The astrocyte marker ALDH1L1 also identifies a stromal cell population in the lymph node

TL;DR

ALDH1L1, known as an astrocyte marker in the brain, also identifies a unique stromal cell population in lymph nodes.

Contribution

ALDH1L1 marks a distinct subset of fibroblastic reticular cells in lymph nodes, not previously recognized.

Findings

ALDH1L1+ cells in lymph nodes do not coexpress CNS or immune cell markers.

ALDH1L1+ cells show significant colocalization with podoplanin, a fibroblastic reticular cell marker.

ALDH1L1+PDPN+ cells are enriched in the paracortex and medulla of lymph nodes.

Abstract

ALDH1L1 is widely used as a marker of astrocytes in the central nervous system (CNS), but its expression and potential roles in the periphery, particularly in lymphoid organs, remain poorly understood. Here, we found that ALDH1L1+ cells comprise approximately 5–9% of the lymph node (LN). To better understand their identity, we investigated whether these cells share characteristics with glial, immune, endothelial or stromal cell populations under homeostatic conditions. Using ALDH1L1/TdTomato reporter mice, immunofluorescence, and flow cytometry, we found that TdTomato+ cells in LNs do not coexpress canonical CNS (GFAP, ACSA-2) or peripheral glial (Sox10) markers. Similarly, using multiple approaches we found minimal overlap with T cells (CD3, TCRβ), B cells and dendritic cells (B220), myeloid cells (CD11b, Iba1), or antigen presenting cells (MHCII). To explore potential stromal associations, we assessed coexpression of TdTomato with endothelial (CD31, LYVE1) and reticular (ER-TR7) markers and found insubstantial overlap. In contrast, TdTomato+ cells showed significant colocalization with podoplanin (PDPN), a marker of fibroblastic reticular cells (FRCs). Interestingly, while many TdTomato+ cells were PDPN+, relatively few PDPN+ cells expressed TdTomato. Notably, ALDH1L1+PDPN+ cells were enriched in the paracortex and medulla but absent from the B cell follicles and subcapsular sinus. These findings suggest that while ALDH1L1 remains a reliable marker of astrocytes in the CNS, it may also delineate a distinct subset of FRCs in the LN, opening new avenues for exploring stromal cell diversity and function in lymphoid tissues.

The online version contains supplementary material available at 10.1038/s41598-026-38426-6.

Figures

ALDH1L1^+^ cells do not colocalize with GFAP or Sox10 in the LN. *Aldh1l1*^Cre/ERT2+^::TdTomato mice were injected with tamoxifen to induce recombination and TdTomato (ALDH1L1) expression, perfused, and tissues were removed and prepared for immunohistochemistry. (A) LNs and (B) sciatic nerve (SN) were labeled for GFAP and Sox10 and nuclei were counterstained with DAPI. Scale bar = 100 μm. (C) The brain was labeled for GFAP and nuclei counterstained with DAPI. Scale bar = 1 mm. (D) TdTomato^+^, (E) GFAP^+^, and (F) Sox10^+^ area were quantified and normalized to the area of the respective tissu

ACSA-2^+^ cells are not detectable in the LN but a high proportion of ACSA-2^+^ cells in the CNS are ALDH1L1^+^. *Aldh1l1*^Cre/ERT2+^::TdTomato mice were injected with tamoxifen to induce recombination and TdTomato (ALDH1L1) expression, perfused, and tissues were removed and prepared for flow cytometry. (A) Cells were identified and gated for (B) singlets, (C) live cells, and ACSA-2 expression in the (D) LN and (E) brain and (F) absolute cell count was quantified. (G) TdTomato positivity was then assessed in ACSA-2^+^ cells and (H) cell count was quantified. Each data point is representative o

ALDH1L1 expression in LN immune cell populations. Following the gating performed in Fig. [2](#Fig2)A-C, cells were gated for (A) TCRβ, (B) B220, and (C) CD11b followed by determining TdTomato positivity (D-F). *Aldh1l1*^Cre/ERT2−^::TdTomato mice were injected with tamoxifen and used as FMO negative controls for TdTomato expression. (G) The proportion and (H) absolute number of cells that were TdTomato^+^ and TdTomato^−^ for each immune cell subset was quantified. Each data point is representative of individual animals and error bars represent mean ± SEM. p* < 0.05, *p* < 0.0001 as determined b

