibrosis was prominent in the tissue parenchyma, the data were stratified based on T cells that were distal from collagen fibers and those that were proximal to collagen, within one cell-length (<12 μm). The mean velocity of T cells migrating distal to collagen were significantly greater than those proximal to collagen in both lymph node ages (Figure 1D, left panel). T cells migrating proximal to collagen were significantly slower in both tissues, with mean average velocities of ∼8–9 μm min⁻¹ (Figure 1D, left panel). However, T cells far from collagen in young slices were still faster than those far from collagen in aged slices, as were T cells close to collagen (Figure 1D, left panel). In young slices, ∼12% of T cells were near collagen fibers, whereas ∼20% of T cells were near collagen fibers in aged slices, indicating that a greater proportion of T cells were slowed in the aged microenvironment. For both ages of lymph node slices, T cells traveled with reduced path straightnesses when proximal to collagen, indicating confined trajectories with little displacement (Figure 1D, right panel). However, T cells both distal and proximal to collagen sources in aged slices traveled in a more confined pattern than their counterparts in young slices (Figure 1D, right panel). Thus, our imaging studies demonstrated that young naïve T cells traveled with less velocity and straightness when placed in an aged rather than a young lymph node microenvironment. Decreased motility was associated with increased deposition of fibrosis into the T cell zone of aged lymph nodes slices, which severely limited the motility of T cells.

As aging in mouse models is often studied at >20 months, we confirmed that structural dysregulation of lymph node compartments occurred at our relatively earlier time points of ≥16 months. In young adult mice, we observed fine compartmentalization of B cell follicles in the subcapsular region from the T cell zone paracortex, by staining B cells using an antibody directed against B220 and platelet-derived growth factor-beta⁺ (PDGFRβ⁺) fibroblasts (Figure 2A, left panel). With age, the lymph node was rounded and compacted, as in our 2-photon imaging, and B cell areas were no longer definable as follicles (Figure 2A, right panel). Rather, PDGFRβ staining was now spread throughout the lymph node and mixed with B220⁺ areas, though its PDGFRβ staining pattern was yet diffuse when compared to the T cell zone of the young lymph node. Thus, aged murine lymph nodes have lost the compartmentalization of the B cell follicles by 16 months of age.

To confirm the SHG fluorescence of fibrosis observed by 2-photon microscopy, we co-stained for collagen IV and PDGFRβ in young and aged lymph node sections. Collagen IV staining the capsular sheath and perivascular cross-sections within the young lymph nodes (Figure 2B, left panel). In the aged lymph node, the capsule was also intensely stained for collagen IV; reticular staining in the paracortex and staining of collagen enwrapping the vessels were more prominent than visualized in young lymph nodes (Figure 2B, right panel). We further stained for alpha-smooth muscle actin (αSMA) against the fibroblast marker podoplanin (PDPN; gp38). Like PDGFRβ, αSMA is a mesenchymal fibroblast specific marker (Chai et al., 2013; Choi et al., 2020). PDPN is expressed by FRCs and lymphatic endothelial cells (LECs), and is a glycoprotein responsible for the contractility of lymph node stroma (Astarita et al., 2015). In the young lymph node, PDPN stains a fine network through the paracortex and in the medullary sinuses, with αSMA interspersed between PDPN⁺ reticular cells of the T cell zone and most prominent in the lining of vessels of the medulla (Figure 2C, left panel and Figure 2C, top row). Though captured at the same settings as the young lymph nodes, the aged lymph node sections stained more intensely for αSMA (Figure 2C, right panel and Figure 2D, bottom row). While the SHG signal captured by 2-photon microscopy emphasized fibers in the subcapsular and medullary areas, staining for αSMA was most dense in the cortical regions. Closer observation of the PDPN channel in the cortex reveal that the network is still maintained, albeit with increased density that is thoroughly intercalated with αSMA (Figure 2D, bottom row, and Figure 2E). Thus, immunofluorescence for fibroblast-associated markers indicated increased deposition of collagen and αSMA within the architecture of the aged lymph node.

