Aged regulatory T cells fail to control autoimmune lacrimal gland pathogenic CD4⁺ T cells

Abstract

CD25KO mice are a model of Sjögren disease. CD25KO mice have severe inflammation and infiltrating lymphocytes to the lacrimal glands (LG). Whether the pathogenicity of CD25KO CD4⁺ T cells can be controlled in vivo by Tregs is unknown. Eight-week-old B6 and CD25KO mice LGs were submitted for RNA bulk sequencing. A total of 3481 genes were differentially expressed in CD25KO LG compared to B6. Tear washing analysis identified CD25KO mice had elevated protein levels of TNF, IFN-γ, and CCL5 and decreased protein levels of IL-12p40 and VEGF-A. Co-adoptive transfer of CD25KO CD4⁺ T cells with either young or aged B6 Tregs was performed in RAG1KO mice. Recipients of CD25KO CD4⁺ T cells alone had higher LG inflammation than naive mice. However, in recipients of young B6 Tregs plus CD25KO CD4⁺ T cells, LGs had significantly reduced inflammation. Recipients of CD25KO CD4⁺ T cells with aged B6 Tregs had more inflamed LGs than young Tregs, suggesting aged Tregs have less suppressive capacity in vivo. Altogether, CD25KO mice have phenotypic and genetic changes resulting in increased inflammation and severe lymphocytic infiltration in the LGs. However, this autoimmunity can be controlled by the addition of young, but not aged, Tregs, suggesting that aging Tregs have dysfunctional suppression.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11357-025-01576-y.

Introduction

Sjögren disease (SjD) is an autoimmune disorder characterized by immune infiltration in the exocrine glands, including the tear-secreting lacrimal gland and the salivary glands. This results in dry eye and dry mouth and is detrimental to the quality of life of patients [1, 2]. There is a significant sex difference in SjD; women are diagnosed with SjD more frequently than men at an average ratio of 9:1 [3]. Despite this, some reports indicate that men have a more severe phenotype of the disease [4, 5]. In a subset of patients, SjD progresses to systemic autoimmunity, including an increased risk of lymphoma [6, 7]. While SjD can occur at any age, the average age of diagnosis is 53 years, and prevalence increases with age [2, 8]. The disease is initially driven by autoreactive CD4⁺ T cells, resulting in destruction of exocrine gland tissues [9].

CD25KO mice are a model of systemic autoimmunity [10–12]. CD25 is the IL-2 receptor α-chain and is essential for IL-2 signaling [13–15]. Due to a lack of IL-2 signaling, CD25KO mice have significantly reduced numbers of Foxp3⁺ Tregs that lack suppressor function and die off due to lack of homeostatic maintenance [16–20]. CD25KO mice are therefore characterized as an autoimmune model without Tregs.

CD25KO mice exhibit many phenotypic signs of SjD. Previous work in our lab has reported that CD25KO mice have increased CD4⁺ and CD8⁺ T cells in the conjunctiva and lacrimal glands, an increase in Th1 and Th17 cytokine mRNA in the lacrimal glands, cornea and conjunctiva, and increased apoptosis in the lacrimal glands [12, 21]. Systemically, CD25KO mice develop ulcerative colitis, hemolytic anemia, enlarged lymphoid organs due to clonal expansion of autoreactive T and B cells, and have increased mortality [16, 22].

In healthy tissue, T regulatory cells (Tregs) regulate autoimmunity by suppressing activated T cells. However, the suppressive ability of Tregs in SjD is deeply contested in the literature [23–28]. Several studies have reported that SjD patients have increased Tregs in the blood [23–26], despite an increase in autoimmune signs and symptoms. Likewise, it has been reported that infiltrating Foxp3⁺ Tregs in the exocrine glands positively correlate with severity of focus score in patients, suggesting the Tregs are not capable of suppressing the autoimmunity [7, 24]. One possible explanation for the incongruence between the presence of Tregs and the exacerbated autoimmunity in SjD patients is that Tregs in aging individuals have less suppressive capacity compared to Tregs from younger individuals.

Similar to the increased presence of Tregs in SjD patients, Tregs accumulate with aging in many tissues [29–33]. The suppressive capability of aged Tregs is likewise controversial. Previously, we have described that aged Tregs have impaired suppressor function in in vitro assays [32, 33]. Similarly, other groups have reported that aged Tregs are less suppressive in vivo in other disease models [34, 35], and one study found aged human Tregs were less suppressive in vitro [36]. On the other hand, some groups report equal or greater suppressive capability in aged Tregs compared to young [31, 37, 38]. Therefore, the suppressive capability of aged Tregs needs to be further elucidated.

The purpose of this work was to characterize the immunological profile of the CD25KO mouse in relation to the lacrimal gland involvement and investigate the capacity of differently aged Tregs to control autoreactivity of their CD4⁺ T cells.

Methods

Animals

The Institutional Animal Care and Use Committee at Baylor College of Medicine approved all animal experiments (AN-6491). In addition, all studies adhered to the Association for Research in Vision and Ophthalmology for the Use of Animals in Ophthalmic and Vision Research and the NIH Guide for the Care and Use of Laboratory Animals [39]. The experiments were performed at the Ocular Surface Center, Department of Ophthalmology, Baylor College of Medicine, Houston, Texas.

Heterozygous breeder pairs of Il2ra^+/− mice on a C57BL/6 background (IL-2ra or CD25, B6.129S4-Il2ra^tm1Dw/J) and breeder pairs of Rag1^+/− mice (Recombination activating gene 1; B6.129S7-Rag1^tm1Mom/J) were purchased from Jackson Laboratories (Bar Harbor, Maine) for establishing breeder colonies. The genotype of CD25KO mice was confirmed according to the Jackson Laboratories’ protocol using a commercial vendor (Transnetyx, Cordova, Tennessee). Young B6 donor mice were purchased from Jackson Laboratories (stock 000664, Bar Harbor, Maine), and aged B6 mice were aged in-house. Mice were housed in specific pathogen-free facilities of Baylor College of Medicine and were kept on diurnal cycles of 12 h/light and 12 h/dark with ad libitum access to food and water and environmental enrichment. Thirty-eight 8-week-old female CD25KO mice, 3 8-week-old male CD25KO mice, 39 female B6 mice under 12 weeks of age, 10 male B6 mice under 12 weeks of age, 24 female B6 mice 24–28 months of age, 87 female RAG1KO mice age 8 weeks to 5 months, and 42 male RAG1KO mice aged 8 weeks to 5 months were used. Efforts were made to use the same mouse in multiple endpoints to decrease the sample size. The final sample size per endpoint can be found in figure legends.