ALDH1L1^+^ cells do not colocalize with several immune cell subsets in the LN. *Aldh1l1*^Cre/ERT2+^::TdTomato mice were injected with tamoxifen to induce recombination and TdTomato (ALDH1L1) expression, perfused, and tissues were removed and prepared for immunohistochemistry. (A) LNs were labeled for CD3 and Iba1. Scale bar = 100 μm. (B) Higher magnification image of the T cell zone in panel A. Scale bar = 10 μm. (C) TdTomato^+^, CD3^+^, and Iba1^+^ area were quantified and normalized to LN area. (D) The Manders’ coefficient was quantified for colocalization of TdTomato with CD3 and Iba1. (E)

A). Notably, *Pdpn* expression was augmented in similar populations (Supplemental Fig. 5B). Further examination of spatially distinct mesenteric LNs described by Gu et al.^[24](#CR24)^ revealed coexpression of *Aldh1l1* with key transcripts in LN stromal cell subtypes identified in Supplemental Fig. 5A including *Nr4a1* and *Ccl19*. We also noted coexpression of *Aldh1l1* with *Tnfsf11*, a marker Rodda et al.^[10](#CR10)^ used to distinguish marginal reticular cells (Supplemental Fig. 5C). Additionally, we isolated TdTomato⁺ cells from LNs of TdTomato::*Aldh1l1*-Cre^ERT2+^ mice using fluoresce

Summary of LN markers evaluated. (A) Schematic representation of the LN, brain, and peripheral nervous system with indicated cell types evaluated and respective TdTomato positivity. Created in BioRender. Williams, J. (2026) [https://BioRender.com/oxw3ubp](https://BioRender.com/oxw3ubp). (B) Quantification of the area occupied by indicated markers, normalized to total LN area, compiled from Figs. [1](#Fig1), [2](#Fig2), [3](#Fig3), [4](#Fig4) and [5](#Fig5). (C) Calculated Manders’ coefficient of TdTomato with indicated markers, compiled from Figs. [1](#Fig1), [2](#Fig2), [3](#Fig3), [4](#Fig4)

Full text

Introduction

Astrocytes are among the most abundant cells of the central nervous system (CNS) and represent a heterogeneous population of specialized glia that carry out diverse and fundamental functions during CNS homeostasis and in disease-related processes^1^. The identification of astrocytes has traditionally relied on molecular markers including glial fibrillary acidic protein (GFAP), S100β, glutamate transporter 1, and more recently, aldehyde dehydrogenase 1 family member L1 (ALDH1L1), which has emerged as a pan-astrocyte marker with high specificity and sensitivity in the CNS^2–8^. ALDH1L1 is a folate enzyme that has important roles in cell division and growth^9^ and has been observed in cells of the lung, liver, spleen, and intestine^2^; however, the potential roles of ALDH1L1^+^ cells outside of the CNS remain in question. Given that Aldh1l1 transcript has been detected in several populations of stromal cells^10^, a cell type shared by nearly all tissues and organs, particularly secondary lymphoid organs, we wondered if this could account for the ambiguity of ALDH1L1^+^ cells described in the literature to date.

Stromal cells in secondary lymphoid organs, such as the lymph nodes (LNs), spleen, and Peyer’s patches, have essential roles in organizing tissue architecture, creating specialized microenvironments and compartmentalizing immune cells, functions that notably overlap with those of astrocytes. Non-hematopoietic in origin, stromal cells can be subdivided into several subsets based on location, function, and phenotype^11^. In the LN, various immune compartments are primarily separated by fibroblastic reticular cells (FRCs), which are highly spatially heterogeneous, having up to ten subtypes^12^. FRCs are commonly identified by expression of podoplanin (PDPN), a transmembrane mucin-type glycoprotein^12,13^. Subsets of LN stromal cells, including FRCs, marginal reticular cells, and medullary fibroblasts, are associated with ER-TR7 staining, which detects collagen type VI in the extracellular matrix and highlights reticular fiber networks within the LN^11,14^. Although it is known that FRC subsets are mesenchymal in origin, the mechanisms governing their differentiation into these distinct lineages are largely unknown^12^.

In addition to immune and stromal cells, LNs are also innervated by sensory neurons that facilitate bi-directional crosstalk between the immune and nervous systems^15^, suggesting the presence of supporting glia. Peripheral glia in other organs, including the intestine, have been described as expressing common astrocyte markers including GFAP and S100β^16^ and exhibit intracellular calcium responses^17,18^. However, despite the similarities to astrocytes, enteric glia do not seem to express appreciable levels of ALDH1L1^19^, suggesting that a level of heterogeneity exists between glia from distinct organ systems.

