Adult C57BL/6 female and male mice were from Jackson Laboratory, Bar Harbor, Maine and maintained under specific pathogen free (SPF) conditions at Indiana University Lab Animal Resource Center. C57BL/6 mice are used in this study because C57BL/6 mice have been vital for our studies of mediators that regulate the development of allergic responses by offspring of allergic mothers (13, 14, 36). The studies were approved by the Indiana University Institutional Review Committee for animals.

αT is necessary for mouse and human placental development (37, 38). Standard basal mouse chow diet contains 45 mg αT/kg of diet and 45 mg γT/kg of diet and supports fetal development in mice. Supplemented αT diets contain 250 mg αT/kg of diet and 45 mg γT/kg of diet (13, 14). Translation of mouse basal α-T doses to humans is calculated as we previously described (page 173 of 39). Taking into account differences in food consumption and metabolism (39), a 45 mg αT/kg of diet for mice is 57 mg αT/day for human adults. For healthy adult humans, 15 mg αT/day is recommended, but asthmatics have low plasma αT (26–29). The 250 mg αT/kg of diet for mice is 285 mg αT/day for human adults, which is well below upper safety limits of 1,000 mg αT/day in human pregnancy and near the 268 mg (400IU) d-αT in pre-eclampsia pregnancy trials (40–45). A relevant dose is a dose that achieves similar fold changes in tissues in mice and humans.

For supplementation of diets with tocopherol, D-α-tocopherol (>98% pure) from Sigma was sent to Dyets, Inc (Bethlehem, PA) to produce the diets with 250 mg αT/kg of diet (catalog#103373) (13, 46). The purity of these tocopherols that were used to make the diets and the tocopherol concentrations in the diets were confirmed by HPLC with electrochemical detection as previously described (13, 46). These αT supplemented diets increase tocopherols 3-fold in mothers and pups (13, 14, 47–49). This is similar to fold tissue changes achievable in humans (15, 16, 21–23, 46).

In our studies, pups received the same allergen or different allergen than the mother because in mice (5–11) and humans (4), allergen responses by offspring are not specific to the allergen to which the mother had been exposed. We induced allergic lung inflammation in mothers or pups with OVA or HDM. OVA is a well-characterized model purified allergen that, in humans, can also be inhaled when exposed to powered egg or when gasping during egg allergic reactions. HDM extract is a model environmental allergen from Greer and has been used in allergy shot induction of tolerance in humans. Mothers with allergic responses and allergic inflammation at the time of mating (15, 46, 50) are mated to non-allergic fathers (13, 14). On gestational day 18 (GD18) [during time of fetal hematopoiesis], we collect mother plasma, placentas, and fetal livers (site of hematopoiesis in the fetus) (13, 14). To assess offspring development of allergy, pups receive a suboptimal allergen sensitization/challenge protocol (Figure 1) (13, 14). There are no differences by sex so data include both sexes (13, 14).

For lung microbiota transfers and microbiota taxa analyses, the microbes were separated from the BAL by differential centrifugation as follows: the BAL was centrifuged for 10 min at 1200 rpm to pellet host cells and supernatant collected. This supernatant was centrifuged at 10 min at 15,000 rpm (500×g) to pellet the microbiome. The supernatant, which contains soluble proteins and mucins (51), was removed and pellet contains the microbiota. The microbiota were 96 ± 0.7% viable as determined by the Biotium Bacterial Viability and Gram Stain Kit with analyses by flow cytometry. Also for visualization of microbiota, separated BAL microbial communities from allergen-challenged pups of allergic and from non-allergic mothers were suspended in a minimal volume, fixed in a small spot on a glass slide, stained with a gram stain for bacteria, and analyzed by microscopy.