Bulk RNA sequencing

One extraorbital lacrimal gland from 8-week-old CD25KO female mice (n = 6) and 8-week-old B6 female mice (n = 4) was excised, minced, and snap-frozen in a tube containing RNA lysis buffer (Qiagen, Valencia, California). Total RNA was extracted using a QIAGEN RNeasy Plus Mini RNA isolation kit according to the manufacturer’s instructions. The concentration was assessed using a NanoDrop 2000 (Thermo Fisher, Waltham, Massachusetts).

A double-stranded DNA library was created at Baylor College of Medicine’s Genomics and RNA Profiling Core. Briefly, using 100 ng of total RNA (measured by pico green), an oligodT primer containing an Illumina-compatible sequence at its 5′ end was hybridized to the RNA, and reverse transcription was performed using a Lexogen kit. Second-strand synthesis was initiated by a random primer containing an Illumina-compatible linker sequence at its 5′ end. The purified double-stranded library was then amplified and purified. The resulting libraries were quantified using Qubit 2.0 (Thermo Fisher), and fragment size was assessed with the Agilent Bioanalyzer (Agilent Technologies, Santa Clara, California). A qPCR quantification was performed on the libraries to determine the concentration of adapter-ligated fragments using Applied Biosystems ViiA7 TM Real-Time PCR System and a KAPA Library Quant Kit (Waltham, Massachusetts). All samples were pooled equimolarly, re-quantified by qPCR, and reassessed on the bioanalyzer. Using the concentration from the ViiA7 TM qPCR machine above, 1.8 pM of the equimolarly pooled library was loaded onto NextSeq 500 high-output flow cell (Illumina, San Diego, California). PhiX Control v3 adapter-ligated library (Illumina) was spiked in at 1% by weight to ensure balanced diversity and to monitor clustering and sequencing performance. A single-read 75 base pair cycle run was used to sequence the flow cell. An average of 34 million reads per sample was sequenced. The FastQ file generation was executed using Illumina’s cloud-based informatics platform, BaseSpace Sequencing Hub (Illumina).

Data was analyzed by ROSALIND® (https://rosalind.bio/), with a HyperScale architecture developed by ROSALIND, Inc. (San Diego, CA). Reads were trimmed using cutadapt [40]. Quality scores were assessed using FastQC [41]. Reads were aligned to the Mus musculus genome build GRCm38 using STAR [42]. Individual sample reads were quantified using HTseq [43] and normalized via Relative Log Expression (RLE) using DESeq2 R library [44]. Read Distribution percentages, violin plots, identity heatmaps, and sample MDS plots were generated as part of the QC step using RSeQC [45]. DEseq2 was also used to calculate fold changes and p-values and perform optional covariate correction. Clustering of genes for the final heatmap of differentially expressed genes was done using the PAM (Partitioning Around Medoids) method using the fpc R library [46]. Hypergeometric distribution was used to analyze the enrichment of pathways, gene ontology, domain structure, and other ontologies. The topGO R library [47] was used to determine local similarities and dependencies between GO terms in order to perform Elim pruning correction. Several database sources were referenced for enrichment analysis, including Interpro [48], NCBI [49], MSigDB [50, 51], REACTOME [52], and WikiPathways [53]. Enrichment was calculated relative to a set of background genes relevant for the experiment.

From the list of differentially expressed genes (DEGs) identified by RNAseq, we used the Metascape database [54] (http://metascape.org/) to identify the altered enriched ontology clusters and biological pathways. To achieve this, we submitted the lists of DEGs and performed genome pathway enrichment analysis and Gene Ontology annotation [55–57].

Measurement of tear volume

Aqueous tear production was assessed using a cotton thread (Quick Thread; FCI Ophthalmics, Marshfield Hills, MA) as previously described [58, 59].

Luminex assay of tears

Tear washings were collected from 2–3-month-old female B6 and female CD25KO mice. Multiple ages were used due to availability since tears had to be pooled from multiple eyes per sample. Briefly, 1.5 µl of phosphate-buffered saline containing 0.1% bovine serum albumin was instilled into the conjunctival sac. The tear fluid and buffer were collected with a 1 µl volume glass capillary tube (Drummond Scientific Co, Broomall, Pennsylvania) by capillary action from the tear meniscus in the lateral canthus and stored at − 80 °C until the assay was performed. One sample consisted of tear washings from both eyes of two mice pooled (4 µl) in phosphate-buffered saline with 0.1% bovine serum albumin (6 µl). Samples were added to wells containing the appropriate cytokine bead mixture that included mouse monoclonal antibodies specific for IL-1β, IL-2, TNF, IL-12p40, interferon-γ, chemokine CXC motif ligand 1 (CXCL1), VEGF-A, chemokine CC motif ligand 5 (CCL5/RANTES), and IL-17A (Millipore, Burlington, Massachusetts), as previously reported [60]. The reactions were detected with streptavidin–phycoerythrin using a Luminex 100 IS 2.3 system (Austin, Texas). Results are presented as mean ± SD (pg/ml). The inferior levels of detection were 0.46 pg/ml (IL-1β), 2.75 pg/ml (TNF), 0.005 pg/ml (IFN-γ), 1.3 pg/ml (IL-2), 1.0 pg/ml (IL-12p40), 1.5 pg/ml (VEGF), 1.92 pg/ml (CXCL1), 3.15 pg/ml (CCL5), and 1.21 pg/ml (IL-17A).

CD4⁺T cell isolation and time course adoptive transfer

Single-cell suspensions of CD25KO lymphocytes were prepared from spleens and draining lymph nodes, and CD4⁺ T cells were isolated by negative selection using the CD4⁺ T Cell Isolation Kit (Miltenyi Biotec, Bergisch Gladbach, Germany) following the manufacturer’s instructions. 2 × 10⁶ cells were injected intraperitoneally into sex-matched RAG1KO mice. These mice were euthanized at one-, two-, three, four-, and five-week timepoints after adoptive transfer, and their tissues were harvested and used for histology and qPCR.

