By use of IMS, we had reported that at least a dozen ether phospholipids - including plasmalogens - were enriched in splenic germinal centers of immunized mice when compared to the rest of the tissue (Jones et al., 2020). Previous work provided evidence that these lipid molecules were important for the homeostasis of short-lived neutrophils (Lodhi et al., 2015b). However, this effect was questioned based on hematological data from patients with inborn errors of metabolism that affect ether lipid biosynthesis (Dorninger et al., 2015). Moreover, the impact of these molecules on adaptive immune cells or functions is not known. Accordingly, we sought to test for effects of an intervention that interferes with a key biosynthetic pathway used for endogenous synthesis of these molecules. To do so, we started with analyses using induced loss of function for Dhrs7b^f/f and the widely expressed Rosa26-CreER^T2 transgene (Friedrich and Soriano, 1991; Shapiro-Shelef et al., 2005) used in Lodhi et al., 2015b, which reduced hepatic generation of a subset of ether lipids and acts in all hematopoietic cells as well.

We modeled initial experiments on the published work with imaging mass spectrometry (Jones et al., 2020). Mice were treated with tamoxifen using conditions less intense than a regimen shown to have no effect on pre-immune splenic and B lineage populations (Higashi et al., 2009; Uhmann et al., 2009), immunized with SRBC [akin to the analyses in Cho et al., 2016; Brookens et al., 2024; Jellusova et al., 2017; Jones et al., 2020], and analyzed 7 days after immunization (Figure 1A). Functional inactivation of Dhrs7b in T and B lymphocytes was confirmed by western blot analysis using anti-PexRAP Ab (Figure 1B). The frequencies and numbers of CD19⁺ B220⁺ B cells in Dhrs7b^f/f; Rosa26-CreER^T2 - only a small minority of which would have been activated by the immunization - were approximately 0.6 the values for controls (Dhrs7b^+/+Rosa26-CreER^T2, i.e. wild-type locus; Figure 1C and D). In contrast to B cell populations, TCRβ⁺ CD4⁺ T cells and TCRβ⁺ CD4^- CD8 T cells were intact in tamoxifen-treated and SRBC-immunized Dhrs7b^f/f; Rosa26-CreER^T2 mice (Figure 1—figure supplement 1, panels A-C). These results suggest that Dhrs7b is required for fully maintaining the B cell population but indicate that the enzyme is dispensable for the main sets of conventional T cells. In light of the earlier finding that a subset of ether lipids was more prevalent in secondary follicles, we analyzed GC in these samples. Microscopy and flow cytometry analyses after immunofluorescent staining found that less GC B cells were present after immunization of Dhrs7b^f/f; Rosa26-CreER^T2 mice, that is, less IgD^- CD95⁺ GL7⁺ (GC-phenotype) B cells (Figure 1E–I; Figure 1—figure supplement 1, panel D). Quantitation of the immune fluorescence micrography with spleens from immunized mice revealed that the numbers of GC per spleen were halved in Dhrs7bΔ/Δ mice compared with controls (Figure 1I; Figure 1—figure supplement 1, panel E). The sizes of those GCs that did form were substantially reduced as well (Figure 1I; Figure 1—figure supplement 1, panel F). The lower number of B cells found in immunized mice - while a less profound effect than the reduction in GC - tempers the result, and it was possible that the function of helper CD4 T cells was impaired even though their numbers were not affected. Consistent with this hypothesis, the frequencies of Tfh and GC-Tfh cells were lower in Dhrs7bΔ/Δ mice compared with controls (Figure 1—figure supplement 1, panels G-I). While not addressing whether or not B cells are part of the requirement for PexRAP in promoting GC, these results indicate that Dhrs7b is necessary for a normal-sized germinal center response.

To determine if there are B-lineage-specific functions of PexRAP, we used a conditionally active Cre transgene that is expressed specifically in mature B cells (Khalil et al., 2012). Dhrs7b^f/f, huCD20-CreER^T2 mice and huCD20-CreER^T2 were treated with tamoxifen to induce Cre-mediated recombination of the floxed alleles after initial establishment of normal B cell populations (Figure 2A). Western blot analyses showed an almost complete absence of PexRAP protein from B cells purified from tamoxifen-treated Dhrs7b^f/f; huCD20-CreER^T2 (hereinafter, B-cell-specific Dhrs7bΔ/Δ or Dhrs7b D B, shown as Dhrs7bΔ/Δ−B) mice compared to similarly treated huCD20-CreER^T2 controls (Figure 2B). The enzyme encoded by this gene catalyzes the synthesis of a lipid precursor to many - but not all - plasmalogens and other ether phospholipids (Lodhi et al., 2015b; Lodhi et al., 2012; Dean and Lodhi, 2018), and Dhrs7b deficiency impacted neutrophil survival (Lodhi et al., 2015b). When we tested if Dhrs7b inactivation affected the steady-state B cell population, the frequencies and total numbers of CD19⁺ B220⁺ B cells in viable lymphocyte gates were about 20% lower in tamoxifen-treated B-cell-specific Dhrs7bΔ/Δ mice compared to tamoxifen-injected, huCD20-CreER^T2, Dhrs7b+/+controls (Figure 2C). With loss of function after production of mature B cells, that is inactivation of Dhrs7b at ages over 6 wk, we observed balanced decreases in the numbers of multiple subsets within the B lineage without evidence of selectivity among main sub-classes of B cell. Moreover, the surface expression of IgM was comparable in Dhrs7bΔ/Δ mice versus WT controls (Figure 2—figure supplement 1, panels A-F). Thus, the frequencies of MZB and FOB amidst the smaller population of splenic B cells in tamoxifen-treated Dhrs7b^f/f; huCD20-CreER^T2 mice were similar (Figure 2—figure supplement 1, panel B), and frequencies of B1 B cells in the peritoneal cavity also were unaffected (Figure 2—figure supplement 1, panel E). Prior work documents normal frequencies and numbers of developing and mature B cells under harsher conditions of tamoxifen-induced deletion with CreER^T2 (Higashi et al., 2009). De novo B cell production rates are too low to yield the ~20% decrease observed 1 week after gene disruption, and few splenic B cells are in cell cycle. Accordingly, we infer that after maturation of a B cell, Dhrs7b promotes its survival, albeit to a modest degree.

