Introduction

Systemic lupus erythematosus (SLE) is the most common autoimmune disease that affects approximately 100 individuals per 100,000 people, 10 times more often in women than men, as manifested by multiorgan involvement primarily due to the deposition of circulating immune complexes1,2. The respiratory involvement in lupus is reported in up to 50–70% of patients, including pleuritis, alveolar infiltration, bronchiolitis obliterans, diaphragmatic involvement, pulmonary hypertension, antiphospholipid syndrome (APS), and diffuse alveolar hemorrhage (DAH)3. Although the incidence of lupus-induced DAH is only 0.5–9% of the patients, autopsies of patients with lupus demonstrate focal collections of red blood cells (RBCs) or more diffuse involvement in 30–66% of cases, implying a common subclinical disease4,5. The young age (mean age 27 years) and an early stage of the disease (mean 35 months with a range of 0–276 months) with recurrent episodes are common resulting in fatal outcomes (average mortality rate of 50%) frequently concomitant with lupus nephritis due to the immune complex-driven etiology6. Although adaptive immune response to immune complex deposition is the major underlying mechanism of DAH, roles of innate immune responses against the continuous exposure of environmental molecules and some oropharyngeal organisms are mentioned7. Accordingly, normal microbiota in the lung comprises several bacteria, and an imbalance of the microbiota, referred to as “dysbiosis”, is mentioned in several lung diseases8. The lung dysbiosis might activate innate immunity and cause more severe lung inflammation9. While dysbiosis activates immune responses, immune responses may also possibly induce dysbiosis because the innate immune cells (neutrophils and macrophages) are important for balancing the normal microbiota10. Not only bacteria, but fungi, especially Candida albicans in humans, are the second most abundant organisms in the oropharyngeal, upper respiratory, and gastrointestinal tracts11,12, which might occasionally pass from the upper into the lower respiratory tract13. Surprisingly, Candida colonization in the respiratory tract is very common, as approximately half of all healthy individuals are positive for the colonization without any symptoms14. Also, the transfer of oropharyngeal Candida into the lung is well-known15.

Although Candida spp. is considered a commensal normal human oral microbiota, with little significance attached to the respiratory specimens, impacts of the presence of Candida spp. in the respiratory tract are mentioned16. Additionally, the alteration of bacterial population by fungi and vice versa is evidenced, partly through symbiosis and antagonistic communication10,17,18,19. Indeed, the impacts of neutrophils on DAH and the increased mortality rate of active lupus from DAH are reported4,5. Then, we hypothesized that the presence of C. albicans (the most common fungi in human oropharyngeal microbiota) in the lung might enhance the severity of DAH, and the intratracheal administration of Candida albicans in pristane-induced lupus mice might worsen DAH partly through Candida-induced lung dysbiosis. We aim to compare the severity and lung microbiota of pristane-induced lung hemorrhage in mice with and without intratracheal Candida administration together with the impact of neutrophil responses against microbial molecules. Hence, the intratracheal administration of C. albicans in the DAH mouse model, using the well-known pristane induction, was performed with the in vitro experiments on the impact of bacteria and fungi on neutrophils to test our hypotheses.

Materials and methods

Preparation of Candida albicans

Candida albicans from the American Type Culture Collection (ATCC) 90,028 (Fisher Scientific, Waltham, MA, USA), a fluconazole susceptible strain (minimal inhibitory concentration of fluconazole 0.25–1 µg/mL), were cultured over-night on Sabouraud dextrose broth (SDB) (Thermo Scientific, Hampshire, UK) and counted in a hemocytometer (Bright-Line, Denver, CO, USA) before use. For the in vitro experiments, heat-killed C. albicans were prepared by immersion in a water-bath at 60 °C for 1 h and sonicated (power amplitude 40%, pulse on and off for 20 and 5 s, respectively, for 60 s) by an ultrasonic Homogenizer (Sonics Vibra Cell, VCX 750, Sonics & Materials Inc, Newtown, Conn) until a homogenous solution was formed18.

