Comparative miRNA transcriptomics of macaques and mice reveals MYOC is an inhibitor for Cryptococcus neoformans invasion into the brain

ABSTRACT Cryptococcal meningoencephalitis (CM) is emerging as an infection in HIV/AIDS patients shifted from primarily ARTnaive to ART-experienced individuals, as well as patients with COVID-19 and immunocompetent hosts. This fungal infection is mainly caused by the opportunistic human pathogen Cryptococcus neoformans. Brain or central nervous system (CNS) dissemination is the deadliest process for this disease; however, mechanisms underlying this process have yet to be elucidated. Moreover, illustrations of clinically relevant responses in cryptococcosis are currently limited due to the low availability of clinical samples. In this study, to explore the clinically relevant responses during C. neoformans infection, macaque and mouse infection models were employed and miRNA-mRNA transcriptomes were performed and combined, which revealed cytoskeleton, a major feature of HIV/AIDS patients, was a centric pathway regulated in both infection models. Notably, assays of clinical immune cells confirmed an enhanced macrophage “Trojan Horse” in patients with HIV/AIDS, which could be shut down by cytoskeleton inhibitors. Furthermore, myocilin, encoded by MYOC, was found to be a novel enhancer for the macrophage “Trojan Horse,” and an enhanced fungal burden was achieved in the brains of MYOC-transgenic mice. Taken together, the findings from this study reveal fundamental roles of the cytoskeleton and MYOC in fungal CNS dissemination, which not only helps to understand the high prevalence of CM in HIV/AIDS but also facilitates the development of novel therapeutics for meningoencephalitis caused by C. neoformans and other pathogenic microorganisms.


Introduction
Fungal invasive diseases are an increasing threat to global public health and lead to more than one million deaths annually, mainly caused by species of Candida albicans, Aspergillus fumigates and Cryptococcus neoformans. Cryptococcal meningoencephalitis (CM) is the leading disease among fungal meningoencephalitis and causes approximately 181,000 deaths each year, with mortality rates of 100% if untreated [1]. Recent studies showed that the prevalence and mortality of CM have not decreased and consequently appeals for action on this disease have been raised [2,3]. Additionally, CM is a major risk for patients with HIV/AIDS, with a prevalence of 77%-90% and leading to approximately 15% of deaths of patients with HIV/AIDS annually [4]. In the era of ART (antiretroviral therapy), CM is also an emerging infectious disease, shifting from primarily ARTnaive individuals to more than 50% of cases being in ART-treated patients [2]. Seriously, the morbidity of cryptococcosis in immunocompetent individuals is increasing rapidly in China, Australia, Canada, and other countries and regions [4][5][6][7][8]. Importantly, some patients with COVID-19 suffered from secondary Cryptococcus infections, suggesting cryptococcosis is an important issue in the post-COVID-19 era [9][10][11][12]. Fungal CNS dissemination, which is vital for both fungal colonization and fungal clearance, is the lethal part of the infection process, but the mechanisms of CNS dissemination are lacking from in vivo or clinical studies.
To date, massive research has focused on CNS invasion during fungal meningitis and has achieved numerous milestones, including that cryptococci interact with brain epithelial cells directly and indirectly by microscopy or electron microscopy [13][14][15][16]. These studies were mainly performed on cell lines, mice, zebra fish or rabbits [17]; however, clinically relevant evidence is limited due to the unavailability of clinical samples. Cryptococcal meningoencephalitis is predominantly prevalent among immunocompromised individuals, especially those with HIV/AIDS [18][19][20]. Thus, elucidating mechanisms from clinical patients with HIV/AIDS is the most direct approach to understand Cryptococcus meningoencephalitis. A dysfunctional cytoskeleton of immune cells is a major pathological feature in patients with HIV/AIDS [21]. Meanwhile, previous studies indicated that C. neoformans infections also disturbed cytoskeleton pathways in mice and cell lines [22,23]. However, there is currently no evidence demonstrating mechanisms between the dampened cytoskeleton and cryptococci brain dissemination. Identification of relationships between the cytoskeleton and fungal infections may reveal novel mechanisms of fungal pathogenesis and provide targets for drug development in mycosis.
Recently, several studies have demonstrated functions of non-coding RNAs during C. neoformans infections in vitro, but there is limited information in vivo [32][33][34][35].
In this study, to mimic human responses, macaques were employed and an in vivo miRNA-mRNA network was constructed. The cytoskeleton pathway was revealed as the core pathway regulated by C. neoformans in both macaques and mice. Moreover, clinical immune cell assays confirmed an enhanced macrophage "Trojan Horse" in patients with HIV/AIDS, and intervention of the cytoskeleton pathway disrupted the "Trojan Horse." Furthermore, the cytoskeleton-associated gene, MYOC, was identified as an important factor for "Trojan Horse" by THP-1 cells and transgenic mice. Collectively, these findings demonstrate global responses at the miRNA-mRNA regulatory level, reveal novel modulators for fungal invasion, and may facilitate the development of novel therapeutics for fungal infectious diseases.