Although glia have been described in several peripheral tissues, those residing in LNs remain largely unexplored. Since Aldh1l1 is now commonly used to drive or delete gene expression in CNS astrocytes, particularly in models of neuroinflammation, it is important to understand which populations may be affected in secondary lymphoid tissues. To better understand its expression profile in peripheral tissues, we generated an inducible TdTomato reporter mouse driven by Aldh1l1-Cre/ERT2. We evaluated TdTomato overlap by comparing canonical astrocyte and satellite cell markers in the central and peripheral nervous systems (PNS) using the brain and sciatic nerve (SN), respectively, with those in the LN. For the immune cell compartment, we used classical markers for T and B lymphocytes, myeloid cells, antigen presenting cells, as well as endothelial and stromal cells. Analyses were performed using immunofluorescent imaging and flow cytometry. As expected, there was significant overlap between TdTomato and GFAP in the brain and SN. However, while nearly all cell populations assessed in the LN had minimal or no overlap with TdTomato, there was a significant population of TdTomato and PDPN double positive cells.

Results

ALDH1L1+ LN cells do not express canonical astrocyte or peripheral glia markers under homeostatic conditions

To characterize populations of cells expressing ALDH1L1 in peripheral tissues, we injected TdTomato::Aldh1l1-Cre^ERT2+^ mice with tamoxifen to induce TdTomato in ALDH1L1-expressing cells. Validation of tamoxifen-induced recombination was assessed by examining LNs from untreated TdTomato::Aldh1l1-Cre^ERT2+^ mice, as well as TdTomato::Aldh1l1-Cre^ERT2−^ mice treated with or without tamoxifen. We observed low reporter activation in these tissues, suggesting minimal spontaneous recombination (Supplemental Fig. 1A-F). Further, comparable levels of TdTomato were detected across LNs collected from different anatomical regions, indicating uniform recombination efficiency (Supplemental Fig. 1G-I). Finally, we confirmed reporter specificity by labeling for ALDH1L1 protein and observing complete overlap with TdTomato (Supplemental Fig. 1J).

To determine if ALDH1L1 was coexpressed with markers associated with CNS and PNS glial lineages, we visualized TdTomato expression and performed immunofluorescent labeling of GFAP, a well-established astrocyte marker, and Sox10, a transcription factor broadly expressed in PNS glial cells, including Schwann cells and satellite glia^20^. Quantifying expression in the LN, SN, and brain, we found that per tissue area, TdTomato, GFAP, and Sox10 were highest in the SN compared to other tissues (Fig. 1A-F) with nearly undetectable levels of GFAP or Sox10 in the LN (Fig. 1E-F). As expected, TdTomato was strongly colocalized with GFAP and Sox10 in CNS and PNS tissues, respectively, but was not significantly coexpressed in the LN (Fig. 1G-H). To confirm these findings, we performed flow cytometry to assess the expression of another classical astrocyte marker, astrocyte cell surface antigen (ACSA)-2, in TdTomato^+^ cells, which constituted 6.97% ± 1.87 of all LN cells digesting with Dispase II (Supplemental Fig. 2A,** B**). Gating on live, singlets (Fig. 2A-C), we determined ASCA-2 expression in LN and in the CNS as a positive control (Fig. 2D-E). Although a significant population of ASCA-2^+^ cells in the CNS expressed TdTomato, ACSA-2 expression was absent in the LN (Fig. 2F-H) and spleen (Supplemental Fig. 3). The lack of colocalization with GFAP, ACSA-2, and Sox10 in the LN suggests that during physiological conditions, glia-associated ALDH1L1 expression is largely restricted to nervous system tissues.

Fig. 1ALDH1L1^+^ cells do not colocalize with GFAP or Sox10 in the LN. Aldh1l1^Cre/ERT2+^::TdTomato mice were injected with tamoxifen to induce recombination and TdTomato (ALDH1L1) expression, perfused, and tissues were removed and prepared for immunohistochemistry. (A) LNs and (B) sciatic nerve (SN) were labeled for GFAP and Sox10 and nuclei were counterstained with DAPI. Scale bar = 100 μm. (C) The brain was labeled for GFAP and nuclei counterstained with DAPI. Scale bar = 1 mm. (D) TdTomato^+^, (E) GFAP^+^, and (F) Sox10^+^ area were quantified and normalized to the area of the respective tissue. The Manders’ coefficient was quantified for colocalization of TdTomato with (G) GFAP and (H) Sox10. Images were collected by confocal microscopy and analyzed using ImageJ software. Each data point is representative of individual animals and error bars represent mean ± SEM. **p < 0.01, ***p < 0.001, ****p < 0.0001 as determined by one-way ANOVA (D, E, G) or t test (F, H).

