IL‐7 derived from lymph node fibroblastic reticular cells is dispensable for naive T cell homeostasis but crucial for central memory T cell survival

The survival of peripheral T cells is dependent on their access to peripheral LNs (pLNs) and stimulation by IL‐7. In pLNs fibroblastic reticular cells (FRCs) and lymphatic endothelial cells (LECs) produce IL‐7 suggesting their contribution to the IL‐7‐dependent survival of T cells. However, IL‐7 production is detectable in multiple organs and is not restricted to pLNs. This raises the question whether pLN‐derived IL‐7 is required for the maintenance of peripheral T cell homeostasis. Here, we show that numbers of naive T cells (TN) remain unaffected in pLNs and spleen of mice lacking Il7 gene activity in pLN FRCs, LECs, or both. In contrast, frequencies of central memory T cells (TCM) are reduced in FRC‐specific IL‐7 KO mice. Thus, steady state IL‐7 production by pLN FRCs is critical for the maintenance of TCM, but not TN, indicating that both T cell subsets colonize different ecological niches in vivo.

The maintenance of the T N pool relies on the accessibility of secondary lymphoid organs (SLOs) where IL-7 is produced by lymphoid stromal cells (LSCs) [9]. In peripheral lymph nodes, for example, fibroblastic reticular cells (FRCs) and lymphatic endothelial cells (LECs) are the main sources of IL-7 [9]. Co-culture experiments demonstrated that FRC-derived IL-7 promotes T cell survival [9]. IL-7 binds to the ECM [10,11] suggesting that it might exert its function in close vicinity to the site of production. Due to the facts that T cell survival is impaired in vivo if either IL-7 action or peripheral LN (pLN) access is blocked [1,4,9,12,13], it has been proposed that circulating T N receive IL-7-dependent survival signals in pLNs [14][15][16][17]. Since T CM and T N have similar migration patterns in vivo [18] and both rely on IL-7 [1,4], pLN FRCs are supposed to be critical for the IL-7-dependent persistence of both T cell subsets in vivo [15,17]. A potential contribution of LEC-derived IL-7 has been suggested as well [19].
However, various non-hematopoietic stromal cells express IL-7 [20][21][22][23] and its steady state levels vary strongly between different organs [24,25]. For example, intestine and skin produce high levels of IL-7 in the steady state while only low levels of Il7 gene activity are detectable in the adult liver [24,25]. Since T N and T CM continuously recirculate between SLOs, blood, and lymph [26], they might utilize IL-7 derived from various organs. Hence, it remained unclear whether the maintenance of peripheral T cell homeostasis relies on the local action of IL-7 in pLNs and/or systemic effects of IL-7 produced by alternative sources.
In order to answer this question, we generated conditional IL-7 KO (IL-7 fl/fl ) mice and inactivated Il7 gene activity in a cell type-specific manner in pLNs. Here, we show that T N numbers remained unaltered in pLNs and spleens of LEC-and FRC-specific IL-7 KO (LEC IL-7 and FRC IL-7 ) mice. In apparent contrast, T CM abundance was significantly reduced in FRC IL-7 mice, an effect that was most pronounced for CD8 + T CM in pLNs. In summary, we provide evidence that FRC-derived IL-7 is dispensable for the systemic survival of T N cells. On the contrary, however, IL-7 produced by pLN FRCs is crucial for the maintenance of T CM homeostasis indicating that T N and T CM occupy different ecological niches in vivo.
Next, we compared the impact of conditional and conventional Il7 gene inactivation on IL-7-dependent lymphocyte homeostasis. While mice harboring one intact Il7 wt allele (PGK-Cre − IL-7 −/wt , PGK-Cre + IL-7 −/wt , and PGK-Cre − IL-7 −/fl mice) had comparable numbers of T and B cells in the spleen, ubiquitous Il7 gene inactivation in PGK-Cre + IL-7 −/fl mice was associated with a strong decrease of T and B cell numbers similar to IL-7 −/− mice (Fig. 1A). Importantly, the lack of IL-7 production in PGK-Cre + IL-7 −/fl and IL-7 −/− mice was accompanied by the selective reduction of CD44 lo CD62L hi CD4 + and CD8 + T N as well as the enrichment of CD44 hi CD4 + and CD8 + memory T cells (T M ; Fig. 1B and C).