Animal infection
Macaques and mice were purchased from Grandforest Co. (Guangxi, China) and Changsheng Biotech (China), respectively, and infections were performed as previously described [22]. Briefly, six female macaques were divided into two groups. Macaques were anesthetized by intraperitoneal injection (IP) with ketamine (10 mg·kg −1 ) and then infected via intratracheal injection with 10 8 cells·mL −1 C. neoformans H99. Controls were injected with the same volume of PBS. Mice were anesthetized and infected intranasally with 10 5 fungal cells. Macaques and mice were monitored for signs of infection and humanely killed at day 7 or 14 post-infection, or used for survival rate determination.

miRNA sequencing and analysis
Total RNA of lung tissues was isolated using TRIzol and assessments of RNA were performed using a NanoDrop 8000. Small RNA-Seq libraries were prepared by using TruSeq® Small RNA Library Prep Kit according to the manufacturer's protocol. Construction and sequencing of miRNA libraries were entrusted to Shanghai Personal Biotechnology Co., Ltd. (China), and then single-end sequencing was conducted using an Illumina NextSeq 500 platform. Raw data were obtained, and clean and unique reads were mapped to corresponding genomes by Bowtie. The expression of miRNAs was identified by using quantifier.pl in Mirdeep2 based on miRBase21. Differentially expressed miRNAs were calculated by DESeq.

Quantitative PCR of miRNAs and mRNAs
For miRNAs, total miRNAs were isolated by miRNeasy Mini Kit (QIAGEN) according to the manufacturer's instructions. First-strand cDNA was synthesized using random oligonucleotides and a miRcuRY LNA Universal RT microRNA PCR Universal cDNA Synthesis Kit II (EXIQON, USA) and U6 was employed as an internal reference. Primers (Table S5) were designed by using miRprimer and the optimum primer pairs were selected [36]. RT-qPCR was performed by using miRcuRY LNA Universal RT microRNA PCR Exilent SYBR master mix (EXIQON) and a StepOne Plus system. For mRNAs, total RNAs were isolated and cDNAs were synthesized. RT-qPCR was performed by using GoTaq®qPCR Master Mix (Promega) and the Roche LightCycler® 480 system.

Histopathology, colony forming units (CFU) and survival rates assessments
Lung tissues from macaques and mice were collected as described previously [22]. For histopathology analyses, tissue samples were fixed with paraformaldehyde, frozen, and processed using a cryostat microtome (CM1850, Leica). Tissue sections of 10 µm thickness were stained with mucicarmine or haematoxylin/eosin. For CFU determination, homogenized lung and brain tissues were diluted, spread on YPD plates and incubated at 30°C for two days, then colonies were counted and calculated. For determination of survival rates, C57BL/6 mice were divided randomly. Bodyweight was examined post infections. A decrease of 15% of initial body weight was identified as the physiological endpoint.