Animals and animal model

In parallel, 8-week-old C57BL/6 female mice, approximate body weight of 22.5 ± 0.7 g (mean ± SEM), were purchased from Nomura Siam International, Bangkok, Thailand, and performed the experiments under the approved protocols number SST 020/2566 from the Institutional Animal Care and Use Committee of the Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand, following the National Institutes of Health (NIH), USA. Only female mice were used due to the clearer lupus characteristics of pristane injection in female mice than in the males20,21, supporting the higher incidence of lupus in females1,2. All mice were maintained in the standard clear plastic cages (3–5 mice per cage) at 22 ± 3 °C with 50 ± 10% relative humidity, and a 12/12-hour light/dark cycle. The mice were divided into 4 groups, including phosphate buffer solution (PBS) alone (Control), C. albicans alone, pristane injection, and pristane plus C. albicans (Pris + Can). Then, 28 mice were used for the survival analysis and another set of 28 mice were used for the 5-week-endpoint experiments. To induce lung hemorrhage, intraperitoneal injection with 0.5 mL of sterile pristane (Sigma-Aldrich, St. Louis, MO, USA) or PBS alone was conducted following a previous publication20,21. After 1 week of pristane injection, the mice were intratracheally injected (3 days/week) with C. albicans at 1 × 106 colony-forming units (CFU) diluted in 50 µL of PBS (Fig. 1A) and monitored for 5 weeks after pristane injection before sacrifice by cardiac puncture under isoflurane anesthesia with blood and organ collection. All animal experiments in this study were performed in accordance with the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines and regulations.

Fig. 1
figure 1

Schema of the experiments (see method) (A) and characteristics of control, pristane injection, Candida administration (Candida), and pristane plus Candida (Pristane + Candida) as indicated by survival (B), bodyweight (C), lupus characteristics (urine protein creatinine index, serum creatinine, and serum anti-dsDNA) (D–F), gut permeability (FITC-dextran assay) (G), serum inflammatory cytokines (TNF-α, IL-6, and IL-10) (H–J) are demonstrated (n = 5–7/group).

Full size image

Mouse sample analysis and gut permeability measurement

Before sacrifice, mice were collected urine for 24 h by placing them into a metabolic cage (Hatteras Instruments, NC, USA) for the determination of urine protein and creatinine by the Bicinchoninic acid assay (BCA Protein Assay Kit; Thermo Fisher Scientific, Wilmington, DE, USA) and QuantiChrom Creatinine Assay (BioAssay Systems, Hayward, CA, USA), respectively. Serum cytokines, including tumor-necrosis factor alpha (TNF-α), interleukin-6 (IL-6), and IL-10, and creatinine were measured by enzyme-linked immunosorbent assay (ELISA; Invitrogen, Vienna, Austria) and QuantiChrom Creatinine Assay, respectively. Serum anti-double stranded deoxyribonucleic acid (anti-dsDNA), an autoantibody, was analyzed by a previous protocol, using coated Calf-DNA (Invitro, Carlsbad, CA, USA)22,23. To determine gut permeability, fluorescein isothiocyanate-dextran (FITC-dextran), a non-absorbable molecule, was measured by orally administered FITC-dextran (molecular weight 4.4 kDa) (Sigma-Aldrich) at 25 mg/mL in 0.25 mL PBS at 3 h before sacrifice. Serum FITC-dextran was measured by a spectrometry microplate reader (Thermo Scientific) with an excitation and emission wavelength at 485 and 528 nm, respectively24. Moreover, fungal identification in the lung samples was conducted through both culture and polymerase chain reaction (PCR), following established protocols25. Briefly, mouse lung tissue was homogenized by a lysis buffer at 65 °C for 3 h, digested (mechanical bead beater for 20 min at 15 vibrations per second) before the extraction of metagenomic DNA by the Phenol-Chloroform method, and amplified by the universal Internal Transcribed Spacer (ITS) eukaryotic primers; ITS3 (forward: 5′-GCATCGATGAAGAACGCAGC-3′) and ITS4 (reverse: 5′-TCCTCCGCTTATTGATATGC-3′). To determine local inflammation, the cytokines from lung tissue were evaluated followed a previous protocol26. In brief, lung samples (0.1 g) were homogenized through sonication (20 s of pulse-on, 5 s of pulse-off on ice for totally 45 s) before centrifugation and evaluated for tissue cytokines (TNF-α, IL-6, and IL-10) using ELISA assays (Invitrogen).

