The generation of single and double KO mice was described preciously (12, 13). All work with animals was conducted in accordance with the Institute of Molecular Genetics guidelines (permit number 12135/2010-17210) and national guidelines (2048/2004-1020). All experiments performed using mice were approved by the ethical committee of the Institute of Molecular Genetics and were conducted in accordance with the ARRIVE guidelines. Mice were housed in top-filter cages and fed a standard diet with freely available water and food at pathogen-free animal facilities at the Institute of Molecular Genetics. Age- and sex-matched mice were used for the purposes of this study. The experiments, where the results were affected by gender (mice weight, spleen size, red blood cell (RBC) numbers and proportions), are presented in gender separated figures. In all other cases, the results obtained using male and female mice were pooled after we confirmed they are independent of gender. The ratio of male and female was maintained in all experiments.

All antibodies, chemicals, mouse strains, commercial kits, software and important laboratory instruments are listed in Supplementary Tables 1 , 2 .

Mice of different genotypes were sacrificed, immune cells were isolated from the spleen, blood and BM of femurs and tibias. Erythrocytes were lysed in ammonium-chloride-potassium (ACK) lysis buffer. Cell suspensions were filtered through 70 μM strainers to remove cellular debris and washed in phosphate-buffered saline (PBS). Cells were stained on ice for 30 minutes with fluorescent-conjugated antibodies diluted in appropriately titrated concentrations. All antibodies used are listed in Supplementary Table 1 . Labeled cells were analyzed on Symphony flow cytometer (BD Bioscience). Data were obtained using Diva software (BD Bioscience) and analyzed using FlowJo Software.

Spleens were isolated from 6 week-old mice, washed in PBS, fixed overnight in 4% formaldehyde solution and stored in 70% ethanol prior to processing. The specimen were processed for 12 hours in a HistoCore PEGASUS Plus (Leica Biosystems) tissue processor using ethanol and xylene for dehydration and histowax paraffin for impregnation. Following embedding in paraffin, the specimen were sectioned into 5 µm samples and transferred onto microscope slides. The samples were deparaffinized and rehydrated in xylene and serial dilutions of ethanol prior to staining with hematoxylin for 5 minutes and counterstaining with eosin for 3 minutes. The stained specimen were visualized on a Leica DM6000 widefield microscope using a 20x/0.70 objective for tile scan images and 10x/0.40 objective for single images.

BM donor cells (5x10⁶) isolated from WT or O1/3dKO mice were transplanted into lethally irradiated CD45.1 congenic mice (recipient) together with support WT CD45.1 congenic BM cells (support, 5x10⁶). To assess short term hematopoietic reconstitution, recipient blood was analyzed 6 weeks after transplantation for the presence of newly developed WBC by flow cytometry. Long term repopulation abilities were analyzed 12 weeks post transplantation. Recipient mice were sacrificed and the presence of newly developed immune cells was analyzed in blood, BM and spleen by flow cytometry.

Single cell suspensions of WT and O1/3dKO splenocytes were prepared as described above. Cells were labeled with CellTrace™ CFSE or Far Red, respectively (Invitrogen) according to the manufacturer´s instructions. Cells were washed and mixed in 1:1 ratio (5x10⁶ cells of each cell type) in 150 μl of sterile PBS. Cell suspensions were injected into tail veins of recipient mice (WT or O1/3dKO). Mice were sacrificed 24 hours later and cells were isolated from the spleens and BM. Single cell suspensions were labeled with fluorescently conjugated anti-B220 antibody to distinguish B cells and analyzed by flow cytometry for the presence of CellTrace-positive cells. The absolute number of homed cells in each organ was calculated based on frequencies of CellTrace-positive cells and organ cell counts.

Blood was collected by submandibular bleeding into EDTA-coated tubes and analyzed using BC5300 Vet auto Hematology Analyzer (Mindray Bio-Medical Electronics Co., Ltd.).