Fig. 2ACSA-2^+^ cells are not detectable in the LN but a high proportion of ACSA-2^+^ cells in the CNS are ALDH1L1^+^. Aldh1l1^Cre/ERT2+^::TdTomato mice were injected with tamoxifen to induce recombination and TdTomato (ALDH1L1) expression, perfused, and tissues were removed and prepared for flow cytometry. (A) Cells were identified and gated for (B) singlets, (C) live cells, and ACSA-2 expression in the (D) LN and (E) brain and (F) absolute cell count was quantified. (G) TdTomato positivity was then assessed in ACSA-2^+^ cells and (H) cell count was quantified. Each data point is representative of individual animals and error bars represent mean ± SEM. ***p < 0.001, ****p < 0.0001 as determined by t test.

Characterization of ALDH1L1 and immune cell marker coexpression

To investigate whether ALDH1L1 is expressed in immune cell populations, we performed flow cytometric analysis on dissociated LN tissues. Following the same gating strategy outlined in Fig. 2A-C, T cells were identified by T cell receptor (TCR)β, B cells and plasmacytoid dendritic cells by B220, and the myeloid compartment by CD11b (Fig. 3A-C). From each of those populations, we generated a histogram to quantify TdTomato expression (Fig. 3D-F), which revealed that the vast majority of TCRβ^+^, B220^+^, and CD11b^+^ cells were ALDH1L1^−^ (Fig. 3G-H). Similar results were observed in the spleen with the exception of CD11b^+^ cells having a greater proportion of TdTomato^+^ cells compared to the LN (Supplemental Fig. 4). To confirm these findings using additional cell type-specific labels, we performed immunofluorescent staining of LN tissue with CD3 and Iba1, markers for T cells and myeloid cells, respectively, and imaged using confocal microscopy. While TdTomato expression per tissue area was relatively low compared to that of CD3 and Iba1, it appeared to occupy the LN medulla, T cell zone, and stroma without visible or quantifiable overlap with T cells or myeloid cells (Fig. 4A-D). Although TdTomato expression was noticeably absent from B cell zones, we verified that both B220 and MHCII, an indicator of all antigen presenting cells, lacked coexpression with ALDH1L1 (Fig. 4E-H). Overall, combining imaging and flow cytometry data, we did not observe convincing overlap or substantial numbers of TdTomato^+^ cells coexpressing canonical immune cell markers including CD3 (T cells), B220 (B cells), CD11b and Iba1 (myeloid cells), or MHCII (antigen presenting cells).

Fig. 3ALDH1L1 expression in LN immune cell populations. Following the gating performed in Fig. 2A-C, cells were gated for (A) TCRβ, (B) B220, and (C) CD11b followed by determining TdTomato positivity (D-F). Aldh1l1^Cre/ERT2−^::TdTomato mice were injected with tamoxifen and used as FMO negative controls for TdTomato expression. (G) The proportion and (H) absolute number of cells that were TdTomato^+^ and TdTomato^−^ for each immune cell subset was quantified. Each data point is representative of individual animals and error bars represent mean ± SEM. *p < 0.05, ****p < 0.0001 as determined by one-way ANOVA.

Fig. 4ALDH1L1^+^ cells do not colocalize with several immune cell subsets in the LN. Aldh1l1^Cre/ERT2+^::TdTomato mice were injected with tamoxifen to induce recombination and TdTomato (ALDH1L1) expression, perfused, and tissues were removed and prepared for immunohistochemistry. (A) LNs were labeled for CD3 and Iba1. Scale bar = 100 μm. (B) Higher magnification image of the T cell zone in panel A. Scale bar = 10 μm. (C) TdTomato^+^, CD3^+^, and Iba1^+^ area were quantified and normalized to LN area. (D) The Manders’ coefficient was quantified for colocalization of TdTomato with CD3 and Iba1. (E) The LN was labeled for B220 and MHCII. Scale bar = 100 μm. (F) Higher magnification image of the border of the T cell zone in panel E. Scale bar = 10 μm. (G) TdTomato^+^, B220^+^, and MHCII^+^ area were quantified and normalized to LN area. (H) The Manders’ coefficient was quantified for colocalization of TdTomato with B220 and MHCII. Images were collected by confocal microscopy and analyzed using ImageJ software. Each data point is representative of individual animals and error bars represent mean ± SEM.