In summary, IL-7-dependent T cell homeostasis is similarly impaired in IL-7 −/− and PGK-Cre + IL-7 −/fl mice thus confirming (i) the efficient Cre-mediated inactivation of the Il7 fl allele and (ii) the crucial importance of IL-7 for T N generation and maintenance. Hence, our IL-7 fl/fl mouse is a suitable tool to study the impact of pLN-specific Il7 gene inactivation on peripheral T cell homeostasis.

LEC-derived IL-7 is dispensable for peripheral T cell homeostasis
In pLNs, CD45 − stromal cells comprise gp38 + CD31 − FRCs, gp38 + CD31 + LECs, gp38 − CD31 + blood endothelial cells (BECs) and gp38 − CD31 − double negative cells (DNs) [28] (Fig. 2A). Lyve-1-expressing LECs produce IL-7 in pLNs and throughout the body [29] and are supposed to be important regulators of IL-7dependent peripheral T cell homeostasis [19]. In order to test this hypothesis, we generated LEC IL-7 mice lacking Il7 gene expression specifically in LECs. For this purpose, Lyve1-Cre-transgenic (Lyve1-Cre + ) mice [30] were crossed to IL-7 fl/fl mice. Lyve1-Cre + mice harboring at least one intact Il7 allele (LEC wt mice) served as controls. CD45 − stromal cells were purified from LNs of LEC IL-7 and LEC wt mice and relative Il7 mRNA levels were quantified by RT-qPCR. In agreement with a previous report [9], LECs produced considerable amounts of Il7 mRNA in control mice, even though tenfold less than FRCs ( Fig. 2A). Of note, Il7 mRNA levels were strongly reduced in LECs from LEC IL-7 mice indicating successful Il7 gene inactivation. On the contrary, Il7 mRNA levels in FRCs, BECs, and DNs were comparable in LEC IL-7 and LEC wt mice.
In order to study whether LEC-derived IL-7 affects peripheral T cell homeostasis, CD4 + and CD8 + T cells were quantified in pLNs and spleens of LEC IL-7 and LEC wt mice. As shown in Fig. 2B, T cell numbers were indistinguishable between both mouse lines. Furthermore, relative frequencies and numbers of CD44 lo CD62L hi T N , CD44 hi CD62L lo T EM , and CD44 hi CD62L hi T CM were comparable in pLNs and spleens ( Fig. 2C-H), although CD4 + T CM frequencies were reduced in pLNs of LEC IL-7 mice ( Fig. 2C and D). In conclusion, Il7 gene inactivation in LECs does not have major effects on quantitative and qualitative aspects of peripheral T cell homeostasis.  Absolute cell numbers of CD3 + CD4 + or CD3 + CD8 + T cells and B220 + B cells were determined in the spleen of the indicated mouse lines. (B and C) Shown are representative contour plots for the CD44/CD62L expression profiles of (B) CD3 + CD4 + or (C) CD3 + CD8 + T cells in spleen. Numbers in contour plots indicate percentages. Frequencies of naive (B) CD3 + CD4 + CD44 lo CD62L hi and (C) CD3 + CD8 + CD44 lo CD62L hi T cells are summarized in bar diagrams. (A-C) The data displayed in bar diagrams represent mean ± SEM of seven to nine mice per group analyzed in two independent experiments by flow cytometry. Statistical significances were tested using a non-parametric two-tailed Mann-Whitney U-test ( ࢩ p ࣘ 0.05; ࢩࢩ p ࣘ 0.01; ࢩࢩࢩ p ࣘ 0.001).