B cell population growth can be impacted by increased cell death and/or decreased proliferation. To test if PexRAP affects proliferation in vivo, Dhrs7bΔ/Δ B cells were stained with CellTrace Violet (CTV) and adoptively transferred into B-cell-deficient µMT recipient mice. Strikingly fewer PexRAP-depleted B cells were recovered (Figure 2D), and the frequencies of Dhrs7bΔ/ΔB cells that were IgD^neg (i.e. had been activated) were half those of controls (Figure 2E; Figure 2—figure supplement 1, panel G). As compared to control B cells, CTV partitioning analyses suggested that division rates in vivo were modestly lower for the Dhrs7bΔ/Δ B cells (Figure 2E and F). Mitogen-stimulated Dhrs7bΔ/ΔB cells that proliferated in culture exhibited substantially less robust division: frequencies of viable cells that had undergone ≥3 divisions were halved in Dhrs7bΔ/Δ B cells and yielded ~1/3 as large a progeny population (Figure 2G–I; Figure 2—figure supplement 1, panel H). Collectively, these data indicate that the Dhrs7b gene product supports homeostatic maintenance of a quiescent (pre-immune) B cell population in vivo and effective proliferation of B cells.

As noted, lipidomic analyses that used two-dimensional image mass spectrometry (2D-IMS) discovered that local concentrations of a subset of ether lipids are enriched in GC compared with the area outside of GC after immunization (Jones et al., 2020). To measure the extent to which PexRAP expression within mature B cells can affect their lipid content and composition, including their ether- and plasmalogen phospholipids, LC-MS-MS analyses were performed with B cells directly isolated from spleens as well as those cultured after mitogenic activation. These analyses showed substantial reductions in many ether phospholipids, as well as lysophosphatidylethanolamines (LPE) generated by phospholipase cleavage of the R2 fatty acid sidechain of an ether phospholipid, in PexRAP-deficient B cells (Figure 3A and B; Figure 3—figure supplement 1, panels A, B). While beyond the scope of this work, this evidence suggests that the generation of potential signaling molecules via phospholipase(s) such as PLA2 may be reduced. Of note, Dhrs7b inactivation reduced levels of a subset of plasmalogens in naive B cells (Figure 3A, B and D; Figure 3—figure supplement 1, panel A), although changes of others were modest or absent in the naive population. A greater impact of PexRAP on quantitative amounts of phospholipids was found in the activated B cells. Several species, skewed towards those with the most detected ions, a number of which were at less than 1/5th the level of non-deleted control B cells (Figure 3C and D). Phospholipid results observed with naive and activated B cells from tamoxifen-treated CreER^T2-expressing mice were similar to those of unmanipulated controls (Figure 3—figure supplement 1, panels C, D). There is no in vitro surrogate that faithfully represents GC or the GC B cell, but collectively these data provide strong evidence that the capacity to rapidly produce increased levels of many ether lipid molecules after lymphocyte activation depends on a functional Dhrs7b gene in B cells. The data provide evidence that Dhrs7b in B cells enhances cell-autonomous biosynthesis of some plasmalogens and other ether lipid species. Especially in resting B cells that are out of cycle, additional uptake or biosynthesis pathways may add to the overall pool (Lodhi et al., 2015b; Lodhi et al., 2012). Notwithstanding these potential contributions, however, the straightforward inference is that PexRAP catalyzes the generation of ether lipid precursors in amounts crucial for regulating the overall ether phospholipid pools, especially after B cell activation.