FRC stromal cell marker overlaps with ALDH1L1+ cells in the LN

Stromal cells can be broadly classified as having endothelial or reticular characteristics with those of endothelial likeness typically lining blood or lymphatic vessels and FRCs forming the network of collagen fibers in secondary lymphoid structures. While these cells can exhibit local diversity, they can generally be distinguished by expression of CD31, LYVE1, and PDPN^21,22^. Further, a subset of FRCs can be recognized using ER-TR7^23^, which can associate with other classes of structural stromal cells and the extracellular matrix within the LN^11^. To determine if TdTomato^+^ cells had an expression profile similar to that of endothelial cells, we labeled LN tissue for CD31 and LYVE1 and found that TdTomato^+^ cells had minimal, if any, significant overlap with either marker (Fig. 5A-D). We next stained the LN using ER-TR7 and again found that while ER-TR7 labels antigens throughout the LN, there was not an appreciable proportion that localized with TdTomato^+^ cells (Fig. 5E-F, I-J). In contrast, when we labeled LN tissue for PDPN and determined the Manders’ coefficient for the proportion of TdTomato^+^ cells that were PDPN^+^, we observed considerable colocalization (Fig. 5G-J), underscoring a significant association between these markers. Notably, quantification of the proportion of PDPN^+^ cells that were also TdTomato^+^ revealed a relatively minimal colocalization coefficient (Fig. 5J) indicating that ALDH1L1 may be restricted to a specific subset of PDPN^+^ stromal cells.

Fig. 5ALDH1L1^+^ cells coexpress PDPN. Aldh1l1^Cre/ERT2+^::TdTomato mice were injected with tamoxifen to induce recombination and TdTomato (ALDH1L1) expression, perfused, and tissues were removed and prepared for immunohistochemistry. (A) LNs were labeled for CD31 and LYVE1. Scale bar = 100 μm. (B) Higher magnification image of the T cell zone in panel A. Scale bar = 10 μm. (C) TdTomato^+^, CD31^+^, and LYVE1^+^ areas were quantified and normalized to LN area. (D) The Manders’ coefficient was quantified for colocalization of TdTomato with CD31 and LYVE1. (E, F) The LN was labeled with ER-TR7 and (G, H) PDPN. Scale bar = 100 μm. (F,** H**) Higher magnification images of the T cell zone in panels E and G, respectively. Scale bar = 10 μm. (I) TdTomato^+^, ER-TR7^+^, and PDPN^+^ area were quantified and normalized to LN area. (J) The Manders’ coefficient was quantified for colocalization of TdTomato with ER-TR7^+^ structures and PDPN. Colocalization of PDPN with TdTomato positive area was also quantified. Images were collected by confocal microscopy and analyzed using ImageJ software. Each data point is representative of individual animals and error bars represent mean ± SEM.

To complement these imaging findings and further define the transcriptional profile of LN-specific ALDH1L1-expressing cells, we examined available stromal cell transcriptomic datasets. Analysis of LN stromal cell subtypes described by Rodda et al.^10^ revealed enriched Aldh1l1 expression in Nr4a1^+^ stromal cells and Ccl19^hi^ and Ccl19^lo^ T-zone reticular cells (Supplemental Fig. 5A). Notably, Pdpn expression was augmented in similar populations (Supplemental Fig. 5B). Further examination of spatially distinct mesenteric LNs described by Gu et al.^24^ revealed coexpression of Aldh1l1 with key transcripts in LN stromal cell subtypes identified in Supplemental Fig. 5A including Nr4a1 and Ccl19. We also noted coexpression of Aldh1l1 with Tnfsf11, a marker Rodda et al.^10^ used to distinguish marginal reticular cells (Supplemental Fig. 5C). Additionally, we isolated TdTomato⁺ cells from LNs of TdTomato::Aldh1l1-Cre^ERT2+^ mice using fluorescence-activated cell sorting (FACS) and analyzed the expression of stromal cell markers by qPCR. Indeed, the stromal cell transcripts Ccl19 and Ccl21 were expressed in sorted cells along with Aldh1l1 (Supplemental Fig. 5D-I). Of note, this expression profile remained in a further analysis of mesenchymal cells within the mesenteric lymph node in which cells expressed Aldh1l1 along with Pdpn, Ccl19, and Ccl21, but were Ptprc (CD45) negative^24^ (Supplemental Fig. 6). Together, this suggests that TdTomato^+^ LN cells may be consistent with stromal cells of a mesenchymal lineage.

ALDH1L1+PDPN+ cells primarily occupy the paracortex and medulla of the LN

Lymph nodes are highly organized structures that facilitate immune interactions. Within them, the stromal compartment is comprised of a diverse set of specialized cells that maintain the architecture of immune cell niches and support immune system function by producing growth factors and secreting chemokines that guide immune cell organization^25^. Spatially, we ruled out the presence of TdTomato^+^ cells in the B cell follicles by immunofluorescent imaging and by assessing coexpression of both B220 and MHCII by flow cytometry (Figs. 3 and 4). We also did not find TdTomato-expressing cells within the subcapsular sinus. In contrast, CD31^+^, LYVE1^+^, and PDPN^+^ cells were dense in this region (Fig. 5). Rather, a vast majority of TdTomato^+^ cells resided in the paracortex and medulla, primarily occupying spaces enriched with T cells without having substantial overlap (Figs. 4A-B and 6A). As expected, immune and stromal cell markers comprised a large majority of the LN tissue area while neither GFAP nor Sox10 labeled the naïve LN by immunofluorescent imaging despite the presence of the canonical astrocyte marker ALDH1L1 (Fig. 6B). Compiling the Manders’ coefficients to determine the degree of colocalization, the coexpression of ALDH1L1 with PDPN was drastically increased compared to all other cell markers analyzed (Fig. 6C), suggesting a cellular association between these two molecules.