Treg sorting and co-adoptive transfer

Single-cell suspensions of young or aged B6 lymphocytes were prepared from spleens and draining lymph nodes, and CD4⁺ T cells were isolated by magnetic beads as before. Cells were counted and then incubated with anti-CD16/32 (BioLegend, San Diego, California) to block Fc receptors for 10 min on ice. Subsequently, the suspensions were stained with the antibodies anti-CD4_FITC (Clone RM4-5, Invitrogen Thermo Fisher, Waltham, Massachusetts), anti-CD25_PE (clone 7D4, BD Biosciences, Franklin Lakes, New Jersey), and anti-CD357(GITR)_APC (clone DTA-1, BioLegend) for 30 min covered on ice. Samples were washed at least three times, resuspended in 4–6 ml of FACS buffer, and sorted on a BD FACSAria™ II Cell Sorter (BD Biosciences). The gating strategy used was as follows: two subsequent singlet gates excluded doublets, followed by dead cell exclusion by using DAPI. CD4⁺ T cells followed by double positive selection of CD25⁺GITR⁺ cells were identified as Tregs. Negative controls consisted of fluorescence minus one from the same single-cell suspensions.

Lacrimal gland inflammation score

Extraorbital lacrimal glands were excised, fixed in 10% formalin, paraffin-embedded, and cut into three levels (at 25, 50 and 75% depth in the tissue), using a microtome (Microm HM 340E, Thermo Fisher, Waltham, Massachusetts). Sections were stained with haematoxylin and eosin (H&E) for evaluating morphology and graded by two masked independent investigators using a modified score described by White and Casarett [61], changing the assigned score to the following rubric as done previously [62]: 0 = indicated no foci of mononuclear cells was observed; 1 = one to five foci (more than 20 cells per focus); 2 = more than five foci were visible but without tissue damage; 3 = indicated more than five foci with moderate tissue damage; 4 = extensive mononuclear infiltration with severe tissue destruction; 5 = extensive mononuclear infiltration with no normal acini visible. One section per level is scored for each gland. These scores are then averaged for a total score for the whole gland.

RNA isolation and quantitative PCR

Extraorbital lacrimal glands from CD25KO mice were excised, and total RNA was extracted using a QIAGEN RNeasy Plus Micro RNA isolation kit (Qiagen, Hilden, Germany) following the manufacturer’s protocol. The concentration of RNA was measured, and cDNA was synthesized using the Ready-To-GoTM You-Prime First-Strand kit (GE Healthcare, Chicago, Illinois). Quantitative real-time PCR was performed with specific minor groove binder probes for TNF (Tnf, Mm99999068), IL-1β (Il1b, Mm00434228), interferon-γ (Ifng, Mm00801778), IL-12A (Il12a, Mm00434165_m1), CD4 (Cd4, Mm00442754), major histocompatibility complex class II (Ciita, Mm00482914), Caspase 3 (Casp3, Mm01195085), Caspase 8 (Casp8, Mm00802247), Caspase 9 (Casp9, Mm00516563), interferon induced protein with tetratricopeptide repeats 1 (Ifit1, Mm00515153_m1), 2′−5′-oligoadenylate synthetase 2 (Oas2, Mm00460961_m1), 2′−5′ oligoadenylate synthetase 1A (Oas1a, Mm00836412_m1), interferon-induced protein 44 (Ifi44, Mm00505670_m1), and hypoxanthine–guanine phosphoribosyltransferase 1 (Hprt1, Mm00446968). The HPRT1 gene was used as an endogenous reference for each reaction. The results of real-time PCR were analyzed by the comparative CT method, and the results were normalized by the CT value of HPRT1.

Statistical analysis

Statistical analyses were performed with GraphPad Prism (GraphPad, Inc., version 10). Non-parametric Mann–Whitney U tests were used to compare B6 mice and CD25KO mice, adoptive transfer recipients of B6 CD4⁺ T cells and adoptive transfer recipients of CD25KO CD4⁺ T cells, and co-adoptive recipients of young or aged Tregs. Kruskal–Wallis test with Dunn’s multiple comparisons test was used to analyze all time course data. Two-way ANOVA were used to analyze the sex differences in the time course data. All experiments were repeated at least three times. The final sample size per endpoint can be found in the figure descriptions.

Results

CD25KO lacrimal glands have upregulated genes in SjD-related pathways

Since SjD disproportionately affects women, we used female mice for our experiments. For our controls in these experiments, we used C57BL/6 (B6) mice, because CD25KO mice have a B6 background. As previously evidenced [10, 12, 21, 63], we confirmed that CD25KO mice had lymphocytic infiltration in the lacrimal and salivary glands by H&E-stained histological sections (Supplemental Fig. 1). To gain a deeper knowledge of pathways that are differentially modulated, our first step was to investigate the gene expression differences in the lacrimal glands of CD25KO mice compared to wild-type B6 controls. Six CD25KO and four B6 extraorbital lacrimal glands were collected and lysed, and total RNA was submitted for bulk RNA sequencing. A total of 3481 genes were differentially expressed (≥ or ≤ 1.5-fold, p-adj < 0.05) in CD25KO mice compared to controls (Fig. 1A, B). Upregulated and downregulated genes were then analyzed to identify altered predicted pathways (Fig. 1C). Many of the upregulated predicted biological pathways involved immune cell pathways, T cell activation, differentiation, and lymphocyte recruitment. Several of these genes from these pathways are highlighted in Fig. 1D. We validated the RNA sequencing data by performing individual qPCR reactions on a selection of the differentially expressed genes (Ifng, Ifit1, Ifi44, Oas2, Oasa1a, and Casp3) and confirmed the bulk RNA sequencing findings (Supplemental Fig. 2). These gene expression results suggest that the pathology of the CD25KO lacrimal gland mimics that of SjD patients.

Fig. 1.