C57BL/6 female mice were maintained on chow diet. The mice were sensitized by intraperitoneal injection (200 μl) of OVA grade V (Sigma-Aldrich Co.) (5 μg)/alum (1 mg) or saline/alum (1 mg) on days 0 and 7 (13, 14). The mice were exposed to nebulized saline or 3% (w/v) OVA in saline for 15 min on 3 consecutive days at 8, 12, and 16 weeks of age and then mated. The pregnant and nursing dams received basal diet (45 mg αT/kg of chow) or αT-supplemented diet (250 mg αT/kg of chow) as indicated in the figures.

In experiments with transfer of lung microbial communities, BAL was collected from neonates at PND5. The BAL microbial communities were separated by differential centrifugation For transfer of microbial communities, the pelleted donor microbiota were suspended in saline and immediately administered to lungs of recipient pups to most closely represent the donor microbiome levels and abundance. The few hours between collection of BAL microbiota from donor pups to administration to recipient pups was within the timeframe described for survival of aerobic microbes, as these aerobic microbes can survive for days in PBS (52). The BAL microbial communities of 2 donors were combined for each recipient and administered in 10 µl intranasally to each PND4 recipient pup for adequate inoculation, similarly to studies by others with viral inoculation of neonates (53).

For allergen challenge of the pups, six-day old pups were sub-optimally sensitized by treating with only one 50 µl i.p. injection (rather than two injections) of 5 µg OVA/1 mg alum (13, 14). At 13, 14, and 15 days old, the pups were challenged for 15 min with 3% OVA. At 16 days old, the pups were weighed, euthanized and tissues collected. Pup bronchoalveolar lavage (BAL) cells were stained and counted as previously described (13, 14). OVA-specific IgE was determined by ELISA as previously described (13, 14).

For microbiome analyses, the microbes were separated from the BAL by differential centrifugation. To limit confounding contributions from contaminant bacteria during collection and sequencing reagents (54, 55), data from the BAL microbiome of pups of allergic mothers were compared to control BAL groups and N = 8–10 pups from 3 to 4 mothers per group. The same sterile reagents were used within an experiment. 16S rRNA gene amplicons were generated via PCR amplification using primers 5′TATGGTAATTGTGTGCCAGCMGCCGCGGTAA3′ and 5′AGTCAGTCAGCCGGACTACHVGGGTWGCTAAT3′ (56). The full sequencing protocol is published by Kumar et al. (56) The 16S rRNA gene sequencing was performed on the Illumina MiSeq platform. The ASV table for the mouse microbiome studies was generated by an analysis pipeline using CLC Genomic Workbench Microbial Module (CLCGW-MM). This includes the preprocessing of V4 16S amplicon (250 bp) reads, mapping to SILVA 16S v.132 SSURef, and filtering of initial ASVs (relative abundance > 1 × 10⁻5) as described in the University of Alabama at Birmingham protocol (56, 57). The nomenclature for Bacteroidetes has recently been updated to Bacteroidota (58), therefore we are using these synonymously in this manuscript. Differential abundant analysis was done by a built-in function in the CLCGW-MM, which generated FDR and log2fold differences in the taxa between two comparison groups. Data are shown as % abundance of 16S rRNA gene amplicon counts of total counts within specific taxonomic levels. Alpha and beta diversity analyses were performed using Quantitative Insights into Microbial Ecology v2 2022.8 (59, 60). Principal component analysis was performed using EMPeror (61).

Total RNA was isolated from 50 to 100 mg lung tissue using the QIAGEN RNeasy Mini Kit (catalog #74136). cDNA was prepared using a MMLV Reverse Transcriptase kit (QuantaBio, catalog #95047) and analyzed by PCR on an ABI 7300 Thermal Cycler (Applied Biosystems). Taqman probes and Taqman Universal Master Mix were used as directed (Applied Biosystems, catalog #4304437). Taqmanprobes used were GAPDH (catalog #4331182) and MUC5AC (catalog#4331182). IL-5, IL-13, IL-33, and CCL11 expression levels were quantified using SsoAdvanced Universal SYBR Green (Biorad catalog# 1725271) with the following primers obtained from Integrated DNA Technologies.