Histology and lung microbiome analysis

The lungs were collected and separated in half for histology and microbiome analysis. For histology, lung tissues were harvested, fixed in 10% paraformaldehyde, paraffin-embedded sections, and stained with hematoxylin and eosin (H&E)27,28. The semi-quantitative score at 200x magnification was used to determine the histopathology of the lungs with the hemorrhagic area in the field using the following scores; 0, no hemorrhage; 1, 0–25%; 3, 25–50%; 3, 50–75%; and 4, 75–100%29. A lung microbiome analysis was performed according to a previous publication30. Briefly, lung tissue (0.1 g) was used for purifying metagenomic DNA as a template for amplifying the V3-V4 region of the 16S ribosomal ribonucleic acid (rRNA)-encoding gene and was sequenced by the Illumina MiSeq sequencing platform (Illumina, San Diego, CA, USA). Universal prokaryotic 16s rRNA primers, 515F (forward: 5’-GTGCCAGCMGCCGCGGTAA-3’) and 806R (reverse: 5’-GGACTACHVGGGTWTCTAAT-3’), supplemented with Illumina adapter and Golay barcode sequences, were employed to construct the bacteriome sequencing library for the 16 S rRNA gene. The sequences of the detected mouse fecal 16 S rRNA were deposited in the NCBI database under accession number PRJNA1078562.

Analyses of immune cells in the spleens and macrophage characteristics

Because the presence of Candida in the lung might alter lupus characteristics in the spleens and change the activity of intraperitoneal pristane-induced macrophages, the spleens were evaluated by flow cytometry analysis for the abundance of adaptive immune cells (B cells and T cells), and antigen presenting cells, especially the dendritic cells (DC), while peritoneal macrophages were analyzed by reverse transcription polymerase chain reaction (RT-PCR). Mouse spleens from all experimental groups were prepared according to the previous publications31,32. In brief, spleens were crushed through a cell strainer to generate a single-cell suspension and resuspended in staining buffer; 0.5% bovine serum albumin (BSA) and 10% fetal bovine serum (FBS) in PBS. Red blood cells were eliminated by lysis buffer; ammonium–Chloride–Potassium (ACK) buffer: ammonium chloride (NH4Cl), potassium bicarbonate (KHCO3), and ethylenediaminetetraacetic acid (EDTA), before washing with PBS and being resuspended in staining buffer. After that, splenocytes (1 × 106 cells) were stained with fluorochrome-conjugated antibodies against different mouse immune cells, including antibodies against the following antigens; cluster of differentiation 4 (CD4), cluster of differentiation 8 (CD8), L-Selectin (CD62L), cluster of differentiation 44 (CD44), cluster of differentiation 3 (CD3), Germinal center (GL7), Immunoglobulin M (IgM), resting B cell (B220), syndecan-1 (CD138), Integrin (CD11c), cluster of differentiation 80 (CD80), and inducible co-stimulator (ICOS) (Biolegend, San Diego, CA, USA) The flow cytometry was performed using BD LSR-II cytometer (BD Biosciences) and analyzed by FlowJo software version 10 (www.flowjo.com) (USA). Then, several immune cells from the spleens, including plasma cells (CD138 and B220 positive), naïve Th cells (CD62L positive with CD44 negative), total control memory T cells (CD3, CD44, and CD62L positive), total Th cell (CD4 positive CD8 negative), central memory Th cell (CD3, CD4, and CD62L positive), effective memory Th cells (CD3 and CD4 positive with CD62L negative), inducible T cell costimulatory (CD4 and ICOS positive), central memory CD8 T cells (CD8 and CD62L positive), adaptive B cells (GL7 and B220 positive), IgM-producing B cell (IgM and B220 positive), dendritic cells (CD80 and CD11c positive), and activated dendritic cells (ICOS and CD11c positive), were analyzed.