CD25KO lacrimal glands have marked transcriptomic changes and significantly upregulated genes involved in type I interferon signaling, T cell activation, and protein processing in the endoplasmic reticulum. A Heatmap of bulk RNA sequencing of female lacrimal glands between wild-type B6 controls (n = 4) and CD25KO mice (n = 6). B Volcano plot of all differentially expressed genes (⋚ 1.5 fold, p-adj < 0.05) between B6 controls and CD25KO mice. Purple DEG (1266 genes) are downregulated in CD25KO mice; green DEG (2215 genes) are upregulated in CD25KO mice. C Gene ontology pathway analysis of differentially expressed genes. D Heat maps of genes involved in the type I Interferon, t cell activation, and protein processing in endoplasmic reticulum pathways. B6 = B6 mice. KO = CD25KO mice. ER = endoplasmic reticulum. All mice were 8 weeks of age at time of lacrimal gland collection

CD25KO mice have differential expression of inflammatory cytokines in tears

As a decrease in tear volume and change in tear composition are hallmarks of dry eye, we investigated the tear film of CD25KO mice. CD25KO mice produced the same volume of tears as wild-type B6 mice (Fig. 2A). To determine if the protein composition of these tears was different, we collected tear washes from female 8-to-12 week-old mice and analyzed them using a Luminex assay panel for inflammatory cytokines and chemokines. Compared to B6 controls, CD25KO tears had significantly higher levels of the cytokines TNF and interferon-γ and the chemokine CCL5. CD25KO tears also had significantly lower protein amounts of VEGF-A compared to B6 mice. Although not significant, there was a trend towards lower levels of IL-2 in CD25KO tears. There were no differences between groups for cytokines IL-1β or IL-17 (Fig. 2B). These results suggest that the CD25KO mouse has an inflammatory tear profile reminiscent of SjD patients.

Fig. 2.

CD25KO tears have significant differences in inflammatory protein expression. A Tear volume in female mice as measured by phenol thread. One dot represents the average from two eyes of the same mouse. Mann–Whitney U test. B Luminex assay on female mice tears. Tear washings were collected from four eyes and pooled into one biological sample (n = 8–11 biological samples/group; each dot represents one biological sample or four eyes). Dotted line represents the limit of detection, VEGF-A’s limit of detection is Y = 1.5 pg/ml. Mann–Whitney U test. B6 = C57BL/6 mice. KO = CD25KO mice

Adoptive transfer of pathogenic CD4⁺ T cells into immunocompromised hosts is sufficient to cause lacrimal gland inflammation

Previously, we have demonstrated that adoptive transfer of CD4⁺ T cells isolated from CD25KO mice into sex-matched immunocompromised hosts was able to recreate the autoimmune dry eye phenotype [62], while transfer of B6 isolated CD4⁺ T cells did not [64]. To confirm these findings, we performed an adoptive transfer using CD4⁺ T cells isolated either from female CD25KO or B6 murine spleens and cervical lymph nodes into female RAG1KO mice. As before, mice were sacrificed at 5 weeks post-adoptive transfer, at which point disease is fully developed. We first evaluated lacrimal gland disease by scoring histological sections of the lacrimal gland with our inflammation scoring metric [62]. Recipients of CD4⁺ T cells from CD25KO mice had significantly higher inflammation scores in the lacrimal gland compared to recipients of CD4⁺ T cells from B6 mice, confirming previous results (Fig. 3A). This indicates greater disease in the lacrimal glands, including substantial infiltration of lymphomononuclear cells, acinar cell atrophy, and the occasional presence of fibrosis (Fig. 3B).

Fig. 3.

Adoptive transfer of pathogenic CD4⁺ T cells results in dacryoadenitis. A Inflammation score of female lacrimal glands from both recipients of B6 CD4⁺ T cells and CD25KO CD4⁺ T cells, assessed using a modified White and Casarett inflammation score [61] as we have done previously [62]. Mann–Whitney U test. Sample size = 6–29 per group. B Representative H&E-stained lacrimal gland sections from recipients that received B6 CD4⁺ T cells (AT-B6) and those that received CD25KO CD4⁺ T cells (AT-KO). C qPCR of broad inflammatory markers Il1b, Tnf, Ifng, T cell-related markers Cd4 and Ciita, and apoptosis-related markers Casp3, Casp8, and Casp9. Mann–Whitney U test. The dotted line represents the average fold expression of naive RAG1KO lacrimal glands. Each dot represents one lacrimal gland. Sample size = 4–8 per group. AT = adoptive transfer. B6 = B6 mice. KO = CD25KO mice

We next assessed inflammatory gene expression in the recipient lacrimal glands by qPCR (Fig. 3C). Recipients of autoreactive CD4⁺ T cells from CD25KO mice had significantly higher expression of the inflammatory markers Il1b, Tnf, and Ifng compared to recipients of B6 CD4⁺ T cells. Likewise, recipients of CD25KO CD4⁺ T cells had higher expression of T cell marker Cd4 and antigen-presenting cell marker Ciita. CD25KO CD4⁺ T cell recipients also had higher expression of Casp3, a marker from the apoptosis pathway, although neither Casp8 nor Casp9 were significantly different between groups.

Altogether, these results indicate that adoptive transfer of CD4⁺ T cells isolated from CD25KO mice results in higher rates of inflammation and more compromised tissue compared to adoptive transfer of CD4⁺ T cells from B6 mice.

Adoptively transferred pathogenic CD4⁺ T cells cause Sjögren-like disease in lacrimal glands at three weeks post-adoptive transfer

To understand the timing of this model, we performed an adoptive transfer of CD4⁺ T cells isolated from CD25KO spleen and cervical lymph nodes and evaluated lacrimal gland pathology once every week for 5 weeks. Three weeks after adoptive transfer, we identified infiltrating immune cell foci in the lacrimal gland, as well as smaller acinar cells (Fig. 4A, B). By 4 weeks, the infiltrating foci took up nearly 75% of most glands, and any remaining acinar cells were found in small clusters. In some glands, fibrosis began to appear. By the fifth week, all glands had high inflammation scores and little healthy tissue left.

Fig. 4.

Adoptive transfer time-course indicates time-dependent autoimmunity in histological sections of the lacrimal gland. A Inflammation score of female lacrimal glands from each time point. Kruskal–Wallis test with Dunn’s multiple comparisons test. B Representative H&E-stained lacrimal glands from each time point. The areas highlighted in yellow are healthy acinar cells. Yellow asterisk indicates infiltrating foci. F indicates fibrosis in the gland. AT = adoptive transfer, w = number of weeks post-adoptive transfer. Sample size = 3–9 lacrimal glands per time point. Scale bar = 100 µm

We then evaluated the expression of mRNA transcripts associated with inflammation in SjD in whole lacrimal gland lysates. Compared to naive RAG1KO mice, there was significantly increased fold expression of Tnf and Il12a by 3 weeks and significantly increased fold expression of Ifng and Ciita by 4 weeks after adoptive transfer (Fig. 5). While not significant by the Kruskal–Wallis test, there was marked increased fold expression of Ifng and Ciita by 3 weeks as well, suggesting that in our adoptive transfer model, 3 weeks is the start of visible inflammation in the lacrimal gland.