FRC-derived IL-7 does not affect size and TCR diversity of the peripheral T cell pool
Prx1-Cre-transgenic (Prx1-Cre + ) mice express Cre in BM stromal cells [31], which are crucial for IL-7-dependent B cell development [32]. Whether this mouse model is suitable for targeting FRCs in pLNs was analyzed next. For this purpose, Prx1-Cre + mice were crossed to ROSA26 reporter mice expressing red fluorescent protein (RFP) upon Cre-mediated activation of the reporter construct [33]. Peripheral LNs of Prx1-Cre + ROSA26 RFP mice were analyzed by flow cytometry to determine the degree of cell type-specific recombination. Among CD45 − stromal cells, around 80% of FRCs expressed RFP while LECs, BECs, and DNs showed only negligible levels of recombination ( Fig. 3A and B). Of note, Cre activity was barely detectable in CD45 + immune cells (Fig. 3A) as well as splenic LSCs (data not shown). Hence, Prx1-Cre + mice allow gene targeting in pLN FRCs.
In order to inactivate Il7 gene activity in pLN FRCs, Prx1-Cre + mice were crossed to IL-7 fl/fl mice. As compared to FRC wt littermate controls, Il7 mRNA levels were reduced by approximately 83% in pLNs of FRC IL-7 mice (Fig. 3C) confirming that FRCs are the major source of IL-7 in pLNs. In contrast, Il7 mRNA levels in the spleen of FRC IL-7 mice remained unaltered (Fig. 3C), probably due to the different developmental origins of splenic and LN FRCs [34,35]. Importantly, FRC-specific Il7 inactivation did not affect frequencies of LSC subsets (Fig. 3D), overall morphology, and chemokine secretion in pLNs (Supporting Information Fig. 2A-D).
When CD4 + and CD8 + T cells were quantified in pLNs and spleens of FRC IL-7 and FRC wt mice, no significant differences C 2020 The Authors. European Journal of Immunology published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.
www.eji-journal.eu Figure 2. LEC-derived IL-7 is dispensable for peripheral T cell homeostasis. (A) Based on their differential expression of gp38 and CD31, live TER-119 − CD45 − pLN LSCs can be subdivided into gp38 + CD31 − FRCs, gp38 + CD31 + LECs, gp38 − CD31 + BECs, and gp38 − CD31 − DNs. Shown is a representative contour plot from LEC IL-7 mice; numbers indicate percentages. The indicated LSC subsets were purified by flow cytometry from LNs of LEC wt (Lyve1-Cre + IL-7 −/wt and Lyve1-Cre − IL-7 wt/wt ) and LEC IL-7 (Lyve1-Cre + IL-7 −/fl ) mice. Three independent sorts with pooled pLNs from three to four mice per group (in total nine to ten mice per group) were performed. Once, cells from two sorts were pooled. Relative Il7 mRNA amounts were determined by RT-qPCR in relation to Hprt. Data displayed in the bar diagram are representative of two data points per group analyzed in two independent RT-qPCR experiments and show mean ± SEM of triplicates. Statistical significances were tested using a nonparametric two-tailed Mann-Whitney U-test ( ࢩ p ࣘ 0.05). (B) Absolute numbers of CD3 + CD4 + and CD3 + CD8 + T cells were determined in pLNs and spleen (Sp). (C, D, F, and G) Frequencies and (E and H) absolute numbers of naive (T N ; CD44 lo CD62L hi ), effector memory (T EM ; CD44 hi CD62L lo ) and central memory (T CM ; CD44 hi CD62L hi ) T cells were determined after gating on (C-E) CD3 + CD4 + and (F-H) CD3 + CD8 + cells isolated from pLNs or spleen. (C and F) Shown are representative contour plots and numbers indicate percentages. (B-H) Data was collected using flow cytometry. (B-H) The data shown in bar diagrams represent mean ± SEM combined from 11-12 LEC wt (Lyve1-Cre + IL-7 −/wt ) and LEC IL-7 (Lyve1-Cre + IL-7 −/fl ) mice per group analyzed in six independent experiments. Statistical significances were tested using a non-parametric two-tailed Mann-Whitney U-test ( ࢩ p ࣘ 0.05).
were detected in either case ( Fig. 3E and G). Furthermore, TCR Vβ repertoires of CD4 + and CD8 + T cells were indistinguishable between FRC IL-7 and FRC wt mice ( Fig. 3F and H). Hence, the size and diversity of the peripheral T cell pool is independent of pLN FRC-derived IL-7.