Additionally, the characteristics of peritoneal macrophages from each experimental group (tissue-resident macrophages) were analyzed modifying from a previous protocol33. Briefly, 10 mL of ice-cold PBS containing 10% FBS was washed into the peritoneal cavity of the mice at 5 weeks after the experiments before the retrieval of all fluid. The isolated peritoneal cells were seeded in Roswell Park Memorial Institute (RPMI)-1640 medium supplemented with 10% FBS and 1% penicillin/streptomycin (Gibco Life Technologies, Paisley, UK) for 12 h. Then, the adherent cells (macrophages) were separated and centrifuged at 1,800 × g, 4 °C for 5 min, and resuspended in RPMI-1640 with 10% FBS. Subsequently, the isolated macrophages at 1 × 106 cells were stained with fluorochrome-conjugated antibodies against M1 pro-inflammatory macrophage polarization markers (CD80 and CD86) and M2 anti-inflammatory macrophage polarization (CD206 and CD163) using BD LSR-II cytometer (BD Biosciences) and analyzed by FlowJo software version 10 (www.flowjo.com) (USA) as mentioned above in the splenocytes. In parallel, the expression of several genes in isolated peritoneal macrophages was also determined using RT-PCR as previously published34. Briefly, the RNA was extracted from cell pellets using a FarvoPrep RNA mini kit (Farvogen, Vienna, Australia). The amount of extracted RNA was measured by a NanoDrop OneC Microvolume UV-Vis Spectrophotometer (Thermo Scientific) and a reverse-transcription assay (Applied Biosystems, Warrington, UK), respectively, using an Applied Biosystems QuantStudio 6 Flex Real-Time PCR System with SYBR® Green PCR Master Mix (Applied Biosystems). The results were demonstrated in terms of relative quantitation of the comparative threshold (delta-delta Ct) method (2 − ΔΔCt) as normalized by β-actin (an endogenous housekeeping gene). The primers for target genes are listed in Table 1. Notably, the M1 polarization indicators were evaluated using RT-PCR for the gene expression of iNOS and IL-1β and using flow cytometry analysis for CD80 and CD86. In parallel, the M2 polarization indicators were Arg-1, Fizz-1, and TGF-β using RT-PCR together with CD206 and CD163 using flow cytometry. Meanwhile, the expression of inflammatory genes (TNF-α, IL-6, IL-10, TLR-4, Dectin-1, and NF-κB) was evaluated using RT-PCR to indicate the pro-inflammatory activity of macrophages.

Table 1 Lists of the primers used in the experiments.
Full size table

The in vitro experiments on neutrophils

Because (i) the influences of macrophages in DAH pathogenesis are well-established35 but the data on the impacts of neutrophils in DAH is still less36 and (ii) the synergistic responses against bacteria and fungi of neutrophils in our model are hypothesized, experiments on neutrophils are interesting. As such, neutrophils were isolated from intraperitoneal injection of 1 mL 3% thioglycolate (Sigma-Aldrich) in the healthy mice and harvested after 3 h post-injection by peritoneal lavages with ice-cold PBS. Then, the samples were centrifuged at 1,800 × g, at 4 °C for 5 min to separate the cells and resuspended in RPMI 1640 Medium supplemented with 10% FBS (Gibco Life Technologies, Paisley, UK). Then, neutrophils at 2 × 106 cells/well in 24-well-plates were incubated with media alone or media with lipopolysaccharide (LPS) (Escherichia coli 026: B6; Sigma-Aldrich) at a concentration of 250 ng/mL or C. albicans at 1:10 (the ratio between the numbers of the cells and the organisms) or LPS plus C. albicans under 5% carbon dioxide (CO2) at 37 °C before collection of supernatant and cells at different time-points. The supernatants were used for cytokine measurement by ELISA (Invitrogen) following the manufacturer’s protocol, and the expression of several genes was determined using RT-PCR as mentioned above.