Fig. 5.

qPCR of whole lacrimal gland lysates from the adoptive transfer time course shows peak in inflammation between 3 and 4 weeks. qPCR of broad inflammatory marker Tnf, T cell-related markers Ifng, Ciita, and Il12a and apoptosis-related marker Casp3 in female whole lacrimal gland lysates. Kruskal–Wallis test with Dunn’s multiple comparisons test. The dotted line represents donor CD25KO lacrimal gland average fold expression. Naive RAG1KO mice served as calibrators for each reaction. AT = adoptive transfer. Time point = number of weeks after adoptive transfer lacrimal glands were collected for analysis. Sample size = 5 per time point. Each dot represents one lacrimal gland per mouse. P value < 0.05 considered significant

Taken together, these results indicate that autoimmune pathology from autoreactive CD4⁺ T cells starts to develop in lacrimal glands 3 weeks after adoptive transfer.

Autoimmunity can be controlled by functional T regulatory cells

Because the CD25KO mouse develops strong autoimmunity, we investigated if it was possible to contain the inflammation from these autoreactive T cells with the addition of Tregs through our adoptive transfer model. Using flow cytometry, Tregs were sorted from healthy, 8-to-12-week-old B6 spleens and cervical-draining-lymph nodes as CD4⁺CD25⁺GITR⁺ cells (Fig. 6). There was 97–98% purity in sorted Tregs (Supplemental Fig. 3). Recipients either received autoreactive CD4⁺ T cells alone or in a co-adoptive transfer with Tregs. Since the 5-week time point had the highest disease severity, we used that as our end point in our follow-up studies.

Fig. 6.

Co-adoptive transfer schematic. Controls consist of RAG1KO recipients receiving only CD4⁺ T cells from CD25KO mice. Young B6 mice are 8–12 weeks at euthanasia. Aged B6 mice are 24–28 months old at euthanasia. Made with BioRender.com

As before, recipients of only autoreactive CD25KO T cells had severe inflammation, few healthy acinar cells, and fibrosis (Fig. 7A, B). Mice that received a co-adoptive transfer of both autoreactive T cells and Tregs had significantly reduced inflammation scores (Fig. 7A, B) and minimal T cell infiltration.

Fig. 7.

Tregs have varying suppressive capabilities based on age. A Inflammation score of female lacrimal glands depending on the group. KO = CD25KO CD4⁺ T cells. Y = Tregs from 8–12 week mice. A = Tregs from 20-month mice. Kruskal–Wallis test with Dunn’s multiple comparisons test. Sample size = 7–29 lacrimal glands per group. B Representative H&E-stained lacrimal gland sections from recipients depending on group. The areas highlighted in yellow are healthy acinar cells. Yellow asterisk indicates infiltrating foci. AT = adoptive transfer. KO = CD25KO CD4⁺ T cells. YTregs = young Tregs. ATregs = aged Tregs. Scale bar = 100 µm

Taken together, these findings indicate that Tregs isolated from young B6 mice can modulate autoreactive T cells in an in vivo Sjögren-like disease model.

Aged Tregs have less suppressive capability compared to young Tregs

Our next question was if Tregs from aged B6 mice had the same suppressive capabilities as Tregs from young B6 mice. To answer this, we performed adoptive transfer experiments with a group that received a co-adoptive transfer of autoreactive T cells and Tregs sorted from 24 to 28 month-old B6 mice (Fig. 6). CD25⁺GITR⁺ Tregs from young and aged groups had similar median fluorescence intensity of Foxp3 (Supplemental Fig. 3), suggesting these populations are equivalent. As before, we waited 5 weeks after the adoptive transfer for all endpoints.

As done previously, we first evaluated the histology of these lacrimal glands. While mice that received a co-adoptive transfer of aged Tregs and autoreactive T cells had reduced inflammation scores on average, there was high variability amongst the cohort. Aged Treg recipients had significantly higher inflammation scores than young Treg recipients (Fig. 7A). However, the aged Treg cohort also had significantly lower inflammation scores compared to the autoreactive T cells alone group (Fig. 7A). When evaluating the histological sections, aged Treg recipient lacrimal glands had more diffuse infiltration of immune cells and more acinar cell death compared to recipients of young Tregs (Fig. 7B). However, the recipients of autoreactive T cells alone generally had much less healthy acinar tissue compared to the recipients of aged Tregs. This suggests that Tregs taken from aged mice have decreased suppressive ability but retain some function.

We then investigated mRNA expression of inflammatory cytokines in whole lacrimal gland lysates to compare young Treg recipients and aged Treg recipients (Fig. 8). The aged Tregs cohort had significantly higher mRNA expression of the inflammatory markers Ifng and Tnf, and the cellular markers Cd4 and Ciita compared to the young Tregs cohort. This suggests that the aged Tregs were incapable of suppressing inflammation as effectively as the young Tregs.

Fig. 8.

Recipients of Tregs from either young (8–12 weeks) or aged (24–28 months) mice determine suppressive capability. qPCR of broad inflammatory markers Il1b, Tnf, Ifng, and T cell-related markers Cd4 and Ciita in female whole lacrimal gland lysates. Mann–Whitney U test. The dotted line represents the average fold expression of the adoptive transfer of CD25KO CD4⁺ T cells into RAG1KO murine lacrimal glands alone. Sample size = 8 lacrimal glands per group. AT = adoptive transfer. KO = CD25KO CD4⁺ T cells. Y = young Tregs. A = aged Tregs

Taken together, these results imply that Tregs from aged mice have less suppressive capability compared to Tregs from young mice, resulting in greater immune infiltration and higher rates of inflammation in the lacrimal glands.

Aged Tregs alone are not sufficient to cause disease in the lacrimal gland

Previously, we have demonstrated that Tregs from aged murine lacrimal glands have a Th1-effector phenotype [32, 33]. This led us to question if aged Tregs are sufficient to cause lacrimal gland disease in our adoptive transfer model. To investigate this, we performed an adoptive transfer using only sorted Tregs from 24-month female B6 mice. At 5 weeks post-adoptive transfer, recipients of aged Tregs alone had significantly lower inflammation scores compared to recipients of autoreactive T cells alone. While all recipients had some inflammation, the average inflammation score of aged Tregs alone was 0.76, which was comparable to the average inflammation score of recipients of naive CD4⁺ B6 cells (0.63) (Fig. 3A). One possibility for this is contamination from other CD4⁺ T cell populations in the recipients. Together, this suggests that despite having a more inflammatory phenotype, aged Tregs alone are not sufficient to cause disease in lacrimal glands in our adoptive transfer model.