Our earlier work (Jones et al., 2020) reported only ion mode features in negative mode and did not analyze the primary follicle (the vast majority of which are resting B cells) relative to the extrafollicular white pulp, red pulp, and GC. To investigate the requirement for specific expression of PexRAP, we started with AID-GFP mice to enhance spatial localization. These measurements confirmed that a substantial number of both positive and negatively charged ions whose exact masses identified them as plasmalogens were substantially concentrated in GC as compared to the rest of the B cell follicle (Table 1; Figure 4—figure supplement 1, panels A, B). Accordingly, we tested if the differential enrichment of any ether lipid species in primary or secondary follicles depends on biosynthesis within a specific lymphoid lineage. Mice with conditional mutations were tamoxifen-treated, immunized with SRBC, harvested seven days thereafter, and analyzed by IMS (Figure 4A). With deletion driven by the widely expressed Rosa26-CreER^T2 transgene, as in earlier work Lodhi et al., 2015b and Figure 1, substantial reductions of the GC enrichment pattern were observed (Figure 4B). For instance, two representative ions identified in negative ion mode (i.e. m/z 752.5545 and m/z 776.5556) were again enriched in GC regions of spleens from immunized WT mice. These features barely increased in splenic samples of immunized Dhrs7bΔ/Δ (tamoxifen-injected Dhrs7b^f/f; Rosa26-CreER^T2) mice (Figure 4B).

To determine if the levels of particular ether phospholipids in GC regions were affected by PexRAP expression in B cells, Dhrs7b^f/f; huCD20-CreER^T2 mice and huCD20-CreER^T2 controls were analyzed after tamoxifen injections followed by immunization and compared to samples from unimmunized mice (Figure 4C). IMS using both positive and negative ion modes identified at least eight ions - including m/z 752.5545, m/z 776.5556, and m/z 872.5749 - much more substantially (~two- to threefold) concentrated in GC of immunized control mice (Figure 4D and E; Table 2) than in the primary follicle of unimmunized mice. Notably, these increased signals in secondary follicles (GC) were reduced or almost completely eliminated in GCs of mice immunized after B-cell-specific PexRAP depletion (Dhrs7b ΔB cells; Figure 4D and E; Table 2; Figure 4—figure supplement 1, panel C). Collectively, these results indicate that Dhrs7b gene expression in activated B cells regulates the spectrum of ether lipids in the micro-anatomic locale of a secondary follicle. Interestingly, immunization also led to modest increases in the levels of some ether and conventional phospholipids in the B-cell-rich primary follicle regions (Table 2). This finding extends the previous report (Jones et al., 2020), which focused on the relationship between the AID-GFP^hi region (i.e. GC) and lipid features. Moreover, the absence of PexRAP blunted most of the immunization-induced increases observed in the primary follicles. These findings and those of Figure 3 suggest that there are secondary consequences of Dhrs7b inactivation (e.g. some di-acyl phospholipids are affected), but provide direct evidence that the enhancement of many ether phospholipid signals in the GC depends on PexRAP in B cells.

 As key components of adaptive immunity, progeny of an activated B cell can differentiate into Ab-secreting plasma cells (PC) or into GC B cells. Over the course of an immune response, the GC reaction diversifies the affinity for and breadth of antigen recognized by the original BCR and improves properties of memory (Good-Jacobson and Shlomchik, 2010; De Silva and Klein, 2015; Cyster and Allen, 2019; Victora and Nussenzweig, 2022; Inoue and Kurosaki, 2024). PCs that develop after a second encounter with antigen increase the amounts of high-affinity Ab. The higher quality and quantity of Ag-specific Abs from plasma cells are integral to humoral immune responses (Inoue and Kurosaki, 2024). An increasing body of evidence indicates that metabolic reprogramming and intermediary metabolites modulate immune cell differentiation and function (Chapman and Chi, 2022; Bacigalupa et al., 2024; Jellusova, 2020; Boothby et al., 2022; O’Neill et al., 2016; Cho et al., 2016; Weisel et al., 2020; Chen et al., 2021; Urbanczyk et al., 2022; Yazicioglu et al., 2023; Brookens et al., 2024). The observed B-cell-intrinsic function of Dhrs7b in the accumulation of many ether lipid species in both primary and secondary follicles prompted us to test the effect of B cell type-restricted PexRAP depletion on GC responses and Ag-specific Ab production. Tamoxifen-injected Dhrs7b^f/f; huCD20-CreER^T2 and control mice were immunized with NP-ovalbumin (NP-OVA), then boosted with NP-OVA to elicit affinity-matured Ab and harvested 1 week thereafter (Figure 5A). Although PexRAP-deficient B cell numbers were reduced by less than 20% a week after starting gene inactivation prior to immunization (i.e. were over 0.8-fold those of controls; Figure 2A–C), frequencies and numbers of IgD^neg B cells were halved when mice with B-cell-specific inactivation of Dhrs7b were harvested (4 weeks after completion of the induced deletion; Figure 5B and C). Moreover, the GL7⁺ CD95⁺ fraction of the B cell population in the dump / IgD^neg gate was substantially reduced (Figure 5D), such that numbers of GL7⁺ CD95⁺ GC B cells were dramatically decreased in Dhrs7b ΔB mice (Figure 5E). Consistent with these findings, GC in Dhrs7b ΔB mice was reduced after SRBC immunization (Figure 5F and G; Figure 5—figure supplement 1, panels A, B), and the numbers of centrocytes and centroblasts were dramatically lower in Dhrs7b ΔB mice compared with WT controls (Figure 5H). LZ B cells (CD86⁺ CXCR4^lo) trended toward being at lower prevalence in Dhrs7b ΔB mice, whereas their DZ counterparts (CD86^neg CXCR4⁺) were comparable to wild-type controls, resulting in a modest decrease in the LZ/DZ ratio when B cells lacked PexRAP (Figure 5—figure supplement 1, panel C). The frequencies of CD138⁺ GL7⁺ IgD^- early plasmablasts were comparable between Dhrs7b ΔB mice and control mice (Figure 5—figure supplement 1, panel D). Although the loss of function was B cell specific, the frequencies of Tfh and GC-Tfh cells were decreased in Dhrs7b ΔB mice compared with control mice (Figure 5—figure supplement 1, panel E). We infer that support for the Tfh populations was reduced due to impact(s) on GC B cell numbers and/or function.