Target	Forward primer sequence	Reverse primer sequence
CCL11	TGTAGCTCTTCAGTAGTGTGTTG	CTTCTATTCCTGCTGCTCACG
GAPDH	GTGGAGTCATACTGGAACATGTAG	AATGGTGAAGGTCGGTGTG
IL-33	AATCACGGCAGAATCATCGAGAAA	GGAGCCAGAGGATCTCCGATT
IL-13	CCAGGGCTACACAGAACCCG	GCTCTTGCTTGCCTTGGTGG
IL-5	ACTGTCCGTGGGGGTACTGT	CCTCGCACACTTCTCTTTTTGG

INHANCE cohort (62, 63) is an urban cohort (birth to 18 months of age) in the Indianapolis area (n = 180, 70% Black or mixed-race Black; NIH K23 AI135094-01 PI Kloepfer). Of these 180 infants, 43 of the 3–5 months of age infants and 50 of the 12–18 months of age infants had nasal 16S microbiome data (62, 63) and sufficient plasma volume available for tocopherol analyses by HPLC.Human 16S rRNA gene sequencing data were analyzed using Quantitative Insights into Microbial Ecology v2 2022.8 (59, 60). Human sequences were aligned with the SILVA 138.1 taxonomy database (57). Also serum tocopherol concentrations at 3–5 months or 12–18 months of age were measured by HPLC with electrochemical detection (64) as previously described (24). Because we have demonstrated that in children and adults that better lung function associates with increasing αT when gamma-tocopherol (γT) concentrations are lower (24, 25, 65–67), the INHANCE cohort infants were placed in groups based on below or above median αT and median γT concentrations (64 and manuscript in preparation) withQ1 (high γT, low αT), Q2 (high γT, high αT), Q3 (low γT, low αT) and Q4 (low γT, high αT). Q4 has an anti-inflammatory profile of tocopherols and Q2 has a pro-inflammatory profile for tocopherol isoforms for allergic lung inflammation and lung function as in our previous reports in children and adults (13, 14, 24, 25, 66, 68).

Data in the figures were analyzed by a one-way ANOVA followed by Tukey's or Dunnett's multiple comparisons test (JMP software, SAS Institute). Data in figures are presented as the means ± the standard errors. Data include both genders because there were no differences in outcomes by gender (data not shown). For analyses of the 16S microbiome from INHANCE cohort infants 3–5 months and 12–18 months of age, cutoffs were set for the data, including removal of 3 samples with insufficient ASV detection (<4% total ASV reads/sample compared to other samples), removal of ASV's that had less than 6/43 samples with reads, and to address extension of findings in mice to human, phyla were included in analyses for phyla observed in the mouse models. These participants were placed in 4 groups based on the median αT and γT concentrations (64). Then an abundance cutoff was set at >0.003% abundance for the sum of the ASV averages for the groups. This yielded 181 ASV for the 3–5 months age and 217 ASV for the 12–18 months age. Based on predetermined results in mice with αT supplementation and low γT, analyses were made in comparison to the group Q4 which had a serum αT concentration above the median and a serum γT below the median concentration. There was no formal adjustment for multiple testing because the analyses were selected based on preclinical mechanistic microbiome outcomes. Furthermore, the associations tested were established a priori at the onset of the project with microbiome as the primary analysis with tocopherol isoforms.