Male CD25KO T cells cause greater expression of inflammatory transcripts than female T cells in RAG1KO recipients

While SjD disproportionately affects women, it also occurs in men. To investigate sex differences in our adoptive transfer model, we isolated CD4⁺ T cells from male B6 and CD25KO murine spleen and cervical draining lymph nodes and adoptively transferred them into male RAG1KO mice. We assessed disease once every week for 5 weeks post-adoptive transfer as we did previously. We identified inflammation in the lacrimal glands at 3 weeks post-adoptive transfer, although the inflammation score was not significantly higher compared to controls until 4 weeks after adoptive transfer (Supplemental Fig. 4A). We noted gradual worsening of the architecture of the lacrimal gland until 5 weeks post-adoptive transfer. By qPCR, we noted a similar trend. Ifng, Tnf, Ciita, and Il12a were all significantly higher 3 weeks after adoptive transfer (Supplemental Fig. 4B). Ifng and Il12a continued to be significantly higher 4 and 5 weeks after adoptive transfer, and Ciita was again significantly higher 5 weeks afterwards (Supplemental Fig. 4B). Altogether, this suggests that, as in the female time course, inflammation develops in the lacrimal gland 3 weeks after adoptive transfer and progressively worsens.

We then combined the time course data from the male and female cohorts to identify sex-specific differences at each time-point. There were no differences between male and female lacrimal gland inflammation scores at any time point, confirming that both sexes develop Sjögren-like disease (Supplemental Fig. 5A). When we compared qPCR data, we observed that male recipient lacrimal glands had higher fold mRNA transcript expression than females (Supplemental Fig. 5B). By 3 weeks after adoptive transfer, Tnf, Ciita, Il12a, and Casp3 were all significantly higher in the male recipients than in the females (Supplemental Fig. 5B). Casp3 was likewise significantly higher in males compared to females at the 2-week time point and the 5-week time point (Supplemental Fig. 5B). This discrepancy between the histology and qPCR indicates that further investigation of sex differences in this animal model is warranted.

Finally, we asked if Tregs sorted from young male mice had the same suppressive ability as young female Tregs. We performed a co-adoptive transfer of autoreactive T cells and young Tregs sorted from male 8-to-12 week-old B6 mice. Young male Tregs have similar suppressive ability compared to the female Tregs (Supplemental Fig. 6A, 6B). The inflammation scores of the young Tregs cohort were significantly decreased compared to the cohort of autoreactive T cells alone (Supplemental Fig. 6A). Recipients of B6 CD4⁺ T cells also had minimal lacrimal gland infiltration (Supplemental Fig. 6A, B), confirming that non-autoimmune CD4⁺ T cells do not cause inflammation. This suggests that sex does not affect the suppressive capability of Tregs from young mice.

Altogether, this suggests that male T cells behave the same as female T cells, despite the increase in gene expression of inflammatory markers.

Discussion

In this study, we evaluated the pathology of the CD25KO mouse. First, we evaluated the effect of a lack of Tregs on the tear-secreting lacrimal gland and the tears it produces. We found that CD25KO mice have significantly more inflamed lacrimal glands and a more inflamed tear profile. We then investigated the pathogenicity of CD4⁺ T cells isolated from spleen and cervical draining lymph nodes of these mice. We found time-dependent autoimmunity in immunocompromised hosts that received an adoptive transfer of CD25KO CD4⁺ T cells. Finally, we showed that this autoimmunity can be controlled by a co-adoptive transfer of young Tregs with autoreactive CD4⁺ T cells, but that aged Tregs were less suppressive.

We first investigated the lacrimal gland of CD25KO mice by bulk RNA sequencing. Like patients with SjD [65, 66], the CD25KO mouse develops lymphocytic infiltration in the lacrimal gland [10, 12]. Through bulk RNA sequencing, we found genes related to type I interferons and T cell activation to be upregulated in the CD25KO murine lacrimal gland. Our lab has previously found upregulation in both these pathways in SjD patients’ conjunctiva [67]. Furthermore, SjD patients have elevated type I interferon signatures in both minor salivary gland biopsies and in peripheral blood [68–74]. In the NOD mouse, a different murine model of SjD, knockout of type I interferon alpha and beta receptor subunit 1 resulted in significantly reduced lacrimal gland inflammation [75]. We also demonstrated significant upregulation of Ifng and Ifngr1, genes related to the type II interferon-γ. We have previously found that genetic knockout of interferon-γ from the CD25KO mouse slows progression of the development of autoimmunity in the lacrimal gland [63]. Likewise, knockout of interferon-γ and its receptor in the NOD mouse ablated signs of lymphocytic infiltration and inflammation in the salivary glands [76]. Together, this suggests the CD25KO murine lacrimal gland mimics SjD autoimmunity.