Statistical analysis

GraphPad Prism version 10.4.1 (532; www.graphpad.com) was used to analyze the data and demonstrated in mean ± SE (standard error) with the one-way analysis of variance (ANOVA) with Tukey’s comparison test. Survival analysis was evaluated by the log-rank test. A p-value less than 0.05 was considered a significant value.

Ethical consideration

The animal protocols in this study were approved by Institutional Animal Care and Use Committee of the Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand, in accordance with the US National Institutes of Health (NIH) guidelines (the ethical approval number: SST 020/2566).

Results

Candida administration worsened lung hemorrhage in pristane-administered mice

Due to the presence of Candida in the lower respiratory tract, either with or without lung diseases16, Candida administration was tested in vivo, as illustrated in Fig. 1A. With intrathecal Candida administration, pristane plus Candida (Pris + Can) demonstrated a more severe model than pristane alone, as indicated by survival analysis and weight retardation (Fig. 1B, C) without the differences in lupus manifestations (proteinuria, serum creatinine, and anti-dsDNA), gut permeability defect (FITC-dextran assay), and serum cytokines (TNF-α, IL-6, and IL-10) (Fig. 1D–J). Indeed, the lung hemorrhage score of Pris + Can was also higher than Pristane alone, while the expression of the fungal genes as measured by RT-PCR using the ITS primer was presented only in Candida-administered mice (Fig. 2A–C). Notably, the fungi could not be identified by culture (3 days after the last dose of Candida administration) in either Candida administration alone or Pris + Can mice (data not shown), implying a low level of Candida in lung tissue and the lack of fungal proliferation after the administration. Pristane rapidly elevated serum pro-inflammatory cytokines (TNF-α and IL-6) (Fig. 1H, I) as early as 1 week, followed by an initiation of lupus manifestation (elevated anti-dsDNA) at 4 weeks (Fig. 1F), and induced lung hemorrhage (Fig. 2A, B) and mortality (Fig. 1B) at 5 weeks after injection, supporting systemic inflammation in lupus-induced DAH6. The natural history of the pris + Can model was similar to pristane alone (Fig. 1B–J); however, lung hemorrhage (Fig. 2A, B), but not serum cytokines (Fig. 1H, I), was more prominent than pristane alone, suggesting an impact of Candida-induced local lung inflammation.

Fig. 2
figure 2

Characteristics of control, pristane injection, Candida administration (Candida), and pristane plus Candida (Pris + Can) as indicated by the representative lung histological pictures as stained by Hematoxylin and Eosin color (A), lung hemorrhage score (B), lung fungal abundance (C), and lung microbiome analysis, including alpha diversity (Chao-1 and Shannon index) with beta diversity (the Principle Coordinate Analysis (PCoA) based on Bray Curtis dissimilarity) (D–F) are demonstrated (n = 5–7/group for A–C and n = 3/group for microbiome analysis).