Next, we assessed whether LEC-derived IL-7 compensates for the lack of FRC-derived IL-7. For this purpose, FRC IL-7 mice were crossed to LEC IL-7 mice to generate double Cre-transgenic FRC/LEC IL-7 mice lacking Il7 gene expression in both, FRCs and LECs. Similar to LEC IL-7 (

FRC-derived IL-7 determines T CM abundance
Bcl-2 is a direct target of IL-7 [1] and is expressed at particularly high levels by CD8 + T M [36]. In order to test whether Bcl-2 expression is altered in the absence of FRC-derived IL-7, CD8 + T N and T M derived from pLNs and spleens were analyzed. As shown in Fig. 4A, frequencies and numbers of CD44 hi Bcl-2 hi CD8 + T M were significantly reduced in pLNs, but not spleens, of FRC IL-7 mice. IL-7 conditions CD8 + T cells for the IL-15-induced upregulation of Eomesodermin (Eomes) [37], a transcription factor promoting CD8 + T M differentiation [38]. As shown in Fig. 4B, CD44 hi Eomes hi CD8 + T M were strongly reduced in pLNs of FRC IL-7 mice. Again, these differences between FRC wt and FRC IL-7 mice were most evident in pLNs. However, there was a tendency of reduced CD8 + T M frequencies and cell numbers in spleens of FRC IL-7 mice ( Fig. 4A and B).
To analyze this IL-7-dependent T M defect in more detail, CD44 and CD62L expression was analyzed on CD8 + T cells from pLNs and spleens of FRC wt and FRC IL-7 mice. Frequencies and numbers of CD8 + CD44 lo CD62L hi T N and CD44 hi CD62L lo T EM were indistinguishable in pLNs and spleens of FRC wt and FRC IL-7 mice (Fig. 4C-E). In apparent contrast, frequencies of CD8 + CD44 hi CD62L hi T CM were significantly reduced in pLNs and spleens of FRC IL-7 mice (Fig. 4D). With regard to absolute CD8 + T CM numbers, this difference between both mouse strains was limited to pLNs (Fig. 4E). CD4 + T N and CD4 + T EM frequencies and numbers were unaltered in pLNs and spleens of FRC wt and FRC IL-7 mice (Fig. 4F-H). Similar to CD8 + T CM (Fig. 4D), frequencies of CD4 + T CM were reduced in FRC IL-7 pLNs but were only slightly affected in spleens (Fig. 4G). Absolute cell numbers were not significantly different in pLNs and spleens of both mouse strains (Fig. 4H). Hence, Il7 gene inactivation in FRCs is associated with a reduction of CD8 + T CM , an effect that was by far less pronounced for CD4 + T CM .
The survival of both, T N and T CM, critically relies on IL-7 [1,4] suggesting that either incomplete Il7 gene inactivation or the presence of non-pLN-derived IL-7 created IL-7 levels in FRC IL-7 pLNs that were sufficient for T N survival but too low for T CM maintenance. However, this assumption would predict different efficacies of IL-7 utilization by T N and T CM . Consistent with this idea and recent data [39], IL-7 treatment induced a more efficient IL-7 receptor α (IL-7Rα; CD127) down-regulation by CD8 + T N compared to T CM (Supporting Information Fig. 3A). IL-7R signaling leads to the phosphorylation of STAT5 that in turn regulates genes controlling T cell survival [40,41]. Interestingly, more pronounced IL-7Rα down-modulation by CD8 + T N (Supporting Information Fig. 3A) correlated with more efficient STAT5 phosphorylation (Supporting Information Fig. 3B). This argues for a more effective utilization of IL-7 by CD8 + T N and provides an explanation for their survival in FRC IL-7 mice. Conversely, CD8 + T CM appear to require higher levels of FRC-derived IL-7 for survival.