We followed this experiment by investigating cytokines in the tears of the CD25KO mouse. Despite the severity of the compromised tissue in the lacrimal gland, there were no differences in tear volume between CD25KO mice and wild-type B6 controls. This is not the only SjD murine model that has lacrimal gland infiltrates but no change in tear volume [77, 78]. Despite the lack of change in tear volume between groups, the tear protein profiles were different. CD25KO mice had significantly higher concentrations of TNF, interferon-γ, and CCL5 than B6 controls. TNF regulates inflammation and is more prevalent in saliva, peripheral blood, and tears of SjD patients compared to controls [79–81]. TNF increases membrane expression of autoantigens SS-A and SS-B in human keratinocytes [82], autoantigens which are prevalent in SjD and used for diagnosis. In age-related dry eye, a TNF inhibitor eye drop was sufficient to reduce the dry eye phenotype [83], suggesting its role is prominent in several dry eye disease states. Interferon-γ is an inflammatory cytokine that is prevalent in SjD patients’ tears, peripheral blood, and minor salivary gland biopsies [69, 84–87]. Ogawa et al. demonstrated that incubation of interferon-γ on cultured epithelial salivary gland cells results in upregulation of T cell attracting chemokines, suggesting interferon-γ plays a crucial role in the development of lymphocytic infiltrations in the exocrine glands [88]. It is likely they also played a role in the massive lymphocytic infiltration and tissue destruction that we observed in the lacrimal glands of our adoptive transfer recipients. Furthermore, interferon-γ induces ferroptosis in human salivary gland epithelial cells [89]. Administration of either TNF or interferon-γ to cultured human salivary gland cells resulted in apoptosis independent of the Fas/FasL pathway [90]. The higher concentrations of these cytokines in CD25KO tears were likely sources of acinar cell apoptosis in the lacrimal glands. Likewise, administration of TNF or interferon-γ to cultured rat parotid gland cells reduced tight junction protein expression and increased cell permeability [91]. CCL5, also known as Regulated upon Activation, Normal T cell Expressed and Secreted (RANTES), is a chemokine that is involved in the recruitment of T cells and lymphocytes. It is found in elevated levels in saliva, tears, and peripheral blood of SjD patients [85, 92–94]. In the NOD mouse, inhibition of RANTES slowed lacrimal gland lymphocytic infiltration [95] and is thought to be part of the pathophysiology of SjD [96]. The higher concentrations of TNF, interferon-γ, and CCL5 in the tears may potentiate immune cell recruitment to the lacrimal gland and perpetuate inflammation and apoptosis of the acinar cells. By acting on the ocular surface, these cytokines might also disrupt tight junction proteins in epithelial cells and increase local inflammation. Altogether, this implies that the inflammation seen in the lacrimal gland is perpetuated by the inflammatory milieu of cytokines and chemokines found in the tears of CD25KO mice.

Next, we investigated the timing of disease progression in our adoptive transfer model’s lacrimal glands. We and others have previously used adoptive transfer models to study SjD [62, 97–99]. One limitation of our study was that we did not perform flow cytometry on the donor T cells we injected into recipients; however, there is literature exploring the phenotype of CD25KO CD4⁺ T cells compared to wild-type controls. CD25KO CD4⁺ T cells have higher expression of CD44 compared to wild-type controls [10, 11, 16, 20, 100], indicating an increase in activated/effector memory CD4⁺ T cells. More recently, CD25KO CD4⁺ T cells have been shown to produce higher amounts of IFN-γ [62, 100]. This suggests that the CD25KO CD4⁺ T cells injected into recipients had previous antigenic encounter and greater activation status compared to B6 CD4⁺ T cells. We identified significant increases in inflammation 3 weeks after adoptive transfer by histology and found a significant increase in the mRNA transcripts Il12a and Tnf, encoding the cytokines IL-12 and TNF, respectively. By 4 weeks after adoptive transfer, there were significant increases in inflammatory transcripts Il12a, Ifng, and Ciita, encoding the proteins IL-12, interferon-γ, and the class II major histocompatibility complex transactivator (CIITA). By 5 weeks, the lacrimal gland tissue was severely compromised, as evidenced by the histology, and so inflammatory transcripts were likely less prevalent because there was less tissue. This suggests that the start of inflammation occurs between 3 to 4 weeks after adoptive transfer. IL-12 can induce Th1 polarization of CD4⁺ T cells that results in secretion of interferon-γ [101]. Furthermore, IL-12 has been suggested to play a role in the pathophysiology of SjD [102], and antibody depletion of IL-12 in germ-free CD25KO mice resulted in decreased lacrimal gland pathology [62]. IL-12’s early increase in the lacrimal glands likely helped further the autoreactive T cell inflammation response that resulted in higher inflammatory cytokines by week four. CIITA is a master regulator of genes involved in MHC II antigen presentation [103]. Increased Ciita implies increased MHC II presentation. Epithelial expression of MHC II increases after addition of interferon-γ [104–108]. Research has demonstrated that autoimmune T cell receptors can be promiscuous for many MHC molecules [109–111]. There is some indirect evidence to suggest that autoreactive T cell generation and expansion can be linked to MHC II expression. In the autoimmune diseases multiple sclerosis and type I diabetes, upregulation of MHC II molecules and their transcription factors has been reported in patients [112–114]. In the non-obese diabetic mouse, islets with CD4⁺ T immune cell infiltration were positive for MHC II mRNA and protein [115]. Aberrant expression of MHC II in the joints of a rat model resulted in a rheumatoid arthritis-like phenotype after administration of bovine collagen type II, a response that did not occur in wild-type rats [116]. It is possible there is an increase in T cell autoreactivity due to the upregulation of Ciita. Together, this suggests that early autoreactive T cells in the lacrimal gland created an inflammatory milieu that perpetuated and enhanced the inflammatory response by week four, resulting in acinar cell death and fibrosis of the lacrimal gland by week five.

We then addressed whether autoimmunity could be controlled by restoring Treg function. Because sorting based on intracellular staining of transcription factors in living Tregs is impossible, we made use of the extracellular glucocorticoid-induced TNFR-related protein (GITR) as a marker for Tregs as we have done previously [32]. GITR is strongly expressed on murine Tregs [117–120] and is an ideal marker for identifying functional Tregs [121, 122]. To confirm this, we performed flow cytometry on representative young and aged populations and confirmed that young CD25⁺GITR⁺ Tregs had similar median fluorescence intensity of Foxp3 compared to aged CD25⁺GITR⁺ Tregs (Supplemental Fig. 3), as we have published previously [32]. We found strong suppression of inflammation and autoreactivity in recipients of young Tregs with autoreactive cells. Similarly, Sharma et al. found that a co-adoptive transfer of Tregs with autoreactive CD4⁺ T cells from the scurfy mouse could curtail autoimmunity in RAG1KO recipients [123]. The relationship between SjD and the prevalence of Tregs remains unclear. Some studies have found increased Tregs in the peripheral blood of SjD patients compared to controls [26, 28]. Others have reported a positive correlation between the number of Tregs in minor salivary gland biopsies and severity of focus score [7, 24]. On the other hand, some studies have reported equal or lower levels of Tregs in the serum of SjD patients [23, 25, 124–126] and lower levels of Tregs in minor salivary biopsy glands with increasing severity of focus score [127]. These discrepancies could be due to the criteria used to identify SjD patients, how long they have had the disease, how Tregs were identified, and if these identified Tregs were equally suppressive between groups. For instance, one study found that while SjD patients had higher numbers of Tregs in the peripheral blood, they had a reduced frequency of pSTAT5⁺ after IL-2 stimulation, implying impaired IL-2 signaling and a reduced suppressive capacity [28]. Since we identified good suppression of autoreactivity with young Tregs, this suggests that in SjD patients, Tregs become dysfunctional. One possible reason for this is the increasing age of patients.