Pups of allergic mothers respond to suboptimal OVA sensitization with allergen and this allergen responsiveness of the offspring is reduced by dietary supplementation of the mother with αT during pregnancy and nursing (13, 14). It is not known whether the pups of allergic mothers also respond to HDM and whether this is modified by αT. Pups of OVA-challenged mothers were responsive to suboptimal sensitization and challenge with OVA (Figures 1A,B) or HDM (Figures 1C,D) with increased numbers of leukocytes (Figures 1B,D) and this was blocked by maternal dietary supplementation with α-tocopherol as compared to pups from allergic mothers with a basal diet (Figures 1B,D). OVA increased numbers of eosinophils, monocytes and lymphocytes in the BAL (Figures 1B,D), OVA-specific IgE (13, 14), and lung cytokines (13, 14) in pups of allergic mothers as compared to pups of non-allergic mothers. HDM challenge of pups of allergic mothers (Figure 1C) also increased numbers of eosinophils, monocytes and lymphocytes in the BAL (Figure 1D) and lung IL-5 expression (Figure 1E). It has been demonstrated that pups of allergic mothers that are challenged with saline do not have allergic lung inflammation and that BAL cell numbers are similar to allergen-challenged pups of non-allergic mothers.(5) There were no sex differences in pup weight or eosinophilia as we previously reported (13, 14), so data include both sexes.

The airway microbiome is altered in adult humans with allergic asthma and in adult mice with allergic lung inflammation (69). This airway microbiome dysbiosis in adults has an increased abundance of Proteobacteria and decreased Bacteroidota (69–71). It is not known whether the lung microbiome is altered in pups of allergic mothers. It is also not known whether the lung microbiome plays a role in regulation of airway response to allergen. Interestingly, at PND4 before pup exposure to allergen, the BAL of pups of allergic mothers with a basal diet had increased abundance of Proteobacteria and decreased abundance of Bacteroidota (log2FC > 0.6, FDR < 0.1) as compared to pups of non-allergic mothers with a basal diet (Figure 2A). To assess whether allergen alters the bacterial microbiome of pups, the pups of allergic mothers and pups of non-allergic mothers were challenged with a purified allergen. OVA was used as a purified protein allergen, thereby avoiding contaminant bacterial 16S in the extracts from HDM. The microbiota, that was separated from the BAL of allergen-challenged PND16 pups, contained gram negative and gram positive microbiota as determined by gram-staining of BAL bacteria fixed to glass slides (Figure 2C). The 16S analyses of the PND16 BAL microbiota of allergen-challenged pups of allergic mothers demonstrated an increase in abundance of Bacteriodota and decrease in abundance of Proteobacteria (Figure 2B). There were increases in several taxa within the Proteobacteria, Firmicutes, Fusobacteria and Verrucomicrobia, but decreases in taxa with the Bacteroidota and several other Firmicutes, Proteobacteria and Archaea as compared to pups of non-allergic mothers and as compared to pups of allergic mothers supplemented with αT (Figure 3).

Neonate bacterial load increases over PND0-14 (72). To address a potential function for the microbiome dysbiosis in allergen responsiveness, the microbial community was obtained from the BAL of PND4 pups without allergen exposure. The donor PND4 BAL microbial community was separated from the BAL and transferred intranasally to recipient PND4 pups. Then, the PND4 pups without donor microbiota and the PND4 pups that received the microbiota transfers were challenged with allergen (Figure 4A). For the transfers, the pup groups are designated as maternal treatment of the donor pups → maternal treatment of the recipient pups. The donor sample 16S microbiome had increased Proteobacteria and decreased Bacteroidota taxa in the lungs of offspring of allergic mothers (log2FC > 0.6, FDR < 0.1) (Figure 2A).

The PND4 donor microbiota were also analyzed for alpha-diversity and beta-diversity. The PND4 donor microbiota groups had a similar Shannon within-group alpha-diversity index; the Shannon Index incorporates total number of bacterial species and relative differences in the abundance of various species in the microbiota community of a group (Supplementary Figure S1A). For beta-diversity analysis, the donor groups did not separate in the Principal Component Analysis (PCA) of the Unweighted Unifrac and Weighted Unifrac between-group beta-diversity analysis of bacterial microbiota (Supplementary Figures S2B,C, left panels); the weighted-Unifrac analysis incorporates only the relative abundance of taxa shared between samples and the unweighted-Unifrac analysis incorporates only the presence/absence of taxa between groups. In contrast, when incorporating both overall abundance per sample and abundance of each taxa of the microbiota communities by the Bray-Curtis beta-diversity distance analyses, there was clustering by PCA for the donor saline groups and for the donor OVA groups, which was unaffected by αT (Supplementary Figure S2A, left panel).