Unimmunized adult mice contain virtual memory CD8 + T cells (CD8 + vT M ), which are generated independently of foreign antigen contact as a result of lymphopenia-induced proliferation (LIP) in the neonatal phase [42][43][44][45][46][47]. As we have shown previously, IL-7 promotes CD8 + vT M formation [45]. Whether and how FRCderived IL-7 also affects the formation/maintenance of foreign antigen-specific CD8 + T CM was tested next. For this purpose, FRC wt and FRC IL-7 mice were reconstituted with TCR-transgenic CD8 + OT-I T cells specific for the OVA-derived peptide SIINFEKL. In order to mimic a viral infection, recipient mice were immunized with a mixture of PolyI:C and SIINFEKL 24 h later. PolyI:C induces pro-inflammatory cytokines such as IFN-α/β and IFN-γ, which promote IL-7 upregulation [24,48] and the subsequent formation and maintenance of T M [8,49]. Thirty days after vaccination the numbers of splenic CD8 + OT-I T M were comparable between FRC wt and FRC IL-7 mice (Fig. 4I). However, similar to the experiments shown above, T CM frequencies were clearly reduced in FRC IL-7 mice whereas T EM appeared to be less dependent on FRC-derived IL-7 ( Fig. 4J and K). This finding indicates that FRC-derived IL-7 helps to maintain both, virtual as well as foreign antigen-specific CD8 + T CM .

Discussion
In steady state, IL-7 is supposed to be produced at constant levels [50], mainly by radio-resistant stromal cells [1,51]. T N and T M express high levels of the IL-7R enabling them to remove IL-7 from the system continuously [50]. As soon as the peripheral T cell pool reaches a critical size, IL-7 production and consumption reach the equilibrium and the survival of additional T cells is prevented. Hence, the maintenance of T cell homeostasis relies on the competition for limiting amounts of IL-7 [14,50,52]  www.eji-journal.eu revealed that FRC-derived IL-7 promotes T N survival [9] suggesting that circulating T N receive IL-7-dependent survival signals in pLNs [14][15][16][17]. Besides its impact on T N homeostasis, IL-7 also promotes the formation and maintenance of other pLN-homing immune cells including CD8 + T CM [1,8,40] and RORγt + type 3 innate lymphoid cells (ILC3) [53,54]. Therefore, a common pool of FRC-derived IL-7 is supposed to regulate homeostasis of multiple immune cells in pLNs [17].
There is accumulating evidence that CD8 + T N and T M pools are regulated independently [55][56][57] indicating that they colonize different ecological niches [55][56][57]. In the immune system, ecological niches are defined by the combination of resources affecting the survival and function of a particular immune cell population [57]. In order to limit competition and enable the simultaneous survival of multiple immune cell types, ecological niches must be segregated. However, niche segregation of CD8 + T N and T CM appears to be incomplete as suggested by their common IL-7 dependence [40,56]. Nevertheless, we do not know yet if niche segregation involves the IL-7-dependent spatial separation of both cell types. The uneven distribution of IL-7-producing FRCs [29,58] suggests that, similar to chemokines [59,60], areas of high and low IL-7 density exist in pLNs. Based on their differential IL-7 demands, this assumption would predict the accumulation of CD8 + T N and T CM in separate pLN regions. Of note, the degree of local CD8 + T N and T CM segregation in pLNs varies strongly between experimental systems [61,62]. Whether this context-dependent effect correlates with the presence or absence of IL-7-producing FRCs in particular regions is still unclear since, at least to our knowledge, reliable reagents for IL-7 protein detection in pLNs are still missing.
Although we cannot fully exclude that different anatomical locations modulate distinct aspects of IL-7-dependent CD8 + T N and T CM homeostasis, our results indicate that variable IL-7 sensitivities of CD8 + T N and T CM contribute to the segregation of their ecological niches. In agreement with a recent study [39], we confirmed that IL-7R signaling is less efficient in CD8 + T CM . In a situation of limited IL-7 availability, this property would provide an explanation for the reduction of virtual as well as foreign antigen-specific CD8 + T CM in FRC IL-7 mice. Furthermore, our results are in line with the current paradigm of IL-7-dependent T cell homeostasis proposing that the optimized utilization of limiting IL-7 amounts is prerequisite for the survival of the greatest possible number of IL-7-dependent immune cells [63]. Based on this model, the degree of competition between different IL-7consuming cells would be restricted and the limited space within pLNs would be used most optimally [14,50].