Ab class-switch recombination (CSR) and affinity maturation can occur through extrafollicular responses. However, the GC reaction significantly increases Ab diversification and high-affinity Ab production, in part later in a primary response but also upon secondary exposure to antigens (MacLennan et al., 2003; Rajewsky, 1996; Boothby et al., 2019; De Silva and Klein, 2015; Cyster and Allen, 2019; Victora and Nussenzweig, 2022). ELISA performed with sera at 3 week post-immunization (Figure 6A–D), just before the boost, showed that B-cell-specific depletion of PexRAP prior to immunization reduced NP-specific IgM and the switched isotype IgG1 (Figure 6A and C). Of note, the capacity to generate high-affinity Ab, detected with low valency NP₂, was even more severely undermined than the overall response - especially for IgG1 (Figure 6B and D). A week after a second immunization - an interval similar to that which followed immunization with SRBC - Ab-secreting cells (ASCs) in the spleen (Figure 6E) and circulating anti-NP IgM concentrations were diminished in Dhrs7b ΔB mice compared with controls (huCD20-CreER^T2, Dhrs7b^+/+ treated with tamoxifen in parallel to the mice with induced loss of PexRAP) (Figure 6F). The ratio of high-affinity (NP₂-binding) to all-affinity (NP₂₀) IgM Ab also was substantially lower in Dhrs7b ΔB mice (Figure 6G), as were the serum levels of high-affinity IgG1 and IgG2c, class-switched isotypes (Figure 5—figure supplement 1, panels F, G). These data reinforce and extend the conclusion that the expression of PexRAP in mature B cells is important for their physiology and function.

Consistent with in vivo results finding reduced Ag-specific ASCs in Dhrs7b ΔB mice compared with controls (Figure 6E), the attenuated population of PexRAP-deficient B lineage cells recovered after transfer into recipient mice (Figure 2D) yielded lower frequencies of CD138⁺ progeny (Figure 6H and I) along with evidence of reduced activation (i.e., lower frequencies of IgD^- progeny [Figure 2—figure supplement 1], panel G). Mitogen-stimulated Dhrs7bΔ/ΔB cells also yielded lower frequencies of CD138⁺ cells in vitro, but the division-specific frequencies of CD138⁺ cells showed only modest decreases in Dhrs7bΔ/ΔB cells (Figure 6—figure supplement 1, panels A, B). Thus, the reduction of CD138⁺ cell differentiation appears to be due mostly to its dependence on survival to a sufficient division count.

Dhrs7b inactivation was initiated prior to immunization in the preceding experiments. Therefore, the impact of PexRAP deficiency on GC and the Ab response in such a setting might be due exclusively to impairment of B cells, for example their reduced population expansion, prior to their entry into GC. Alternatively, the effects could also in part involve a requirement for PexRAP-dependent metabolites within GC B cells. To test if Dhrs7b functions within GC, we used conditional deletion of Dhrs7b driven by the S1pr2-CreER^T2 transgene whose expression at high levels marks GC B cells (Shinnakasu et al., 2016). Of note, experiments with a fate-marking reporter allele showed that activated B cells that lack the GC B phenotype were not marked by this conditional Cre after immunization with the NP-carrier approach (Shinnakasu et al., 2016). Dhrs7b^f/f; S1pr2-CreER^T2 and control S1pr2-CreER^T2 (control) mice were immunized with SRBC, and tamoxifen was injected at a time point after the initiation of GC to test more directly that the inactivation of Dhrs7b would be in GC B cells rather than pre-GC blasts (Figure 7A). Frequencies of GC B cells were substantially lower in tamoxifen-treated Dhrs7b^f/f; S1pr2-CreER^T2 mice (Figure 7B and C). While this finding does not exclude that there may be an additional effect of impaired clonal expansion in B cells activated by immunization but not yet resident in the GC, the results indicate that PexRAP functions within GC B cells to support a full population of this subset. All together, these data provide evidence that Dhrs7b expression influences the lipidome and function of GC B cells. Notably, an optimal GC response requires PexRAP function in B cells, and the B-cell-intrinsic functions of Dhrs7b promote the quantity and affinity maturation of Ab responses, in part via net proliferation of B cells.