Alpha-diversity was also assessed for the PND16 pup microbiota from the BAL of allergen-challenged pups with and without microbiota transfer. Without the microbiota transfers, the BAL of allergen-challenged PND16 pups from allergic mothers (the no donor → OVA/B group in Supplementary Figure S1B) had decreased alpha-diversity as compared to the saline groups (the no donor → Sal/B group and the no donor → Sal/αT group) (Supplementary Figure S1B). With the microbiota transfers, the BAL of allergen-challenged PND16 pups of allergic mothers (OVA) as either donor or recipients (designated as microbiota donor → recipient pairs of pups) had reduced alpha-diversity (Supplementary Figure S1B, red box) as compared to several control groups, including the no donor → Sal/B, the Sal/B → Sal/B, the no donor → Sal/αT or the OVA/αT → OVA/αT (Supplementary Figure S1B).

Beta-diversity was assessed for the PND16 pup microbiota from the BAL of allergen-challenged pups with and without microbiota transfer. There was minimal separation of the PND16 groups in the PCA plot of the Unweighted Unifrac and Weighted Unifrac between-group beta-diversity bacterial microbiota analyses (Supplementary Figures S2B,C, right panels). In the PCA plot of the Bray-Curtis beta-diversity distance analyses of the allergen-challenged pups without donor microbiota transfers (Supplementary Figure S2A, right panel with cone-shaped symbols), there was some separation of the no donor → OVA/B group as compared to the other no donor groups. In the PCA plot of the Bray-Curtis beta-diversity distance analyses of the allergen-challenged pups that received donor microbiota transfers, there was unique clustering of microbiota from pups of allergic mothers with basal diet (OVA/B) as either donors or recipients (Supplementary Figure S2A, right panel with sphere-shaped symbols); these are the groups with allergic inflammation in Figure 5.

Notably, when either the BAL microbial community of the donor pup or the recipient pup was from an allergic mother with basal diet (OVA/B), the recipient pup BAL had an increase in abundance of the class Gamma-proteobacteria and decrease in abundance of the class Bacteroidia (Figure 4B), as compared to the saline/B → saline/B group of pups (Figure 4B). In Figure 4C, when the BAL microbial community was from a group with a donor pup or the recipient pup from an allergic mother with basal diet (OVA/B), there was an increase in a Proteobacteria and a Fusobacteria and a decrease in several Bacteriodota taxa and a Firmicute. These data suggest that the BAL microbial community of the pups of allergic mothers with basal diet was dominant.

The BAL cells were assessed for the pups in (Figure 4). Without microbiota transfers, the pups of allergic mothers had increases in BAL eosinophils, monocytes and lymphocytes (Figure 5), no-donor groups). After intranasal microbiome transfers and airway allergen challenge, there were increased numbers of BAL eosinophils, monocytes and lymphocytes in the pups that were either recipient pups of mothers in the OVA,basal group or were pups receiving microbiome from pups of mothers in the OVA,basal group (i.e. OVA/B as donor or recipient) as compared with the pups of the control saline/B → saline/B group (Figure 5). The donor → recipient pup groups without an OVA/B group in the donor or the recipients did not develop lung eosinophilia after allergen exposure (Figure 5). Interestingly, the mice were in a specific-pathogen-free facility and dysbiosis of the transferred microbial communities was sustained in recipient pups (Figure 4) and only exhibited in pups with allergic lung responses (Figure 5). These novel transfer studies demonstrate that the dysbiotic microbiome of pups of allergic mothers enhances pup responsiveness to allergen.