Isolation of human monocytes and induction of monocyte-derived macrophages (MDMs)
Peripheral blood mononuclear cells (PBMCs) from patients with HIV and healthy individuals were isolated by Ficoll-Paque PLUS density gradient media (Cytiva). Blood samples were centrifuged in a swing bucket rotor at 400g for 30 min at 25°C with acceleration set at 5 and break at 0, followed by purification in cold PBS twice and centrifugation at 300g for 10 min at 4°C with acceleration and break set at 5. The isolated PBMC were seeded into 48-well plates at 1 × 10 6 cells per well in RPMI-1640 media without FBS for 1 h to allow adherence of monocytes, then the medium was exchanged for fresh RPMI-1640 supplemented with 10% FBS, 1% penicillin/streptomycin and 50 ng·mL −1 M-CSF. The medium was further renewed on days 3 and 6 during differentiation of the monocytes, and the MDMs were ready to use on day 7.

Phagocytosis effectivity, killing and transmigration assessments
MDMs or TDMs were seeded on 48-well plates at 1 × 10 5 cells per well 24 h before fungi interaction. GFPexpressing strains of C. neoformans H99 were incubated overnight at 30°C, washed, and opsonized by 18B7 (1mg·L −1 ) at room temperature for 30 min. Fungal cells and macrophages were incubated overnight in a CO 2 incubator at a multiplicity of infection (MOI) of 1:10. The cells were then washed five times with PBS, digested with trypsin, and phagocytosis effectivity was detected by flow cytometry. For killing assays, the supernatant was collected after incubation and cells were washed and collected. Total lysates and supernatants were mixed, diluted, spread on YPD agar plates and incubated for 48 h at 30°C, then colonies were counted and calculated. Migration of THP-1 cells and MDMs were performed by trans-well assay.

Microscopy
Phagocytes of TDMs contained with C. neoformans were washed, and then fixed using ice-cold 4% paraformaldehyde (PFA) in 0.1 M PBS for 30 min room temperature, and stained for F-actin (phalliodin-Alexa647, A22287, Invitrogen) for 30 min at room temperature. Imaging was conducted using the laser confocal microscope ZEISS LSM 980 with Airyscan 2.

Construction of MYOC-transgenic mice
MYOC-transgenic mice were produced by Beijing View Solid Biotechnology, China. The plasmid pCAG-MYOC was linearized by BstEII restriction enzyme (NEB) digestion, purified, and injected into zygotes of C57BL/6 mice in M2 media (Millipore) using a FemtoJet micromanipulator (Eppendorf, Germany). Microinjected zygotes were transferred into pseudo-pregnant female mice. All mice were maintained in a specific pathogen-free facility. Genotype identification was performed by PCR and sequencing from 2-week-old newborn mice (Table S5). Transgenic mice were mated with wild-type C57BL/ 6 mice to obtain heterozygous mice and colony expansion.

Bioinformatics analysis
Targeted genes of miRNAs were predicted using miR-Walk 3.0 [37]. Regulatory network miRNA-mRNA was constructed by Cytoscape. Gene ontology and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed using R version 4.1.2, clus-terProfiler v4.2.0, org.Hs.eg.db version 3.14.0 and org.Mm.eg.db version 3.14.0 packages and were plotted by ggplot2 version 3.3.5. KEGG of miRNAs was performed by using the web-based application miEAA [38] and were plotted by ggplot2. A heatmap of expression was generated by Pheatmap version 1.0.12. Homology analysis was performed based on the miRbase search engine database (https://www. mirbase.org/search.shtml).

Statistics and reproducibility
All experiments were performed at least biological triplicates to ensure reproducibility. Statistics of phagocytosis effectivity, RT-qPCR and CFU determinations were calculated with an unpaired or paired student t-test using GraphPad Prism 9.0. All results were shown as mean ± standard error of the mean (SEM). When the p-value was less than 0.05, statistical significance was recognized.