Full size image

Lung microbiota alteration in pristane with versus without Candida

Because of the well-known interaction between fungi and bacteria, bacterial dysbiosis from Candida administration might partly be responsible for the more severe DAH. As such, Candida administration altered alpha diversity, as indicated by the elevation in Chao-1 richness estimation (the number of species in the defined ecosystem), Shannon evenness estimation (the similarity of the abundance of the species), and beta diversity, as measured by the principle Coordinate Analysis (PCoA) based on Bray Curtis dissimilarity metrics (a simplified method to demonstrate the similarity among groups using the 2-dimensional distances from the axis) (Fig. 2D–F). However, there were some differences in bacterial abundance among groups (Fig. 3A–F) as Candida administration (Pris + Can and Candida alone) reduced bacteria in the phylum Proteobacteria (mostly Gram-negative aerobes) (Fig. 3G), especially in the order Burkholderiales (Fig. 3C). Meanwhile, the phylum Bacteroidota (Bacteroidetes; mostly Gram-negative anaerobic bacteria) in Candida alone was most prominent, without the difference in Firmicutes (Fig. 3G). The lowest abundance of Cupriavidus spp. (Gram-negative, non-glucose-fermenting aerobes in the Burkholderia family and Proteobacteria phylum categorized as an opportunistic pathogen)37 was presented in Candida alone, while the highest abundance of Lactobacillus spp. was identified in Pris + Can (Fig. 3H). There was no difference in Lachnospiraceae among all groups (Fig. 3H). These data support the impacts of fungi on bacterial population38,39,40. Surprisingly, there was no difference in the abundance of bacteria in the lung between control group and intraperitoneal pristane-injected mice (Fig. 3A–H). These data support a limited influence of pristane on the lung bacteria, despite the positive pristane-induced lung hemorrhage (Fig. 2A, B).

Fig. 3
figure 3

Microbiome analysis from the lung of control, pristane injection, Candida administration (Candida), and pristane plus Candida (Pris + Can) as indicated by bacterial abundance in different levels (phylum, class, order, family, genus, and the average of genus) (A–F) with the selected graph presentation of only the high abundance groups in phylum (G) and genus (H) are demonstrated (n = 3/ group).

Full size image

Non-different systemic inflammation with more prominent lung inflammation in pristane plus Candida compared with pristane alone

Although systemic inflammation (serum cytokines) was not different among groups (Fig. 1H–J), Candida administration might activate some immune cells in the spleen. Then, flow cytometry analysis of splenocytes was used to determine systemic immune responses (Fig. 4A–L). As such, the elevation of plasma cells, naïve Th cells, total Th cells, central memory CD8 T cells, and IgM-producing B cells was demonstrated in all experimental groups compared with control (Fig. 4A–L). Meanwhile, central memory cells (total T cells and Th cells), effective memory T cells, inducible T cells, activated B cells, and dendritic cells were elevated in pristane and Pris + Can, but not Candida alone, suggesting the responses against pristane-induced lupus (Fig. 4A–L). There was no difference between Pris + Can and Pristane alone, supporting a limited systemic response against intratracheal Candida administration (Fig. 4A–L). In parallel, Pris + Can demonstrated the most prominent lung inflammation as indicated by the elevation of cytokines (TNF-α, IL-6, and IL-10) in the lung tissue (Fig. 4M–O), correlating with more severe lung hemorrhage in Pris + Can (Fig. 2A, B). The non-different cytokine levels in the lung tissue between control and Candida-administered mice (Fig. 4M–O) suggest a less potent inflammatory inducer of Candida compared with bacteria41,42. Furthermore, the peritoneal tissue-resident macrophages of all experimental groups were compared (Fig. 5) to explore the impacts of intratracheal Candida on pristane-induced macrophages. Accordingly, intraperitoneal pristane injection induced M1 pro-inflammatory macrophages, as indicated by the elevation of iNOS and IL-1β genes (PCR) together with the enhanced CD80, and CD86 on the cell surface (flow cytometry) without altering M2 anti-inflammatory markers, including Fizz-1, Arg-1, TGF-β, CD206, and CD163 (Fig. 5A–I). Also, the pristane-induced macrophages prominently expressed pro-inflammatory cytokine genes (TNF-α and IL-6), partly through nuclear factor kappa B (NF-κB) gene activation, without the differences in IL-10, TLR-4, and Dectin-1 genes (Fig. 5J–O). However, the intratracheal Candida administration did not alter peritoneal macrophages, neither in the control nor the pristane groups (Fig. 5A–O), implying a lower systemic response against the intratracheal Candida.