Increased susceptibility to cell death after activation would be a mechanism that could cause the reduced population expansion, GC, and Ab production. Consistent with earlier analyses of LPS-stimulated B cells (Price et al., 2018), we had noted that ROS levels were higher in GC B cells, memory B cells (MBCs) and ASCs than in the naive B2 B cell pool (Brookens et al., 2024). ROS can be produced in activated B cells via diverse sources that include mitochondrial electron transport chain function, NADPH oxidase activated by BCR engagement, and fatty acid oxidation in peroxisomes (Capasso et al., 2010; Wheeler and Defranco, 2012; Tsubata, 2020). ROS can positively mediate signal transduction, but their steady-state concentrations need to be calibrated because overproduction can have deleterious effects on cells, for instance via lipid peroxidation at excessive rates (Jiang et al., 2021; Nordgren and Fransen, 2014; Liang et al., 2022). Endogenous antioxidant properties have been imputed to plasmalogens because the vinyl ether bond is susceptible to reaction with reactive oxygen, thereby scavenging ROS (Dean and Lodhi, 2018; Braverman and Moser, 2012; Gorgas et al., 2006; Jiménez-Rojo and Riezman, 2019), so we tested if Dhrs7b function contributes to redox control in B cells. ROS levels were ~twofold higher in GC B cells from immunized Dhrs7b ΔB mice compared to those from immunized WT controls (Figure 7D). Dhrs7bΔ/Δ GC B cells also showed higher cell death and attenuated BrdU incorporation (Figure 7E and F; Figure 6—figure supplement 1, panel C). Moreover, CD40-stimulated Dhrs7bΔ/ΔB cells exhibited ~threefold higher signal and about a doubling when stained with fluorescent sensors of cellular and mitochondrial ROS, respectively, compared to WT B cells in vitro (Figure 8A–D). Substantially increased ROS were also measured after activation by BCR and CD40 cross-linking (Figure 6—figure supplement 1, panel D). Iron- and copper-dependent lipid peroxidation is considered to be a biological mechanism for ROS-mediated cell death (Nordgren and Fransen, 2014; Liang et al., 2022). Lipid peroxidation measured by a fluorescent indicator (C11-Bodipy) in vitro was higher in Dhrs7bΔ/ΔB cells compared with WT controls after each type of mitogen activation (Figure 8E; Figure 6—figure supplement 1, panel E). Chemical elicitation of increased ROS and lipid peroxidation in B cells, independent from B cell activation or proliferation, rapidly led to reduced cell viability and B cell numbers (Figure 8F; Figure 8—figure supplement 1, panel A). Furthermore, in vitro analyses indicated that B cells lacking PexRAP also exhibited increases in the activated executioner caspase, cleaved caspase-3 (CC3), along with early apoptotic cells that are annexin V⁺ but exclude 7-aminoactinomycin D (Figure 8G and H). These data indicate that Dhrs7b function supports not only protection against cell death, but also control of cellular and mitochondrial ROS levels and their impact on lipid modification (Figure 8—figure supplement 1). Of note, while H₂O₂ did not increase this early apoptotic population (Figure 8—figure supplement 1, panel C), increasing ROS with menadione drove B cell death preceded by annexin V⁺ state without an increase in C11-Bodipy signal (Figure 8—figure supplement 1, panels D-G). We infer that rather than a single mode of death, increased ROS resulting from PexRAP depletion probably stimulates at least two distinct modes of B cell death.

Consistent with the increased steady-state ROS and death, activated Dhrs7bΔ/ΔB cells exhibited a defect in B cell population growth in vivo and in vitro (Figure 2; Figure 8A–H). The well-established ROS scavenger, N-acetyl-L-cysteine (NAC), exerted a concentration-dependent effect that improved the survival and population growth of PexRAP-deficient B cells (Figure 8I). We exploited the sub-maximal rescue by a lower concentration of NAC (1 mM) to explore if peroxisomal oxidative metabolism might contribute to the toxicity observed when B cells are PexRAP-depleted. Although thioridazine on its own provided no increase in the B cell population growth, its combination with 1 mM NAC treatment improved growth of the Dhrs7bΔ/ΔB cell population compared to either agent on its own (Figure 8J). Analyses of B cells activated by BCR crosslinking along with anti-CD40 also found that PexRAP expression in B cells supports their population increase (Figure 6—figure supplement 1, panel F). Moreover, the inclusion of NAC in the cultures partially mitigated the inactivation of Dhrs7b (Figure 6—figure supplement 1, panel F). Of note, measurements of CTV partitioning identified a modest but statistically significant decrease in division counts of PexRAP-deficient B cells under these conditions and found that NAC did not increase this component of proliferation (Figure 6—figure supplement 1, panels G, H). These data support the conclusion that sufficient generation of PexRAP-dependent products is crucial for B cell survival. Moreover, this enzyme contributes - directly or indirectly - to a resolution or detoxification of ROS that is critical for population growth of activated B cells.

Our finding that mitochondrial ROS were increased prompted us to investigate how PexRAP affects physiological functions by measuring respiration in activated B cells. Mitochondrial stress tests with in vitro activated B lymphoblasts measured small but definite decreases in both basal and maximal respiration (oxygen consumption rates, or OCR; Figure 9A–C). The altered performance of mitochondria was associated with reductions in calculated ATP generation and proton leak as well as a small decrease in spare respiratory capacity (SRC) (Figure 9D–F, respectively). In contrast, the rates of glucose-stimulated extracellular acidification - cytosolic reactions and a surrogate that can approximate glycolytic activity - were unaffected (Figure 9G). Thus, the effects of PexRAP on B cell metabolism are not global, in that altered respiration did not reflect a decrease in glucose-stimulated acidification. Inasmuch as peroxisomes and mitochondria form contacts and functionally interact with the ER, we measured relative ER mass and found that the signal of the fluorophore ERTracker was consistently reduced in PexRAP-depleted B cells (Figure 9H and I). We infer that both mitochondrial function and ER mass in B cells are promoted by the product of the gene.