MicroRNA transcriptomics in macaque and mouse during cryptococcal pneumonia
Transcriptional responses during C. neoformans infections in macaque and mouse were unveiled in a previous study by our group and several regulators and pathways were identified [22]. To investigate global responses at post-transcriptional level, miRNA transcriptomes were performed in the current study using lung tissues isolated from macaque and mouse infection models ( Figure 1). Furthermore, a miRNA-mRNA network was constructed to reveal a comprehensive and clinically relevant responses in cryptococcosis.
Pathologies of macaque and mouse lung tissues were confirmed by histopathology observations using mucicarmine and haematoxylin/eosin staining to show the capsular structure of C. neoformans (Figure 2(A,B)). Total RNAs of macaque and mouse lung tissues were extracted and miRNA transcriptomes were determined. Total and unique reads from omics were calculated, with a number of more than 10 7 of total read and 10 6 of unique read ( Figure S1A). Reads around 22 nucleotides were the most abundant (Figure S1 B, C). The clean read files were used to map the corresponding genome based on species specificity in miRBase21. However, macaque only has a few miRNA annotations available, therefore, miRNA annotations of humans were employed for macaque samples. As a result, 1038 and 1166 miRNAs were identified in lung tissues from macaques and mice, respectively (Table S1). Principal component analyses (PCA) were examined by using the read number of miRNAs, and both infected and uninfected samples were reproducible ( Figure 2C, D). Differentially expressed miRNAs were determined using DESeq. Heatmaps of the differentially expressed items revealed the presence of 32 miRNAs (4 downregulated and 28 upregulated) in macaques and 29 miRNAs (5 downregulated and 24 upregulated) in mice following C. neoformans infection, shown as two clusters by column ( Figure 2E, F).

Core responses of the host-derived from miRNA-mRNA integrative analyses
To mine and mimic responses in human cryptococcosis, homology analyses were performed ( Figure 3A). Eight miRNAs were co-regulated in macaque and mouse during C. neoformans infections. Information about the eight miRNAs were displayed, including IDs, foldchanges and homologous e-values, and RT-qPCRs for the eight miRNAs were performed in mice and were consistent with the miRNA-Seq data ( Figure 3B, C, D).
Targets of the eight co-regulated miRNAs were predicted, then mapped to the RNA-Seq data, and omics of miRNA and mRNA were integrated ( Figure S2). The core regulatory target mRNAs, co-regulated by the eight co-regulated miRNAs, were selected for a mini miRNA-mRNA network. This network identified a total of 223 target mRNAs, including OCSTAMP, DC-STAMP, IL17A TNF and LIF, which demonstrated anti-microbial activities ( Figure  3E). KEGG and GO analyses were performed on the eight co-regulated miRNAs and targets (Figure 4, Figure S3, Tables S2 and S3). KEGG pathways associated with the immune system, lung diseases, and infectious diseases were significantly enriched at both miRNA and mRNA levels. These pathways included the TNF signalling pathway, Cytokine-cytokine receptor interaction, Osteoclast differentiation, small cell lung cancer, non-small cell lung cancer and Influenza A ( Figure S3). GO analyses identified terms involved in cytokine activity, histone, immune cells, monocyte, myeloid leukocyte, and cell cycle were significantly enriched during C. neoformans infections (Figure 4). Moreover, actin binding and microtubule and their associated complexes constitute the cytoskeleton system, were highly enriched in GO analyses, including cellular component, molecular function, and biological process. These findings indicated a potential function of the cytoskeleton in cryptococcosis.