Our 2-photon imaging demonstrated that T cell motility was impaired in the aged lymph node, especially when in proximity to collagen deposition; however, there was no obvious change in the capacity for young naïve T cells to home to the aged lymph nodes, and many T cells were able to migrate at high velocity when far from collagen fibers. To validate our aged mouse model as a context in which naïve T cell survival was impaired, we recapitulated experiments conducted by Becklund and colleagues to confirm that naïve T cell homeostatic proliferation was indeed decreased in aged mice (Becklund et al., 2016). Cell Trace Violet (CTV)-labelled congenic CD4⁺ naïve T cells from young mice were adoptively transferred into young and aged mice. Hosts were sub-lethally irradiated before adoptive transfer so that T cells would undergo non-activating homeostatic proliferation in order to repopulate T cell numbers. Naïve T cells diluted CTV in response to homeostatic proliferation to a greater extent when place within young rather than aged hosts, indicating that they underwent more rounds of homeostatic proliferation (Figures 3A,B). Notably, the proportion of transferred young naïve T cells detected in young hosts was also significantly greater than that in aged hosts (Figure 3B). To determine whether changes in homeostatic proliferation were due to differences in signaling through IL-7 receptor, we measured the mean fluorescence intensity (MFI) of CD127 (IL-7Rα chain) in the transferred cells, as the receptor is internalized in response to IL-7 signaling (Martin et al., 2017). However, we did not measure any significant difference in the MFI of CD127 for the congenic naïve T cells within young or aged hosts (Figure 3C, left panel), nor did we measure any significant difference in T cell phosphorylation of STAT5, downstream of IL-7R signaling (Figure 3C, right panel). Thus, homeostatic proliferation of young naïve T cells was impaired within the context of the aged microenvironment, but not associated with any obvious differences in IL-7 signaling.

In order to determine if decreased homeostatic proliferation observed in vivo was due to cell-specific changes to aged stroma, we adopted a well-established lymph node stroma co-culture assay (Link et al., 2007; Fletcher et al., 2011). Young congenic lymphocytes were incubated with CD45-depleted young or aged lymph node stromal cells, and the survival of naïve T cells was measured after 3 days. We measured no significant difference in the survival of naïve CD4⁺ or CD8⁺ T cells when co-cultured with stroma from young or aged lymph nodes (Figure 3D). Addition of a low concentration of recombinant murine IL-7 (rIL-7) significantly increased naïve T cell numbers when co-cultured with stroma, and increased survival was negated by inclusion of IL-7-blocking antibody (αIL-7) (Figure 3D). We concluded from our in vitro assays that bulk aged stroma do not have an inherent defect in supporting naïve T cells.

To determine whether aging impacted the homeostatic function of the lymph node stroma, we surveyed the global expression of Il7, a T cell maintenance cytokine, and chemokine ligands Ccl19 and Ccl21, which promote T cell homing to and interactions within the secondary lymphoid organs (Link et al., 2007). We furthermore quantified stromal expression of Ltbr, as signaling through this receptor enforces development of lymph node stroma (Bénézech et al., 2010; Chai et al., 2013). We found no significant difference in the expression of Il7 or Ccl19 when comparing young and aged stroma (Figure 4A). However, we found significant decreases in the aged stromal expression of Ccl21 and Ltbr (Figure 4A). Both Ltbr and Ccl21 were highly expressed in FRCs from sorted adult stroma (Supplementary Figure S1), suggesting that the FRC population may play a predominant contribution to the age-associated decline observed. Since Ltbr expression was decreased with age, we measured the expression of lymphotoxin ligands Lta, Ltb, and LIGHT in bulk splenocytes from young and aged mice. As stroma constitute <1% of the cellularity of the secondary lymphoid organs, they would have a negligible impact in the measurement of these ligands in hematopoietic cells. We found no significant difference in the expression of Lta, Ltb, and LIGHT between young and aged cells (Figure 4B). This indicated that decreased Ltbr expression by aged stroma was not associated with changes in ligand expression. Thus, we determined that aging was associated with decreased transcript expression of Ltbr in stroma, and the homeostatic chemokine Ccl21.