The extraction of sphingolipids and their analysis by tandem mass spectrometry (LC-ESI-MS/MS) was performed as described previously (13). Briefly, splenocytes or BM cells (6 × 10⁶) were sonicated. Protein concentrations were measured using a Pierce BCA protein assay kit according to the manufacturer’s instructions. Aliquots of 40 µg of proteins were transferred into glass vials containing internal standards (d17:1 sphingosine [Cat. No. 860640P] and d18:1/17:0 ceramide [Cat. No. 860517P], Avanti Polar Lipids) and mixed with 1 ml of 2:1 (v/v) chloroform/methanol solution. The samples were incubated for 60 minutes at room temperature under gentle agitation and filtered through Millex LH 0.45 µm filters (Merck Millipore) prior to LC-ESI-MS/MS analysis.

All statistical analysis was performed using GraphPad Prism software version 7.3. Data are presented as mean ± standard error of the mean (SEM) of at least three independent experiments. Data were tested for normality using the Shapiro-Wilk normality test. Comparison between two groups was analyzed by unpaired two-tailed Student’s t-test. Comparison of more than two groups was evaluated using one-way analysis of variance (ANOVA) with Tukey’s posttest for normally distributed data or Kruskal-Wallis test with Dunn’s posttest for non-parametric data as indicated in the figure legend. P-values of < 0.05 were considered to be statistically significant (*P < 0.05; **P < 0.01; ***P < 0.001).

Suppressed expression of ORMDL3 has been associated with the increased occurrence of autoimmune diseases (23, 24, 26). Deletion of the ORMDL3 protein in experimental animals led to a slight reduction of B cell numbers in their spleens (20). It has been shown that changes in phenotype observed in O3KO mice or cells can be potentiated by the simultaneous deletion of ORMDL1 or ORMDL2 proteins (11–13). To find out if this also applies to B cell development and homeostasis, we took advantage of our recently prepared single and double Ormdl KO mice (12, 13). In contrast to single KO and other double KO mice, O1/3dKO mice are smaller in size (11–13). We observed that spleen sizes of O1/3dKO mice were markedly reduced with a diminished immune cell content ( Figures 1A, B ). This was not a consequence of lower murine body weights of the mice as was evident from reduced spleen/body weight ratios ( Figure 1C ). Unlike in female mice, significantly smaller spleens were also observed in O3KO and O2/3dKO male mice ( Supplementary Figure 1A ). However, in these mice, the number of immune cells was not altered ( Supplementary Figure 1B ). The reduced spleen weight observed in O3KO and O2/3dKO males correlated with a lower body weight of the mice as was indicated by the unchanged spleen/body weight ratios ( Supplementary Figure 1C ).

Since the observed decrease of immune cell numbers was most pronounced in O1/3dKO mice, we decided to focus our study on the role of ORMDL proteins in B cell development and homeostasis using O1/3dKO mice, their single KO littermates (O1KO and O3KO) and WT mice. It should be noted that a significant number of newborn O1/3dKO pups die shortly after birth, which could be attributed to their reduced ability to obtain food and water due to their reduced coordination and neuromuscular status (our observations and (11)). Only a small proportion of the O1/3dKO mice survive beyond 10 weeks of age. Hence, in our studies we used 6 to 8 week-old mice.

Detailed analysis of individual immune cell populations showed that all cell types, B cells, T cells, as well as myeloid cells were reduced in the spleens of O1/3dKO animals ( Figure 1D ). Further analysis revealed that the ratio between B cells and myeloid cells was slightly but significantly shifted in favor of myeloid cells in O1/3dKO mice ( Figure 1E ). However, despite the highly reduced immune cell content, the spleen architecture of O1/3dKO animals remained relatively normal ( Figure 1F ). B cells are crucially involved in the pathogenesis of multiple autoimmune diseases (41). Given the diminished B cell numbers in the spleen, we analyzed which developmental stages are most affected using detailed analysis of B cells in the spleen. Detailed analysis of splenic B cells revealed that all expected populations of B cells were present. Flow cytometry analysis showed that the number of mature B cells (IgD⁺) was significantly reduced at the expense of the immature transitional T1 (IgD^-, IgM⁺, CD23^-) population in spleens of O1/3dKO mice ( Figure 2A ).