The macrophage "Trojan Horse" was enhanced in patients with HIV/AIDS and could be dampened by a cytoskeleton pathway inhibitor Dysfunctional cytoskeleton of immune cells is one of the major pathological features of patients with HIV/ AIDS, who are the predominate population for CM. Based on PSM (Propensity Score Matching, PSM), 100 HIV patients and 200 healthy individuals were compared, which revealed that the number and percentage of monocytes were significantly enlarged in patients with HIV ( Figure 5B). To reveal functions of the host cytoskeleton during the battle between host and C. neoformans, phagocytosis and transmigration of MDMs from nine patients with HIV/AIDS and eight healthy volunteers were compared ( Figure 5A). Phagocytosis effectivity was enhanced in patients with HIV/AIDS ( Figure 5C), and furthermore, the lifted MFI (Mean Fluorescent Intensity, MFI) indicated more fungal cells were phagocytosed or underwent more intracellular proliferation in patients with HIV/AIDS. There was no change in transmigration of MDMs between patients with HIV and healthy volunteers (data not shown), but MDM transmigration was positively correlated with phagocytosis capacity ( Figure 5D).
To confirm the association of the phenomenon of C. neoformans infection in patients with HIV/AIDS with the cytoskeleton, R10015a cytoskeleton pathway inhibitorwas employed, who targets LIM Kinase (LIMK) in the cytoskeleton pathway [39]. A growth curve was generated to select the appropriate inhibitory dose of R10015, which was 14.815 µM ( Figure S4A). TDMs and human MDMs were pretreated with R10015 for 2 h, washed and then incubated with opsonized C. neoformans overnight. R10015 inhibited the phagocytosis capacity and migration of macrophages derived from both THP-1 cell lines and primary human monocytes (Figure 5E, F and G, H). However, R10015 did not affect the killing in TDMs and MDMs ( Figure 5I, J). These results confirmed the vital roles of the cytoskeleton in the macrophage "Trojan Horse." MYOC is an inhibitor for cryptococci brain dissemination by modulating macrophage "Trojan Horse" To identify functions of the cytoskeleton during C. neoformans infections, genes associated with the cytoskeleton were screened, and myocilin was selected. Myocilin is a tubulin-binding protein encoded by the gene MYOC, which was one of the centric modulators in miRNA-mRNA regulatory network and was downregulated significantly during C. neoformans infections in both macaque and Three biological replicates were performed. An unpaired student's t-test was performed to assess statistical significance; *p < 0.05, ** p < 0.01 and ***p < 0.005. E. miRNA-mRNA core regulatory network. Differentially expressed target genes of eight co-regulated miRNAs are shown. The whole network is presented in Figure S3. mouse ( Figure 6A, B, Table S4). In addition, expression of MYOC was elevated in patients with HIV/AIDS ( Figure 6C). Previous studies demonstrated that the protein was involved with cell migration and adhesion [40].
To evaluate functions of MYOC in vivo, a MYOCtransgenic mouse was generated ( Figure 7A) and confirmed by PCR ( Figure 7B). Six-eight weeks old mice were used for CFU assessments and survival rate determination. Compared to wild-type mice from the same cage, CFUs in brains were enhanced in MYOC-transgenic mice in both male and female groups ( Figure 7C). CFUs in lungs were decreased in female mice, while no changes in the male group ( Figure 7D). Consistent with CFU assays, MYOCtransgenic mice showed reduced survival times compared with wildtype mice in survival rate tests ( Figure  7E, Figure S4F).