We used quantification via flow cytometry to determine whether aging perturbs the composition of stromal subsets. CD45-depleted cells from enzymatically digested lymph nodes had been separated into skin-draining (sdLN) and mesenteric (mLN) groups, as site-specific differences have been previously reported for young adult mice (Fletcher et al., 2011). FRC, LEC, blood endothelial cell (BEC), and double-negative (DN) subsets were resolved based on PDPN/gp38 and CD31 gating (Figure 5A). We found no significant difference in the frequency of all stromal subsets between young and aged sdLNs (Figure 5B, left panel). This was further shown in the similar numbers of stromal subsets between young and aged sdLNs (Figure 5C, left panel), which reflected the similar overall cellularity of the total lymph node that we observed at this age range (Supplementary Figure S2A). However, we found that FRCs were significantly decreased in frequency within aged mLNs, which was compensated by increased proportions of BECs and DNs (Figure 5B, right panel). This corresponded to a decrease in the FRC numbers in the aged mLNs, though the increase in BEC and DN proportion did not result in a statistically significant increase in their numbers (Figure 5C, right panel). We also observed a trend towards decreased LECs in the aged mLN, though these differences did not rise to the level of statistical significance (Figures 5B,C).

Heterogeneity has been noted within the FRC compartment (Rodda et al., 2018; Pradhan et al., 2021), and MAdCAM-1 expression delineates marginal zone reticular cells (MRCs), which are found in the subcapsular zone and mainly destined to become follicular dendritic cells (Jarjour et al., 2014); we therefore gated out MAdCAM-1⁺ cells from our analysis of FRCs that modulate naïve T cell activity (Supplementary Figure S2B). We subsequently used VCAM-1 upregulation to distinguish stages of FRC differentiation (Bénézech et al., 2010; Fletcher et al., 2015) in the FRCs from young and aged lymph nodes (Figure 5D). We found significantly decreased proportions of VCAM-1^lo FRCs in aged sdLNs, with a concomitant increase in aged VCAM-1^int FRC frequency (Figure 5E, top panel). Though the overall proportion of FRC from mLN stroma was impacted by aging (Figure 5B, right panel), there was no difference in the distribution of mLN FRC subsets with age (Figure 5E, bottom panel). Thus, we determined that aging only impacted FRC composition at certain lymph node sites, which include the mesenteries. For the skin-draining sites, there was no overall change in FRC proportion, but differences were observed within the compartment when subsetted by VCAM-1 expression.

Flow cytometric analysis further suggested that not only are aged FRC subsets quantitatively different with age, but qualitatively different. We measured surface expression of LTBR, which is highly expressed on FRCs and their progenitors in order to enforce their maturation (Bénézech et al., 2010; Fletcher et al., 2015). In young VCAM-1^lo FRCs, LTBR was most highly expressed, and gradually decreased during VCAM-1 upregulation in the sdLN (Figure 5F, top panel). Interestingly, while LTBR expression was similar in young and aged VCAM-1^hi FRCs, it was maintained at the same expression level throughout VCAM-1 upregulation (Figure 5F, top panel), indicating that aged FRC progenitors did not have a period of upregulated LTBR expression. Interestingly, the trajectory of LTBR expression was different in the mLNs; in young FRCs, LTBR expression remained constant throughout VCAM-1 upregulation (Figure 5F, bottom panel). LTBR expression was also fairly constant through the aged FRC stages but was about half the magnitude of that in young FRCs (Figure 5F, bottom panel). We further analyzed surface expression of the cell renewal marker SCA-1 (Ly6a), which is also associated with mesenchymal stem cells (Morikawa et al., 2009). In both the sdLN and mLN, SCA-1 was highly expressed in VCAM-1^lo FRCs and gradually decreased with VCAM-1 upregulation (Figure 5G). However, the surface expression of SCA-1 was significantly lower in aged VCAM-1^lo FRCs at both sites and remained at low levels throughout VCAM-1 upregulation (Figure 5G). Thus, we have detected differences in the frequency of and surface expression of characteristic markers such as LTBR and SCA-1 in VCAM-1^lo FRCs, the precursors to mature T cell zone FRCs.