Monoclonal antibodies (mAb) against mouse CD40, CD90.2, B220 (CD45R), and other mAbs (purified, biotinylated, or fluorophore-conjugated) were from BD Biosciences or Tonbo Biosciences (San Diego CA) unless otherwise indicated. BAFF was from AdipoGen (San Diego, CA). Recombinant mouse IL-4 and recombinant mouse IL-5 were from Peprotech (Rocky Hill, NJ). Glucose, 2-deoxyglucose, oligomycin, rotenone, antimycin A, carbonyl cyanide 4-phenylhydrazone (FCCP), N-acetyl-cysteine (NAC), thioridazine hydrochloride, H₂O₂, and 4-hydroxytamoxifen (4-OHT) were from Sigma-Aldrich Chemicals (St. Louis, MO), and menadione from Cayman Chemicals (Ann Arbor, MI). Tamoxifen (Tmx) was from APExBio Technology (Houston TX). NP-BSA and NP-PSA (for capture in ELISA) as well as NP-OVA (for immunization) were obtained from Biosearch Technology (Novato, CA). SRBCs (sheep red blood cells) were from Thermo Fisher Scientific (Waltham, MA). CellTrace Violet cell proliferation kit, CM-H₂DCFDA, MitoSOX Red, and BODIPY 581/591 C11 were obtained from Invitrogen (Waltham, MA).

All animal protocols were reviewed and approved by the Vanderbilt University Institutional Animal Care and Use Committee. Mice were housed in ventilated micro-isolators under Specified Pathogen-Free conditions in a Vanderbilt University Medical Center mouse facility and used both male and female mice at 6–8 weeks of age. For cell-type-specific inactivation of the Dhrs7b gene, Dhrs7b^f/f mice (Lodhi et al., 2015b) were crossed with Rosa26-CreER^T2 (Lodhi et al., 2015b), huCD20-CerER^T2 (O’Neill et al., 2016), or S1pr2-CreER^T2 strains (Dorninger et al., 2022; Kuerschner et al., 2012), all on C57BL/6 genetic background. Tamoxifen was administered as previously reported (Mayer et al., 2017; O’Neill et al., 2016). To control for potential Cre toxicity, CreER^T2 mice were similarly injected to use as wild-type (WT) controls. Mice (ages ~8 weeks) were immunized with SRBCs (2×10⁸ cells per mouse) and analyzed 1 week after immunization as described (Mayer et al., 2017; Jones et al., 2020). Alternatively, mice were immunized and boosted by i.p. injections (each of 100 µg NP₁₆-OVA [Biosearch Technologies, Novato, CA] emulsified in 100 µL of Imject Alum [Thermo Fisher Scientific, Pittsburgh, PA]) as described previously (Mayer et al., 2017; O’Neill et al., 2016), and harvested for analyses 1 week after the second injection. Isotype-specific relative levels of Ag-specific Ab were quantitated by capture ELISA using Ag (NP₂₀-BSA or NP₂-PSA for all- or high-affinity hapten-specific Abs, respectively) followed by SBA Clonotyping System (Southern Biotech, Birmingham AL), as described previously (Mayer et al., 2017; O’Neill et al., 2016). As previously described (O’Neill et al., 2016), frequencies of Ab-secreting cells (ASCs) were analyzed by ELISpot and quantitated using an ImmunoSpot Analyzer (Cellular Technology, Shaker Heights, OH).

Splenic B cells were purified (90–95%) using negative selection as previously described (O’Neill et al., 2016) or by positive selection with anti-mouse B220 microbeads (Miltenyi Biotech, Auburn, CA). B cells (5×10⁵ cells in 1 mL) were activated with anti-CD40 (1 μg/mL) and cultured with BAFF (10 ng/mL), IL-4 (10 ng/mL), IL-5 (10 ng/mL), and 4-OHT (50 nM) in Iscove’s Modified Dulbecco’s Medium (IMDM) supplemented with 10% Fetal Bovine Serum (FBS), 100 U/mL penicillin, 100 µg/mL streptomycin (Invitrogen), 3 mM L-glutamine, nonessential amino acids (Invitrogen), 0.1 mM 2-ME (Sigma-Aldrich). To analyze the effect of ROS scavenger and inhibition of peroxisomal lipid oxidation on population growth, bead-purified B cells from tamoxifen-treated huCD20-CerER^T2 or Dhrs7b^f/f; huCD20-CerER^T2 mice were activated with anti-CD40, cultured with BAFF, IL-4 and IL-5 in the presence or absence of NAC (1 mM and 5 mM) and/or thioridazine⋅HCl (100 µM) for 5 days. For enhancement of ROS in WT B cells, purified B cells were analyzed 3 days after activation as for ROS scavenging, but vehicle, H₂O₂ (200 μM) or menadione (8 μM) was added at three different time points (overnight, 3 hr, and 1 hr) prior to processing for flow cytometry to counting and measurements of DCFDA, annexin V, and C11 Bodipy signals.