As we observed reduced numbers of B cells in the spleens of O1/3dKO mice, we decided to test whether this was the result of increased apoptosis. We measured the proportion of viable, early apoptotic, late apoptotic and dead B cells in the spleens of both single and O1/3dKO mice. We did not observe any significant changes in the proportion of viable or apoptotic B cells between various genotypes ( Figure 2B ). In addition, the proliferation of splenic B cells was not affected by the loss of ORMDL proteins ( Figure 2C ).

Our results showed that the simultaneous deletion of ORMDL1 and ORMDL3 proteins leads to a significant decrease in the number of immune cells, particularly B cells, in the spleen. Although proliferation and apoptosis remained unchanged, B cell maturation was compromised in the spleens of O1/3dKO mice.

Given the reduced number of immune cells in the spleen, we investigated whether their circulation was impaired. We measured the amount of white blood cells (WBC) and red blood cells (RBC) in blood using Hemavet veterinary hematological analyzer and flow cytometry. The amount of WBC, particularly lymphocytes and monocytes, was decreased in the blood of O1/3dKO mice ( Figures 3A–D ). In the blood, like in the spleen, the ratio of B cells to myeloid cells was slightly but significantly shifted in favor of myeloid cells in O1/3dKO mice ( Figure 3E ). The amount of RBC did not vary between the different genotypes in female mice ( Figure 3F ). We noted a reduction in RBC count in male O3KO mice ( Supplementary Figure 2A ). RBCs and RBC parameters from male and female mice were purposefully analyzed separately as differences in red cell mass and hemoglobin concentrations vary between males and females. In a physiological steady state, venous hemoglobin levels are lower in female mammals than males (42).

Analysis of RBC parameters revealed that the average volume of RBCs sampled from the blood of O1/3dKO mice was lower than that of WT mice, as reflected by the lower mean corpuscular volume (MCV) values ( Figure 3G , Supplementary Figure 2B ). As MCV values are an indication of the size of RBCs, we observed that the RBCs sampled from the blood of O1/3dKO mice were smaller than those of WT mice. This microcytosis, or smaller RBC size, was observed in both, male and female, O1/3dKO mice ( Figure 3G , Supplementary Figure 2B ). Consistent with the lower MCV values, the average mass of hemoglobin per RBC, or mean corpuscular hemoglobin (MCH), was decreased in the RBCs of O1/3dKO mice compared with the MCH of WT mice ( Figure 3H , Supplementary Figure 2C ). The concentrations of hemoglobin per RBC, or mean corpuscular hemoglobin concentration (MCHC), was comparable between the various genotypes ( Figure 3I , Supplementary Figure 2D ). Though we noted a trend toward reduced overall hemoglobin levels in female O1/3dKO mice, total hemoglobin concentrations (HGB) were significantly decreased in male mice only ( Figure 3J , Supplementary Figure 2E ). Overall, our results showed that blood homeostasis is significantly impaired in the concurrent absence of ORMDL1 and ORMDL3 proteins.