Fig. 4
figure 4

Flow cytometry analysis of the spleens from control, pristane injection, Candida administration (Candida), and pristane plus Candida (Pris + Can) as indicated by the abundance of plasma cells (CD138 and B220 positive) (A), naïve Th cells (CD62L positive with CD44 negative) (B), total control memory T cells (CD3, CD44, and CD62L positive) (C), total Th cell (CD4 positive CD8 negative) (D), central memory Th cell (CD3, CD4, and CD62L positive) (E), effective memory Th cells (CD3 and CD4 positive with CD62L negative) (F), inducible T cell costimulatory (CD4 and ICOS positive) (G), central memory CD8 T cells (CD8 and CD62L positive) (H), adaptive B cells (GL7 and B220 positive) (I), IgM-producing B cell (IgM and B220 positive) (J), dendritic cells (CD80 and CD11c positive) (K), and activated dendritic cells (ICOS and CD11c positive) (L) are demonstrated (n = 5/ group). Lung tissue cytokines (TNF-a, IL-6, and IL-10) (M–O) are also demonstrated (n = 5/ group).

Full size image
Fig. 5
figure 5

Characteristics of tissue-resident macrophages (peritoneal macrophages) of mice from control, pristane injection, Candida administration (Candida), and pristane plus Candida (Pris + Can) as indicated by the biomarkers of M1 pro-inflammatory macrophages, including the expression of iNOS and IL-1β (using PCR) with CD80 and CD86 (flow cytometry), (A–D), M2 anti-inflammatory macrophages using Arg-1, Fizz-1, and TGF-β (PCR) with CD206 and CD163 (flow cytometry), (E–I), and expression of inflammatory genes (TNF-α, IL-6, IL-10, TLR-4, Dectin-1, and NF-κB) (J–O) are demonstrated (n = 5 /group).

Full size image

Proinflammatory impact on neutrophils with the presence of Gram-negative bacteria and fungi through LPS and beta-glucan

Because (i) the presence of Proteobacteria and Bacteriodota (the prominent Gram-negative bacteria in the lung) with Candida might induce more severe inflammation than bacteria alone, and (ii) neutrophils are important immune cells in DAH pathogenesis36, LPS (a Gram-negative bacterial molecule) and beta-glucan (BG; a major cell wall component of fungi) were used to stimulate neutrophils in vitro43,44,45. As such, BG plus LPS induced more severe neutrophil responses than LPS alone, as indicated by the higher expression (at least one time point) of inflammatory cytokine genes (TNF-α, IL-6, IL-1β, and IL-10), Dectin-1 (a BG receptor gene), and NF-κB (a transcriptional factor gene) (Fig. 6A–F). Meanwhile, there was a non-different down-regulation of TLR-4 gene in both LPS plus BG and LPS alone (Fig. 6E), supporting TLR-4 activation by LPS46. These data indicate a synergistic pro-inflammatory effect of fungi and bacteria compared with bacteria alone on the activation of pulmonary neutrophils.

Fig. 6
figure 6

Characteristics of thioglycolate-induced neutrophils after the activation by lipopolysaccharide (LPS) or Candida preparation (see method) (Candida) or combined LPS plus Candida (LPS + Candida) or untreated control as indicated by expression of cytokine genes (TNF-α, IL-6, IL-1β, and IL-10) (A–D), TLR-4 gene (a LPS receptor) (E), Dectin-1 gene (a receptor of beta-glucan, an important component of Candida cell wall) (F), and NF-κB gene (an important transcriptional factor) (G) are demonstrated (results from isolated triplicated experiments).