The insensitivity of T N to FRC-specific Il7 inactivation may be due to the fact that T N are anyway capable of surviving short phases of IL-7 deficiency [63]. Indeed, IL-7 binding induces the down-modulation of IL-7R expression by T N more rapidly than by T CM rendering them insensitive to further IL-7 signals [63]. This effect is transient and appears to fulfill at least two functions. First, the amount of IL-7 consumed by a single T N is restricted thereby optimizing IL-7 availability for other immune cells [63]. Second, permanent IL-7R signaling would cause chronic T cell activation and subsequent activation-induced cell death [64]. Keeping in mind that (i) multiple organs produce IL-7 [24,25] and (ii) T N continuously circulate through the body, they may tolerate the partial IL-7-deficiency in FRC IL-7 pLNs because they received critical IL-7 signals elsewhere. Alternatively, incomplete Il7 gene inactivation in FRC IL-7 pLNs may allow the production of residual IL-7, which is just sufficient to promote local T N survival. In any case, our data demonstrate that T CM and T N do not tolerate the reduction of IL-7 in FRC IL-7 pLNs equally well. As shown for polyclonal CD8 + T cells in the steady state and for CD8 + OT-I T cells after vaccination, T CM prove to be particularly sensitive to IL-7 ablation in FRCs. However, CD8 + T CM are only partially reduced in FRC IL-7 mice. Whether this is due to the survival of CD8 + T CM subsets with reduced IL-7 demands remains to be shown in the future.
When we compared FRC IL-7 and FRC wt pLNs, we did not observe any obvious differences (Supporting Information Fig. 2A-C). T and B lymphocyte distribution, stromal cell localization, and relative distances between FRCs and lymphocytes appeared normal in FRC IL-7 mice (Supporting Information Fig. 2A-C). Furthermore, chemokine expression was comparable between FRC IL-7 and FRC wt pLNs (Supporting Information Fig. 2D) and ILC3 contributing to the IL-7-dependent regulation of T cell homeostasis [51,54,65] were similarly abundant (Supporting Information Fig. 2E). Hence, our findings argue for normal LN development and function in the absence of FRC-derived IL-7. This strongly suggests that the reduction of T CM in FRC IL-7 results from a lack of IL-7-dependent homing/survival signals rather than structural and/or functional alterations of FRC IL-7 pLNs.
In summary, we provide evidence that IL-7 produced by pLN FRCs regulates T cell homeostasis. As opposed to the current model, our data demonstrate that pLN FRC-derived IL-7 is of limited importance for the local and systemic survival of T N . On the contrary, the maintenance of T CM critically relies on steady state levels of FRC-derived IL-7 suggesting that T N and T CM colonize different ecological niches in vivo.

Cell isolation
To obtain single cell suspensions from pLNs and spleens, organs were forced through metal strainers in PBS/2 mM EDTA (Carl Roth) and erythrocytes were lysed. For erythrocyte lysis, spleen cells were re-suspended in ammonium-chloride-potassium lysis buffer for 90 s followed by the addition of RPMI 1640 (Biochrom) containing 10% (v/v) FCS (PAN Biotech) and 1% (v/v) penicillin/streptomycin (P/S; Gibco). After centrifugation, spleen cells were re-suspended in PBS/2 mM EDTA and filtered through 40 µm cell strainers (Corning, Durham, NC).

Flow cytometry cell sorting of LSCs
LSCs were isolated from peripheral and mesenteric lymph nodes as described above. LN cells were incubated with purified anti-CD16/32 ( TER-119 − CD45 − LSC subsets were sorted based on their differential gp38/CD31 expression. Purities of the indicated LSC subsets were >73.3 % (data not shown).