Similarly to the spleen and peripheral blood, we found a reduced number of B cells in the BM of O1/3dKO animals as well ( Figure 4A ). However, this decrease of B cell numbers could be caused by the reduced size of O1/3dKO mice. The proportion of B cells within the WBC population in the BM remained unchanged ( Figure 4B ). To find out whether early development of B cells in the BM was affected by the deletion of ORMDL1 and ORMDL3 proteins, we analyzed the different developmental stages of B cells in the BM ( Figures 4C, D ). We found an alteration of B cell development in O1/3dKO BM, where the ratio of B220^high to B220^low cells was significantly shifted toward less mature B220^low cells ( Figure 4C ). In depth analysis of these populations showed reduced numbers of mature B cells (IgD⁺, IgM^-) in O1/3dKO mice while no significant changes in the numbers of B cell precursors (IgD^-, IgM^-) and immature B cells (IgD^-, IgM⁺) were observed ( Figure 4D ). Since the presence of mature B cells in the BM is dictated by the circulation of mature B cells from secondary lymphoid organs and peripheral blood, it is likely that the reduced number of mature B cells in the BM of O1/3dKO mice is a consequence of the reduced number of mature cells found in their blood and spleen ( Figures 1D , 3E ). As such, it appears that B cell development was not disrupted in the BM of O1/3dKO mice. Deletion of SPT long-chain base subunit 1 (SPTLC1), activity of which is controlled by ORMDL proteins, resulted in compromised myelopoiesis, characterized by diminished numbers of common myeloid progenitors (CMPs) and granulocyte-macrophage progenitors (GMPs), although the proportion of hematopoietic stem cells was increased (17). To find out whether early development of lymphocytes is affected by the loss of ORMDL proteins, we studied the proportion of common lymphoid progenitors (CLPs), CMPs and GMPs, as well as megakaryocyte/erythrocyte progenitors (MEPs) in BM. We did not find any significant differences in the number of different immune cell progenitors between the various genotypes ( Supplementary Figure 3 ). Our data show that the deletion of ORMDL proteins does not affect hematopoietic progenitor distributions in the BM.

Given the markedly reduced cellularity of the spleen of O1/3dKO mice in contrast to only limited reduction of immune cell numbers in the blood and BM, we reasoned that a defect in lymphocyte recirculation may underlie the observed phenotype. We therefore evaluated whether the simultaneous deletion of ORMDL1 and ORMDL3 affected B cell homing to the spleen and BM. Lymphocytes isolated from the spleens of WT and O1/3dKO mice were labeled with Cell Tracers (CFSE and CFSE-Far Red, respectively), injected into WT or O1/3dKO mice and allowed to migrate to the recipients’ organs for 16 hours ( Figure 5A ). Flow cytometry was used to identify the relative frequencies of adoptively transferred lymphocytes and to distinguish B cells. The obtained data showed that adoptively transferred splenocytes from both genotypes were able to home to the spleen and, with less efficiency, to the BM. However, no significant differences in homing were observed between WT and O1/3dKO donors and recipients. Our data revealed that the significant reduction of B cells in the spleen of O1/3dKO mice is not caused by impaired homing.

Members of the ORMDL family are expressed not only in immune cells but also in other cell types (5). It has been shown that under pathological conditions, the degree of expression of ORMDL proteins is cell-type specific. For example, upon allergen challenge, ORMDL3 expression was increased more than 100 times in lung epithelial cells and macrophages, but remained unchanged in blood neutrophils (22). To clarify whether the observed defects in B cell development and homeostasis were caused by cell intrinsic or extrinsic factors, we performed a competitive BM transplantation assay. BM cells isolated from WT or O1/3dKO mice were transplanted into lethally irradiated congenic mice together with support WT congenic BM cells in a 1:1 ratio ( Figure 5B ). To assess short term hematopoietic reconstitution, recipient blood was analyzed 6 weeks after transplantation for the presence of newly developed lymphocytes. We did not observe a difference in the percentage of total CD45.2+ cells reconstituted from WT and O1/3dKO donors in the blood of recipient mice 6 weeks post transplantation ( Figure 5C ). Long term repopulation abilities were analyzed 12 weeks post transplantation. In the blood, similarly to the result observed 6 weeks post transplantation, no differences in total CD45.2+ cells reconstituted from WT and O1/3dKO donors were noted ( Figure 5C ). Comparable results were noted in the BM, where no significant changes in CD45.2+ engraftment between WT and O1/3dKO cells were observed ( Figures 5D, E ). In addition, no differences in the engraftment of individual WT and O1/3dKO myeloid cells or lymphocytes were detected ( Figure 5F ). In the spleen, however, we noted a decreased number of splenocytes in mice transplanted with cells isolated from the BM of O1/3dKO mice compared with cells from WT donors ( Figure 5G ). The deletion of ORMDL1 and ORMDL3 proteins affected the overall engraftment of CD45.2+ cells in the spleen, as well as individual immune cell types, myeloid, B cells and T cells, equally ( Figures 5H, I ). From these experiments we conclude that the development of O1/3dKO lymphocytes in the spleen is, at least partially, dependent on intrinsic factors inherent to hematopoietic cells.