Full size image

Discussion

Despite the well-known importance of adaptive immunity to lupus-induced DAH, innate immunity against micro-organisms is also important, as the intratracheal Candida administration worsened pristane-induced lung hemorrhage without inducing systemic inflammation. Chronic peritoneal inflammation (type I interferon and circulating immune complex)-induced lung hemorrhage in the pristane model35 was supported by M1 pro-inflammatory peritoneal macrophages that increased systemic inflammation and altered immune cells in the spleens. Then, intratracheal Candida induce only a subtle systemic impact only in the spleen (elevation in plasma cells, Th cells, and IgM-producing B cells), supporting a cross-talk between the lung and other organs47. Meanwhile, intrathecal Candida administration enhanced lung inflammation (lung cytokines) and lung hemorrhage in the pristane model, supporting a prominent local lung effect with very little systemic impact of lung Candida.

The continuous contact between the lung and the environment (inhalation of organisms) and the ecological niches of the respiratory system (nasal cavity, nasopharynx, oropharynx, trachea, and lungs) induces the balance of lung microbiota48. Among several causes, Candida albicans (the most common fungi in the human oropharynx) frequently induce microbiota imbalance (asymptomatic Candida colonization in the sputum49, saliva50, and the lung15). Although the abundance of intratracheal Candida here was not enough to be detected by culture at 3 days after the last dose of Candida administration, the fungi were detected by PCR using ITS primers (Fig. 2C) which alter bacterial microbiota in the lung (increased richness of bacterial species). Here, Candida administration, in either regular mice or pristane mice, reduced the abundance of Proteobacteria and elevated Bacteroidota (in Candida alone). Indeed, C. albicans may stimulate Bacteroides growth via aerobic respiration and antioxidant production51. Subsequently, the enhanced growth of some bacteria might antagonistically reduce the growth of other bacteria. Surprisingly, Candida administration increased the abundance of Lactobacilli in the Pristane plus Candida group but not in the Candida or Pristane alone, partly due to the fungi-induced anaerobic microenvironment (the use of oxygen by fungi) plus an adaptation against lung inflammation (as mentioned in cystic fibrosis)52,53. However, an elevation of Lactobacilli (immunomodulation probiotics) in pristane plus Candida mice could not neutralize lung hemorrhage here.

Although lung Candida did not obviously elevate pathogens, the presence of LPS from proteobacteria (the high-abundance Gram-negative bacteria presented in all experimental groups) and fungal molecules (Candida preparation; see method) synergistically activated immune cells, especially neutrophils (Fig. 6)54,55,56,57. Indeed, LPS plus beta-glucan (BG; a major molecule of the Candida cell wall) activates Toll-like receptor 4 (TLR-4) and Dectin-1, the main receptors for LPS and BG, respectively, more prominently than LPS alone58,59. Although BG alone is not a potent immune activator42, BG synergistically enhances the action of LPS, a potent innate immune stimulator, partly through increased cell energy responses60. For a clinical perspective, lupus-DAH might be more severe in patients with the presence of Candida in the lung which might be monitored through Candida colonization (culture method) or nucleic acid abundance (PCR using ITS primer) from the sputum or saliva. Also, the increased Candida colonization after the use of immunosuppressive drugs in lupus should be concerned. Additionally, the blockage of neutrophils or LPS responses (TLR-4 inhibitor) in severe lupus-induced DAH might be clinically helpful.

Several limitations should be mentioned. First, the sample size was small due to the minimization of the number of animals according to our animal study protocol. Second, the duration of the study might not be sufficient to observe the long-term effects of Candida administration on lung dysbiosis. Third, limited mouse parameters were performed; for example, only 3 inflammatory cytokines were tested. The experiments with more parameters will produce more information. Fourth, the in-depth mechanisms of the worsening of lung hemorrhage by Candida administration in the pristane model were not performed. The additive neutrophil responses against LPS plus BG over LPS alone might be clearer through gene knockout experiments. Further studies to elucidate the underlying pathways are needed.

In conclusion, a proof of concept on the enhanced lung hemorrhage from the presence of Candida in the lung, partly through the neutrophil responses against bacteria plus fungi was demonstrated. More studies are warranted.