Discussion
CM is an emerging disease with high mortality, even under current ART conditions, and warrants more attentions in the post-COVID-19 era [1][2][3][9][10][11][12]. Brain dissemination is the lethal infection process. However, clinically relevant mechanisms of brain dissemination in CM are currently limited [13][14][15][16][17]. In this study, the landscapes at transcriptional and post-transcriptional levels were determined in mouse and macaque infection models, which were employed to mimic and unveil responses at miRNA-mRNA regulatory levels in humans. The mouse is the most used animal model for CM in previous studies, but there are marked variations between mice and humans, such as in the process of immune cells maturation. Therefore, data in the current study from M. fascicularis, which shows 92.83% sequence identity in humans, was hypothesized be more clinically relevant and may serve as a database for cryptococcosis or mycosis. To investigate core responses during C. neoformans infections, GO and KEGG analyses of combined miRNA-mRNA data were performed.
Eight key miRNAs were identified, including miR-146a, miR-223 and miR-155, which were induced in monocytes by co-cultured with C. neoformans in vitro [32]. The functions of unique miRNAs were not explored in the current study, instead, the cytoskeleton pathway was characterized as a core regulatory modulator based on enrichment analyses of miRNAs and their targets.
Our data identified the fundamental roles of the cytoskeleton during macrophage "Trojan Horse" formation, and furthermore highlighted potential targets for novel "orphan drug" development. Studies have indicated relationships between cytoskeleton and fungal infections [23,[41][42][43][44]. However, the mechanisms of interaction have yet to be elucidated and there is a lack of in vivo studies confirming the relationship. Important roles of "Trojan horse" by macrophages during fungal CNS invasion have previously been reported, whose processes were cytoskeleton-dependent [45]. The current study illustrated the basic functions of cytoskeleton dynamics on capacities of phagocytosis and migration of macrophages, which can be blocked by cytoskeleton inhibitors. However, small-molecule drugs, such as vanadate, cytochalasin D, Y27632 and R10015 are often toxic to humans [23,39,46]. This toxicity involves fundamental roles of innate immunity and acquired immunity of macrophages, NK cells and T cells, which also play important roles during fungal, bacterial and viral infections. Overcoming toxicity of cytoskeletal inhibitors is the main barrier to clinical application.
Further analyses revealed that a dysfunctional cytoskeleton may account for the high prevalence of CM in patients with HIV/AIDS. Studies indicated a similar variation of cytoskeleton pathway during HIV infection and C. neoformans or A. fumigates infections [23,42,43,47,48]. Indeed, C. neoformans is an environmental yeast that is globally ubiquitous and easily accessible for all individuals [49]. However, why do these happen much more in HIV/AIDS? Previous studies suggested the decreased number of CD4 T cells was a high-risk factor; however, this is contradictory to the high prevalence of high-level CD4 T cells who are ART experienced [50][51][52]. Another hypothesis is the increased exposure to a C. neoformans environment, but this is inconsistent with the low prevalence observed in other people in the same environment [53]. Data from the current study suggest the dampened cytoskeleton structure may be the reason for the high prevalence of CM in patients with HIV. In patients with HIV/AIDS, numerous studies have proved the cytoskeleton was dampened in immune cells of PBMCs [39,48,54], which contributes to fungal invasion of the CNS by the macrophage "Trojan Horse." Furthermore, vomocytosis was recently found to be enhanced by HIV infection in macrophages [55], which process was also regulated by the cytoskeleton.
In addition, the current study identified MYOC as a novel modulator for a fungal invasion via regulation on macrophages. Our data showed repression of MYOC in C. neoformans infections and induction in HIV infection. Myocilin co-localizes with microtubules, the endoplasmic reticulum (ER), and the Golgi apparatus [56,57], and overexpression of MYOC induces a loss of actin stress fibres [58]. Furthermore, myocilin promotes cell migration and phagocytic activities of human trabecular meshwork cells [40,59,60]. In the present work, MYOC overexpressed THP-1 cell line was generated, and expression of genes associated with actin polymerization and phagocytosis indicated an enhanced phagocytosis and killing capacity in MYOC-overexpressed THP-1 cells. Moreover, overexpression of MYOC in TDMs elevated phagocytic activities and migration, which induced more intracellular cryptococci and enhanced effects of the macrophage Genomic DNA was isolated and then PCR and electrophoresis were performed. Target PCR product was 498 bp. C, D. CFU assessments of brain (C) and lung (D) tissues in MYOC-transgenic mice. Seven MYOC-transgenic mice (three male and four female) and eight wild-type mice (four male and four female) from the same cages were used for CFU assays. An unpaired Student's t-test was performed to assess statistical significance; * p < 0.05. E. Survival tests of MYOC-transgenic mice. Eleven MYOC-transgenic mice (five male and six female) and nine wild-type mice (three male and six female) from the same cages were used for survival assays. Log-rank (Mantel-Cox) test was employed for statistical analysis; * p < 0.05. "Trojan Horse." This indicated that MYOC may be the effector for cryptococcoses secondary to HIV/AIDS.
In conclusion, this study described global miRNA-mRNA regulatory responses during C. neoformans infections in primate and rodent animal models, which serves as a clinically relevant database for fundamental and clinical research. We highlight the importance of MYOC and cytoskeleton pathways during Cryptococcus meningoencephalitis and underscores their critical functions in the formation of "Trojan Horse" (Figure 8). This study demonstrates the critical roles of the cytoskeleton in fungal CNS invasion, reveals a potential direct explanation for the high prevalence of CM in patients with HIV/AIDS, and may facilitate the development of novel anti-fungal drugs.