Fig. 6. Summary of LN markers evaluated. (A) Schematic representation of the LN, brain, and peripheral nervous system with indicated cell types evaluated and respective TdTomato positivity. Created in BioRender. Williams, J. (2026) https://BioRender.com/oxw3ubp. (B) Quantification of the area occupied by indicated markers, normalized to total LN area, compiled from Figs. 1, 2, 3, 4 and 5. (C) Calculated Manders’ coefficient of TdTomato with indicated markers, compiled from Figs. 1, 2, 3, 4 and 5. Each data point is representative of individual animals and error bars represent mean ± SEM. ****p < 0.0001 as determined by one-way ANOVA.

Discussion

The Aldh1l1-Cre/ERT2 transgenic mouse strain is widely used to study astrocytes due to its high specificity and sensitivity in the CNS^2–8^. However, some studies have reported expression in various peripheral tissues^2,26,27^. Since Aldh1l1 is also expressed in certain stromal cell populations^10^, cells common to all tissues, we wondered if this could account for the observed peripheral expression. Despite reports of ALDH1L1 expression in various peripheral tissues, its presence in the LN has not yet been clearly defined. To address this gap, here we report ALDH1L1 expression across LN cells, CNS astrocytes, and peripheral satellite glia during homeostatic conditions using a TdTomato reporter. Given that peripheral glia are found in many lymphoid tissues, we first evaluated the coexpression of TdTomato with glia-associated markers and found minimal overlap. Additionally, using flow cytometric analysis of the LN, we did not observe TdTomato expression in ACSA-2^+^ cells, suggesting that the ALDH1L1-expressing cells in the lymph node do not exhibit conventional astrocyte-associated markers. Next, we focused on the immune cell compartment and found that an overwhelming majority of TCRβ^+^ T cells, B220^+^ B cells and dendritic cells, MHCII^+^ antigen presenting cells, Iba1^+^ and CD11b^+^ myeloid cells were TdTomato^−^ by both histological and flow cytometric quantification. Finally, we assessed the expression of endothelial and stromal cell marker expression and found that ALDH1L1^+^ cells had high levels of coexpression with PDPN. Although a significant proportion of ALDH1L1^+^ cells were also PDPN^+^, not all PDPN^+^ cells expressed ALDH1L1, suggesting that Aldh1l1-Cre/ERT2 may drive recombination in a specific subset of LN stromal cells.

Although ALDH1L1 is primarily thought of as an astrocyte marker, it also functions as a key enzyme in a variety of biological processes, suggesting that its expression in the lymph node may have functional significance. Specifically, ALDH1L1 has a critical role in folate metabolism by converting 10-formyl-tetrahydrofolate to tetrahydrofolate (THF) and carbon dioxide in an NADP^+^-dependent reaction that generates NADPH, a key antioxidant, as a by-product^9,28,29^. FRCs in the LN are exposed to periods of oxidative stress in several circumstances, including aging. Age-associated oxidative stress in FRCs reduces the expression of T cell survival factors, such as interleukin (IL)-7, resulting in impaired maintenance of naïve T cell populations^30^. Moreover, aged LN stromal cells exhibit defects in proliferative expansion during infection, compromising immune responses. Strikingly, pharmacological treatment with antioxidants or compounds that reduce mitochondrial reactive oxygen species restored the capacity of aged FRCs to support naïve T cell survival, highlighting the importance of LN stromal cells in redox balance^30^.

In addition, FRCs respond to inflammatory cues that drive their expansion and metabolic reprogramming. For example, IL-17 produced in inflammatory settings such as experimental autoimmune encephalomyelitis and colitis stimulates LN stromal cells, particularly FRCs, driving their proliferation. This expansion is required for maintaining the structural scaffold that supports inflating lymphoid populations^25,31,32^. Consequently, FRCs require increased energy to maintain this proliferation^32^. The ALDH1L1-mediated breakdown of folate regenerates THF, which feeds into one-carbon pathways such as methionine and purine synthesis, processes essential for supporting cellular energy metabolism and ATP generation^33,34^. Thus, beyond maintaining redox homeostasis, ALDH1L1 expression in FRCs may also help sustain the energy demands required to support effective proliferation and immune responses in the LN although further analyses are required to fully appreciate the function of ALDH1L1^+^ FRCs.