Flow cytometry
The following reagents were purchased from BioLegend: anti-

Reverse transcriptase PCR (RT-PCR) and real-time quantitative PCR (RT-qPCR)
Colon samples were transferred into CK14 2 mL tubes (Peqlab/VWR) containing 700 µL TRIzol reagent (Invitrogen) and homogenized in a Precellys 24 homogenizer (Peqlab/VWR). Peripheral LNs were transferred into CK14 0.5 mL tubes (Peqlab/VWR) containing 200 µL TRIzol reagent and homogenized. Sorted LSCs were re-suspended in 500 µL TRIzol reagent. For RNA extraction, chloroform (Sigma-Aldrich) was added and total RNA was isolated according to the manufacturer's instructions. Isolated RNA was quantified by photometric Nanodrop (Thermo Fisher Scientific) measurement. RNA was reverse-transcribed using random hexamer primers and the advantage RT-for-PCR kit (Takara Clontech) according to the manufacturer's instructions.
For RT-PCR analyses of colon samples, the Taqman R Gene Expression Master Mix (Thermo Fisher Scientific) and the following TaqMan R Gene Expression Assays (Thermo Fisher Scientific) were used according to the manufacturer's instructions: Il7 (FAM-MGB probe Mm01295804 m1) and Hprt (FAM-MGB probe Mm00446968 m1). PCR products were analyzed by agarose gel electrophoresis.

Automated multidimensional fluorescence microscopy by multi-epitope-ligand cartography
Multi-epitope-ligand cartography (MELC) was performed as described previously [70]. Briefly, pLNs were embedded into Tissue-Tek R O.C.T. TM compound (Sakura Finetek), frozen on dry ice, and stored at -80°C. Ten micrometer cryo-sections adhered to silane-coated cover slides (Thermo Fisher Scientific) were fixed with PBS/2% (w/v) paraformaldehyde (Sigma-Aldrich), permeabilized with PBS/0.2% (v/v) Triton-X-100 (Carl Roth) and blocked with PBS/1 % (w/v) BSA (Sigma-Aldrich) + 30% (v/v) normal goat serum (Invitrogen). Tissue samples were transferred to an inverted wide-field fluorescence microscope (Leica DMi8, 20× air lens NA 0.80; Leica Microsystems). The automated cyclic robotic process started with the incubation of the first fluorochrome-labeled antibody (tag). After a series of washing steps, the fluorescence signals and a corresponding phase contrast image were acquired by a cooled charge-coupled device camera (Apogee KX4; Apogee Instruments). The specific signal of the given tag was removed by bleaching the fluorescent dye followed by recording of post-bleaching fluorescence signals and repetition of incubation-imaging-bleaching-cycle. The appropriate working dilutions, incubation times, and positions within the MELC experiment of the used tags (anti-mouse CD3 (17A2), CD31 (390), CD8a (53-6.7), gp38 (8.1.1), CD45 (30-F11), CD54 (YN1/1.7.4), CD44 (IM7), and CD45R/B220 (RA3-6B2) were purchased from BioLegend, anti-mouse CD4 (RM4-5) from BD Biosciences, PI from Sigma-Aldrich) were validated systematically using conditions suitable to MELC [70]. The series of fluorescence images produced by each tag were aligned pixel-wise using the corresponding phase contrast images. The automated algorithm reaches an alignment accuracy of 0.1 pixels. Illumination faults of the images were corrected using flat-field correction. Post-bleaching images were subtracted from the following fluorescence tag images. Section artifacts were excluded as invalid by a manual mask-setting process. We developed pipelines for the Cell Profiler software package [71] in order to detect (i) all cells within the tissue section using the staining of PI, CD45, CD44, and CD54, and (ii) to create masks for gp38 and CD31 positive signals. Using these masks of gp38 and CD31 the FRC region was defined as gp38 + and CD31 − . For each cell, the mean fluorescent intensity and the smallest distance to the reference region FRC was calculated. The resulting matrix of intensities and distances were exported into an FCS file and uploaded to the online cytometry analysis platform "cytobank.org" for multiparametric analysis.