While SjD can occur at any age, the average age of diagnosis is mid-50s [8, 128]. One study found that the prevalence of SjD was seven times higher in a cohort of 70-year-old patients compared to a cohort of 40 year olds, regardless of criteria used to define the disease [2]. However, parsing the differences between age-related sicca symptoms and autoimmunity can be difficult, as aging is accompanied by an increase in systemic low-grade inflammation termed inflammaging [129]. Despite this increase in inflammation, aging is also accompanied by an accumulation of Tregs [29–31]. One explanation to explain this discrepancy is that aged Tregs are less functional than young Tregs. To investigate this question, we demonstrated that Tregs isolated from aged mice had less suppressive capacity than Tregs from young mice. Our findings confirm our earlier reports that aged Tregs are less suppressive [32, 33]. They are also in agreement with other literature that evinces aged Tregs are less suppressive in vivo and in vitro [34–36]. There are some findings that report Treg activity is equally or more suppressive in aged hosts compared to young [31, 37, 38, 130]. These discrepancies may be due to different methods to investigate Treg suppression, what markers they used to identify Tregs, different concentrations of T effector and Tregs, and differing definitions of “aged.” For instance, it is possible that in vivo the Tregs become “ex-Tregs,” T cells that lose Foxp3 expression and become pro-inflammatory due to the inflammatory environment, which may not be identified by in vitro assays [131, 132]. It is possible that in the presence of the pro-inflammatory autoreactive T cells, the aged Tregs became ex-Tregs or effector-Tregs [131, 132]. In our aged Tregs alone group, we identified minimal inflammation (median inflammation score of 0.76), which might suggest contamination by other populations of CD4⁺ T cells or a conversion of Tregs to ex-Tregs. One limitation of this work was that we did not perform flow cytometry on recipient lacrimal glands of the aged Treg alone group, so we cannot confirm if this conversion occurred. However, conversion to ex-Tregs is not likely in the absence of inflammation [131, 132]. Furthermore, literature has found that aged Tregs do not lose Foxp3 expression in thymus, bone marrow, spleen, and adipose tissue with aging [29, 133]. Altogether, our findings found variable suppressive capability of aged Tregs, suggesting that while they retain some function, they were not capable of suppressing the inflammation brought on by the severe autoimmunity of CD25KO CD4⁺ T cells. This biological variability is likely also at play in human patients.

Because SjD disproportionately affects women more than men at a ratio of 9:1 [3], we investigated any potential sex differences in our adoptive transfer model. In our previous work characterizing the CD25KO lacrimal gland, we did not find any sex differences [12], so we were interested to know if sex differences would be present in the pathology of RAG1KO recipients. The trend of increasing inflammation was similar between sexes, but male lacrimal glands had increased expression of all inflammatory transcripts investigated (Tnf, Il12a, Ifng, Ciita) at 3 weeks, suggesting in males the start of inflammation may be earlier than in females. When compared by two-way ANOVA, male recipient lacrimal glands had significantly higher concentrations of inflammatory transcripts than female recipients during week three for Tnf, Il12a, and Ciita. The apoptotic marker Casp3 was significantly higher in males than females at weeks two, three, and five. Some literature suggests that SjD is more severe in men than women, despite women having a higher chance of developing it [4, 5]. It has been reported that men have higher frequency of extraglandular manifestations [134] and higher levels of the autoantibodies SS-A compared to women [4, 5]. Furthermore, men are at a higher risk of developing lymphoma than women [135]. Our adoptive transfer model recapitulates some of these sex differences, but not all. On a macro scale, male and female T cells follow the same patterns of autoreactivity, despite differences in inflammatory gene expression, warranting further investigation.

Taken together, our results indicate that Tregs can play an important role in controlling autoimmunity in a SjD model. Furthermore, they excitingly highlight one potential reason for the development of autoimmunity with aging and provide further evidence that aged Tregs are dysregulatory. Targeting aged Tregs may be a key therapeutic for addressing this disease.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

CSdP and JGG conceived of the study. CSdP designed the study. KKS, LS, GG, ZY, JGG, and CSdP participated in acquisition of data. KKS, LS, JGG, and CSdP analyzed and interpreted the data. KKS wrote the initial manuscript; CSdP, LS, and JGG revised the manuscript. All authors contributed to the article and approved the submitted version.

Funding

This work was supported by the National Institutes of Health/National Eye Institute R01EY030447 (CSdP), EY-002520, and P30 EY021725 Center Core Grant for Vision Research (Core Grant for Vision Research Department of Ophthalmology at Baylor College of Medicine), NEI Training Grant in Vision Sciences T32 EY007001 (KKS); Research to Prevent Blindness (Dept. of Ophthalmology), The Hamill Foundation, The Sid Richardson Foundation, and by Baylor College of Medicine Pathology Core (NCI P30CA125123). This project was supported by the Cytometry and Cell Sorting Core at Baylor College of Medicine with funding from the CPRIT Core Facility Support Award (CPRIT-RP180672), the NIH (CA125123 and RR024574), and the expert assistance of Joel M. Sederstrom. Jeremias Galletti received a Fulbright Visiting Scholar Award to participate in this study, and he is funded by Wellcome Trust 221859/Z/20/Z and Agencia Nacional de Promoción Científica y Tecnológica (Argentina, PICT 2020–00138, PICT 2021–00109). CSdP receives salary support from NIH/NEI R01EY035333 and Caroline Elles Professorship (Baylor College of Medicine). The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Availability

The data obtained from the RNA bulk sequencing can be found in the GEO repository (accession ID GSE269883).

Declarations

Ethics approval and consent to participate

The Institutional Animal Care and Use Committee at Baylor College of Medicine approved all animal experiments (AN-6491). In addition, all studies adhered to the Association for Research in Vision and Ophthalmology for the Use of Animals in Ophthalmic and Vision Research and the NIH Guide for the Care and Use of Laboratory Animals [39].

Consent for publication

All authors consent to the manuscript being published.

Competing interests

CSdP was a consultant for Spring Discovery (May to August 2022). All other authors have no additional disclosures.