Mediators of allergic inflammation were measured including serum allergen-specific antibodies, the chemokines and cytokines that mediate eosinophilia, and the mucin Muc5ac. We have reported that anti-OVA IgE is increased in the OVA/B group and this is reduced by OVA/αT (13). The serum of pups in the OVA/B → OVA/B group and OVA/B → Sal/B group had elevated anti-OVA IgE after allergen exposure (Figure 6), suggesting that the transfer of microbial communities of pups of allergic mothers with basal diet is sufficient to mediate enhanced induction of anti-OVA IgE in these pups. In contrast, there were no increases in anti-OVA IgG2b and anti-OVA IgG1 (Figure 6). The chemokine CCL11, which mediates recruitment of eosinophils, and IL-33, which is important in induction of allergic inflammation, was increased in the groups with OVA/B as donor or recipient and in the no donor → OVA/B group (Figure 7). Similarly, IL-5 and IL-13 had a significant increase in the no donor → OVA/B group and had either a trend or significant increase in most of the microbiota transfer groups with pups of OVA/B-treated moms that were either the donor or recipient of the microbe transfers (Figure 7). Muc5ac was increased in several groups that had OVA/B as donor or recipient (Figure 7). The pups with transfers of microbial communities from pups of saline-treated mothers did not have an increase in CCL11, IL-13, IL-5 or Muc5ac (Figure 7). There was also no increase in IL-33 for the recipient pups with saline-treated mothers, except a small increase for the OVA/αT → saline/B group (Figure 7). Thus, transfer of microbial communities with the OVA/B group as the donor or recipient regulated these mediators of allergic inflammation.

We have demonstrated that in children and adults that better lung function associates with increasing αT concentrations when the gamma-tocopherol (γT) concentration is lower (24, 25, 65–67). To extend our microbiota studies in mice to humans, it was determined whether infants within the INHANCE cohort with an anti-inflammatory tocopherol isoform profile (high αT with low γT levels) had an altered microbiota composition compared to infants with a pro-inflammatory tocopherol isoform profile (high γT levels). To assess infant microbiome associations with an anti-inflammatory tocopherol profile, the INHANCE cohort infants that had 16S microbiota data and sufficient plasma volume for tocopherol analysis were placed in groups based on median αT and γT concentrations (64). The median serum tocopherol concentrations at 3–5 months of age were 28 µM αT and 2.6 µM γT and at 12–18 months of age were 19 µM αT and 2 µM γT (Table 1). The higher medians for 3–5 months infants are consistent with increased tocopherol concentrations during pregnancy (24) that will influence early life tocopherol concentrations in infants. The four groups are Q1 (high γT, low αT), Q2 (high γT, high αT), Q3 (low γT, low αT) and Q4 (low γT, high αT) were defined using the median serum tocopherol concentrations (Table 1). Thus, the microbiome of groups Q1, Q2 and Q3 groups were compared to Q4 because Q4 had the anti-inflammatory profile (low γT, high αT) for allergic lung inflammation, lung function and wheeze (13, 14, 24, 25, 66, 68) and had the highest lung function (64). To examine the associations of αT without elevated γT, as this was the condition in the mouse studies in (Figures 1–7), Q4 was compared to Q3. In infants 3–5 months of age, there was a significance or trend for higher % abundance in some Firmicutes and Bacteroidota taxa in Q4 compared to Q3 (Figure 8A). As infants, the airway microbiome matures from birth to 1 year of life (73–75). In INHANCE infants at 12–18 months of age, there was significantly lower % abundance in a Firmicute in Q4 compared to Q3 (Figure 8B). Moreover, for the group Q2, which has a pro-inflammatory tocopherol isoform profile with allergic lung inflammation and function (13, 14, 24, 25, 66, 68), there was a significantly higher % abundance in taxa of a Firmicute and an Proteobacteria (Figure 8B). These data suggest that tocopherol profiles associate with altered microbiome abundance of several taxa in infants.