Within the LN, stromal cell subsets are highly specialized and extraordinarily complex. PDPN expression is a hallmark of FRCs, a spatially defined fibroblastic cell type with multiple subsets. Although FRCs are known to tile the LN and other secondary lymphoid tissues, their precise functions and defining markers have yet to be fully elucidated. FRCs are typically located in the paracortex but absent from B cell zones in secondary lymphoid organs^35^. Additionally, they are known to express PDPN, but lack CD31 expression^36^. We observed that TdTomato^+^ cells are largely absent from LN B cell follicles, but are prevalent in T cell zones, and lack CD31 and ER-TR7 labeling, which are typically expressed by T zone reticular cells^12^. This may suggest that Aldh1l1^+^ LN cells exhibit functional similarities with a subset of FRCs, although further examination using additional stromal cell markers is necessary to define this subset.

In contrast to most connective tissue fibroblasts, which are typically embedded in the extracellular matrix, LN FRCs ensheathe strands of ECM to maintain close contact with immune cells^37^. FRCs are core to LN structure and organization in part due to their high expression of small leucine-rich proteoglycans including lumican, fibromodulin, osteoglycin, decorin, biglycan, and prolargin^38^. In addition, expression of adhesion molecules, such as vascular cell adhesion molecule-1, intercellular adhesion molecule-1, and CD44 by FRCs facilitates strong immune cell attachment to the stromal network^39^. Because FRCs are tightly associated with several cell types through multiple adhesion mechanisms, their isolation for analysis requires specialized dissociation protocols that combine mechanical separation with DNase I, Collagenase P (or D), Dispase II, and even depletion of CD45^+^ hematopoietic cells for more complete and representative recovery^10,40^. Since our objective was to determine the cellular source of ALDH1L1 within the LN, depletion of specific cell types would introduce substantial confounds. However, although we optimized the dissociation protocol prior to flow cytometric analysis, residual stromal cell attachment cannot be fully excluded and may account for some TdTomato overlap with immune cells. This is particularly true for the spleen where FRCs are more numerous and complex due to the presence of the red pulp^41^. Indeed, the Manders’ coefficient quantification from imaging analyses and the distinct TdTomato^+^ cell morphology in the LN would suggest that TdTomato^+^ cells are likely separate from classical immune cells. Although flow cytometric analysis of ALDH1L1^+^ stromal cells would be informative, dissociation of LN tissue with Dispase II limited PDPN expression in TdTomato^+^ cells, albeit without a demonstrable change in TdTomato expression (Supplemental Fig. 7). This is thus a limitation of the current study.

Based on the markers used here, we hypothesize that ALDH1L1^+^ cells in the LN are part of a stromal lineage and, more specifically, most closely align with FRCs. This is largely attributed to their expression of PDPN, spatial distribution, cellular morphology, and the absence of markers canonically associated with astrocytes, satellite cells, immune cells, and endothelial cells. However, further analysis is needed to fully support this conclusion. Notably, while PDPN can be expressed by astrocytes^7,42–44^, we did not observe substantial overlap between conventional astrocyte markers and TdTomato^+^ cells in the LN. Further, astrocytes originate from the neural ectoderm^45^ while LN stromal cells are made up of both mesenchymal (FRCs) and endothelial cells^12,36^, suggesting two distinct cell types. Overall, while the Aldh1l1-Cre/ERT2 strain is well-suited for studying astrocytes in the CNS, it may also serve as a tool to identify and study distinct subsets of stromal cells in secondary lymphoid tissues. Further analysis of this cell population via proteomic, transcriptomic, and functional analysis will be important for investigating the potentially significant roles that these cells hold in LN physiology.

Methods

Animals

B6.Cg-Gt(ROSA)26Sor^tm14(CAG−tdTomato)Hze^/J (TdTomato) (strain no. 007914) and Aldh1l1-Cre/ERT2 (strain no. 031008) mice were obtained from The Jackson Laboratory. Mice were bred as heterozygous pairs to generate Aldh1l1^Cre/ERT2^::TdTomato mice and both sexes were used for all experiments. Tamoxifen (75 mg/kg) (Sigma) was dissolved in corn oil (Sigma) and injected into 12–16-week-old mice i.p. for 5 consecutive days to induce recombination. All mice used were maintained on a 14/10 h light/dark cycle and had ad libitum access to food and water. Mice (housed 2–5/cage) did not have any prior history of drug administration, surgery, or behavioral testing. All housing, breeding, and experimental procedures were approved by the Institutional Animal Care and Use Committee at the Cleveland Clinic Foundation (Cleveland, OH) using protocol number 1862, and adhered to ARRIVE guidelines. This study was conducted in accordance with relevant guidelines and procedures.