Sphingolipids are able to regulate the migration of several cell types, including lymphocytes (43). ORMDL proteins are important regulators of sphingolipid biosynthesis (3–5). Among individual ORMDL proteins, deletion of ORMDL3 causes the highest elevation in sphingolipid levels, which can be further potentiated by the deletion of other family members, depending on the cell type (11–13, 19). To define the role of ORMDL proteins in sphingolipid synthesis in the BM and spleen, we extracted sphingolipids from BM cells and splenocytes obtained from WT, O1KO, O3KO and O1/3dKO mice and analyzed them by tandem mass spectrometry. We measured levels of dihydrosphingosine (d18:0), generated exclusively as an early intermediate in de novo sphingolipid biosynthesis, as well as sphingosine (d18:1) and different ceramides generated by both de novo and salvage sphingolipid biosynthesis pathways ( Figure 6A ). Simultaneous deletion of ORMDL1 and ORMDL3 proteins led to the greatest increase in the levels of all sphingolipids measured. This combined deletion influenced both de novo and recycling pathways of sphingolipid biosynthesis in the spleen and BM ( Figures 6B–K ). The effect of individual ORMDL family members on the synthesis of different sphingolipid subtypes was distinct between the spleen and BM. In the spleen, the deletion of ORMDL3 caused an increase in de novo sphingolipid synthesis, as was evident from the measured concentration of dihydrosphingosine (d18:0) ( Figure 6C ). The amount of sphingosine (d18:1), which is produced by salvage pathways, was not affected ( Figure 6C ). Deletion of ORMDL3 also led to an increased production of ceramides in the spleen ( Figures 6D, E ). The increase in ceramides was primarily noted among molecular species with long (C16-C20) saturated fatty acid chains ( Figures 6E, F ). In the BM, single deletion of ORMDL1 led to an increase in total sphingosines and ceramides comparable to the changes observed with the single deletion of ORMDL3 ( Figures 6G, I ). In contrast to the spleen, the increase in sphingolipids was mainly driven by the salvage pathway, as the amount of de novo intermediate product, dihydrosphingosine, was not significantly altered in single KO cells ( Figure 6H ). It should also be noted that synthesis of ceramides with very-long fatty acid chains (C22-C24) in both, the spleen and BM, was mostly affected by the simultaneous loss of ORMDL1 and ORMDL3 proteins ( Figures 6E, J ). This was reflected by an increased ratio of synthetized ceramides with very-long vs. long acyl chains ( Figures 6D, K ). These data confirm that the influence of individual ORMDL family members on the production of different sphingolipid species is cell-type dependent. The simultaneous deletion of ORMDL1 and ORMDL3 augments the effects seen with the loss of individual ORMDL proteins.

In this study, we analyzed the effects of deleting ORMDL proteins in immune cell development. Increased mortality of O1/3dKO mice shortly after birth and weaning precludes some in vivo experiments. Thus, it is not possible to demonstrate the role of an ORMDL1 and ORMDL3 deficient environment in immune cell development by transplantation assay. It is also challenging to study the function of the immune system of O1/3dKO mice in vivo, given their baseline health status. ORMDL family members are highly homologous and they are at least partially redundant. The degree of these compensatory mechanisms during development is highly variable. Thus, detailed understanding of the mechanism by which the deletion of individual ORMDL proteins affects the immune system requires further study.