For the analysis of proliferation in vitro, B cells were purified by depleting CD90.2⁺ T cells and CD138⁺ cells using biotinylated anti-CD90.2 Ab and biotinylated anti-CD138 Ab followed by streptavidin-conjugated microbeads (iMag; BD Bioscience), labeled with 5 µM CellTrace Violet (CTV, Invitrogen), activated and cultured as above. To analyze proliferation in vivo, B cells were purified, labeled with CTV, and injected intravenously into B-cell-deficient µMT recipient mice. Mice were harvested at 4 days after adoptive transfer. To measure proliferation rates of GC B cells, in vivo BrdU incorporation was performed by injecting mice with BrdU (2 mg/mouse; BD-Pharmingen, San Jose, CA) at 16 hr and 4 hr before harvesting spleens from SRBC-immunized mice. Cells were stained for surface markers (B220, IgD, GL7, CD38) followed by fixation, permeabilization, and staining with anti-BrdU-APC Ab using the BrdU Flow Kit (BD-Pharmingen) according to manufacturer’s protocol.

GC B cells were identified as GL7⁺ CD95⁺ events in the viable B220⁺ dump^- gate (dump channel consisting of one fluorophore for CD11b, CD11c, F4/80, Gr1, TCRβ, and 7AAD), while plasmablasts were defined as B220⁺ CD138⁺ TACI⁺ dump^- cells. For flow analyses of total intracellular ROS and mitochondrial superoxide, B cells (1×10⁶ cells) were washed with PBS and stained with 1.25 µM CM-H₂DCFDA or 5 µM MitoSOX Red in PBS (20 min at 37 °C), respectively, then washed with 1% BSA containing PBS, and further stained with anti-B220, anti-CD19, anti-CD138, and Ghost-BV510. To measure the lipid peroxidation, cultured B cells (1×10⁶) were washed with PBS and stained with 1.25 µM C11-BODIPY in PBS (20 min at 37 °C), and washed with 1% BSA containing PBS followed by surface staining as above. To analyze the level of endoplasmic reticulum (ER), B cells (1×10⁶) were stained with 0.5 µM ER-Tracker Green (Molecular Probes, Eugene OR) in Hank’s Balanced Salt Solution with calcium and magnesium (HBSS/Ca/Mg; Gibco) for 30 min at 37 °C.

For measurements of cell death (Hossain et al., 2022), cultured B cells (1x10⁶) were washed twice with PBS, once with Annexin V binding buffer (10 mM HEPES, pH 7.4, 140 mM NaCl, 2.5 mM CaCl₂), and then incubated (15 min at 20 C in the dark) with Annexin V-rPE, 7AAD, APC-conjugated CD138, APC-Cy7-conjugated B220, e450-CD19. Cells were then washed with Annexin V binding buffer and analyzed on the flow cytometer. To measure levels of cleaved caspase-3 (Mayer et al., 2017), an activated apoptosis executor, the cultured cells were rinsed with PBS after harvest, stained (15 min at 4 C) with 7AAD, APC-conjugated CD138, APC-Cy7-conjugated B220, e450-CD19, washed (FBS, 1% v/v in PBS), then fixed with paraformaldehyde (4% w/v, 10 min at 20 C), permeabilized with permeabilization buffer (saponin, 0.2% w/v with FBS 1% in PBS), stained with FITC-conjugated Ab specific for cleaved caspase-3 (BD Biosciences), rinsed, and analyzed by flow cytometry.

Mice were immunized with SRBC, and spleens were harvested at 1 week after immunization. Spleens were snap-frozen on dry ice and mounted to a cryostat chuck with a minimal amount of OCT (Thermo Fisher Scientific, San Jose, CA). Frozen spleens were cut at 12 µm thickness using a Leica CM3050S (Leica, IL, USA) at –20 °C, fixed with 1% paraformaldehyde for 10 min, washed twice with PBS, and blocked with M.O.M (Vector Lab) followed by incubation with GL7-FITC, anti-IgD-PE, and anti-CD35-biotin Ab followed by streptavidin-conjugated Alex647 at 4 °C. After the tile scanning of spleen sections, GC size and numbers were quantified with FIJI Image J software.