Immunofluorescent labeling

Aldh1l1^Cre/ERT2^::TdTomato mice were intracardially perfused with PBS followed by 4% paraformaldehyde (PFA) and tissues were removed and fixed in 4% PFA at 4 °C for 24 h. Brachial and inguinal LN, sciatic nerve, and brain tissue was then cryopreserved in 30% sucrose and frozen in O.C.T. compound (Fisher HealthCare). Frozen sections (10 μm) were slide-mounted and stored at -80 °C. Tissue sections were blocked with 10% goat serum and 0.1% Triton X-100 (Southern Biotech) for 1 h at room temperature and then incubated with anti-GFAP (1:200; Invitrogen; 13–0300), -Sox10 (1:1000; Abcam; ab155279), -CD3 (1:250; Invitrogen; 14-0031-82), -Iba1 (1:250; Wako Chemicals; 019-19741), -B220 (1:400; BD Pharmingen; 557390), -MHCII (1:100; Abcam; ab7856), -CD31 (1:100; Millapore Sigma; MAB1398Z), -LYVE1 (1:800; Novus Biologicals; NB600-1008), -ER-TR7 (1:200; Novus Biologicals; NB100-64932), -PDPN (1:400; Invitrogen; 14-5381-82), and ALDH1L1 (1:100; Cell Signaling; 85828) primary antibodies overnight at 4 °C. Appropriate secondary antibodies conjugated to AlexaFluor 405, 488, 555, or 647 (1:400; ThermoFisher Scientific) were applied for 1 h at room temperature. Nuclei were counterstained with DAPI (ThermoFisher Scientific). Sections were imaged using the 10x objective of a light microscope BX-X710 (Keyence) and the 63x objective of a confocal microscope LSM 800 (Carl Zeiss). 10x images were stitched using ImageJ v1.54p. Stitched images are representative of 9–15 tissue sections per individual mouse. 63x images shown are representative of 3–4 images taken from at least two sections per individual mouse. The mean positive area and Manders’ coefficient of colocalization were determined by setting thresholds using appropriate controls and quantified using ImageJ v1.54p.

Flow cytometry

Following cardiac perfusion with PBS, combined brachial and inguinal LN, spleen, and CNS tissues were isolated from Aldh1l1^Cre/ERT2^::TdTomato mice. Tissues were digested in DMEM containing 10 mM HEPES (Gibco), 1% Penicillin/Streptomycin (Life Technologies Corp), 1.4 U/mL Collagenase VIII (Sigma) and 100 U/mL DNase I (Millipore Sigma), and lymphoid tissue was additionally digested in 2 U/mL Dispase II (Sigma) for 40 min shaking at 37 °C, triturating after the first 20 min. Tissue was pushed through 70 µM strainer to generate a single cell suspension. CNS tissue was separated using a 30% Percoll (Cytiva) gradient to remove myelin debris. Cells were then counted, stained for live cells with Zombie NIR (1:1000; BioLegend; 423105) and blocked with anti-mouse CD16/32 (1:50; BioLegend; 101302) before staining with fluorescently conjugated antibodies specific for ACSA-2 (1:100; Miltenyi; 130-117-386), TCRβ (1:100; BioLegend; 109241), B220 (1:100; Fisher Scientific; BDB553087), CD11b (1:100; BD Pharmingen; 550993), and PDPN (1:100; BioLegend; 127410). Stained cells were fixed using Fixation Buffer (BioLegend) for 15 min at room temperature. Data was collected using a 6-laser, 21-color CytoFLEX LX (Beckman Coulter) or sorted using a BD FACSAria Illu and analyzed using FlowJo software v5.00000. All gates were based on marker-specific FMO negative controls.

Published data set analyses

Single-cell RNA sequencing data from inguinal, brachial, and axillary LN stromal cells described by Rodda et al. were analyzed using the software provided in the published manuscript (http://scorpio.ucsf.edu/LNSC/)^10^. BioTuring Lens software was utilized for bioinformatics analysis of target genes using the single-nucleus RNA sequencing dataset of mesenchymal lymph nodes from Gu et al.^24^. The UMAP analysis function was used to identify and plot target genes.

Statistics

Data were analyzed using a two-tailed Student’s t test or one-way analysis of variance (ANOVA) with correction for multiple comparisons where appropriate. All statistical analyses were performed using GraphPad Prism software version 10.1.2 (GraphPad). A p-value of less than 0.05 was considered statistically significant. Data points in graphs represent individuals.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1