For Matrix-Assisted Laser Desorption/Ionization (MALDI) Imaging Mass Spectrometry (IMS), frozen spleens were cut as above and mounted onto indium tin oxide (ITO) coated microscope slides (Delta Technologies ETC). Slides were then vacuum desiccated for at least 30 min before matrix application. Pre-IMS autofluorescence (AF) images were acquired prior to matrix application on a Zeiss Axio Scan Z1 Slide Scanner using the brightfield channel and the DAPI (ex. 340–380 nm; em. 435–485 nm, blue), EGFP (ex. 450–490 nm; em. 500–550 nm, green), DsRed (ex. 538–562 nm; em. 570–640 nm, red) filter cubes. After taking AF images, a custom in-house developed sublimation device was used to apply the matrices 2,5-dihydroxyacetophenone (DHA) and 1,5-diaminonaphthalene (DAN) (Sigma Aldrich, St. Louis, MO, USA) to tissue sections for positive and negative ion mode analyses, respectively (Hossain et al., 2022). For immunized AID-GFP mice, the fluorescence emissions identified GC localization within the section used next for IMS. Alternatively, two serial sections of immunized mice spleens were used, one for 2D-IMS and the contiguous section for immunohistochemistry (IHC). Fluorescence data were registered with the 2D-IMS visualizations of phospholipids to align images (Jones et al., 2020), allowing quantitation of the ion intensities in specific m/z peaks within defined micro-anatomic portions of the spleen (Figure 4B). MALDI IMS data were acquired with a 10 µm pixel size (laser spot size 8 µm) in full scan mode using a Bruker trapped ion mobility time-of-flight (timsTOF) Flex mass spectrometer (Bruker Daltonics Billerica, MA, USA). Data were acquired with 250 shots per pixel and a mass range of m/z 200–2000. SCiLS (software Bruker Daltonics) was used to process the data, normalize ion intensity, visualize ion images, and merge the images.

To prepare samples, a one-phase method was used to extract lipids (Pellegrino et al., 2014). Briefly, 0.5 mL of MeOH/MTBE/CHCl₃ mix (1.3:1:1) was added to a frozen pellet of B-cells (1x10⁶ cells total), spiked with 10 µL EquiSPLASH-lipidomics internal standard mix (Avanti Research), briefly vortexed and shaken gently for 20 min, followed by centrifugation at 20,000 x g for 15 min at 10 °C. The supernatant was transferred to a clean Eppendorf tube, evaporated under a gentle stream of N₂ gas, and resuspended in 100 µL methanol/CHCl₃ (9:1), and 2 µL were used for LC-HRMS (high-resolution MS) analysis. Each sample was injected two times - one injection in positive ESI mode followed by one in negative mode. Pooled QCs were injected to assess the performance of the LC and MS instruments at the beginning, in the middle and at the end of each sequence. Discovery lipidomics data were acquired using a Vanquish UHPLC (ultrahigh performance liquid chromatography) system interfaced to a Q Exactive HF quadrupole/orbitrap mass spectrometer (Thermo Fisher Scientific).

Chromatographic separation was performed with a reverse-phase Acquity BEH C18 column (1.7 mm, 2.1x150 mm, Waters, Milford, MA) at a flow rate of 250 µL/min. Mobile phases were made up of 10 mM ammonium formate and 0.1% formic acid in (A) H₂O/CH₃CN (40:60) and in (B) CH₃CN/iPrOH (10:90). Gradient conditions were as follows: 0–1 min, B=20%; 1–8 min, B=20–100%; 8–10 min, B=100%; 10–10.5 min, B=100–20%; 10.5–15 min, B=20%. The total chromatographic run time was 15 min. Mass spectra were acquired over a precursor ion scan range of m/z 200–1600 at a resolving power of 60,000 using the following HESI-II source parameters: spray voltage 4 kV (3 kV in negative mode); capillary temperature 250 °C; S-lens RF level 60 V; N₂ sheath gas 40; N₂ auxiliary gas 10; auxiliary gas temperature 350 °C. MS/MS spectra were acquired for the top-seven most abundant precursor ions with an MS/MS AGC target of 1e5, a maximum MS/MS injection time of 100ms, and a normalized collision energy of 15, 30, 40. High-resolution mass spectrometry data were processed with MS-DIAL version 4.70 in lipidomics mode (Tsugawa et al., 2020). MS1 and MS2 tolerances were set to 0.01 and 0.025 Da, respectively. Minimum peak height was set to 30,000 to decrease the number of false positive hits. Peaks were aligned on a quality control (QC) reference file with RT tolerance of 0.1 min and mass tolerance of 0.015 Da. Default lipid library was used (Msp20210527163602_converted.lbm2), solvent type was set to HCOONH₄ to match the solvent used for separation, and the identification score cut-off was set to 80%. All lipid classes were made available for the search. After lipid identification was completed, MS-DIAL results were exported into Excel and cleaned using minimum RSD for QC samples set to 20% and minimum ratio of QC to Blank set to 10. For species identified in both PexRAP-depleted (Dhrs7b Δ/Δ) B cells and controls, mean levels (areas under peak curves) and their variance were analyzed in Excel.

Oxygen Consumption Rate (OCR) was measured using Seahorse XF96 extracellular flux analyzer (Agilent Technology, Santa Clara, CA) as described previously (O’Neill et al., 2016). Briefly, in vitro activated B cells were washed twice, resuspended in XF Base Media (Agilent Technologies) supplemented with 2 mM L-glutamine, and equal numbers of B cells (2×10⁵) were plated on extracellular flux assay plates (Agilent Technologies) coated with 2.5 µg/mL CellTak (Corning) according to the manufacturer’s protocol. Before extracellular flux analysis, B cells were rested (25 min at 37 °C, atmospheric CO₂) in XF Base Media. OCR and ECAR were measured before and after the sequential addition of 1.5 µM oligomycin, 0.5 µM FCCP and 0.5 µM rotenone/antimycin A.