Evidence for Immune Activation in Podoconiosis Pathogenesis PDF

Article https://doi.org/10.1038/s41467-024-46347-z Evidence for immune activation in pathogenesis of the HLA class II associated...

Article https://doi.org/10.1038/s41467-024-46347-z Evidence for immune activation in pathogenesis of the HLA class II associated disease, podoconiosis Received: 1 December 2023 Mikias Negash 1,2,3 , Menberework Chanyalew2, Tigist Girma2, Fekadu Alemu 2, Diana Alcantara1, Ben Towler 4, Gail Davey1,5, Accepted: 23 February 2024 Rosemary J. Boyton 6, Daniel M. Altmann 7, Rawleigh Howe2 & Melanie J. Newport 1 Check for updates 1234567890():,; 1234567890():,; Available evidences suggest that podoconiosis is triggered by long term exposure of bare feet to volcanic red clay soil particles. Previous genome-wide studies in Ethiopia showed association between the HLA class II region and disease susceptibility. However, functional relationships between the soil trigger, immunogenetic risk factors and the immunological basis of the dis- ease are uncharted. Therefore, we aimed to characterise the immune profile and gene expression of podoconiosis patients relative to endemic healthy controls. Peripheral blood immunophenotyping of T cells indicated podoco- niosis patients had significantly higher CD4 and CD8 T cell surface HLA-DR expression compared to healthy controls while CD62L expression was sig- nificantly lower. The levels of the activation markers CD40 and CD86 were significantly higher on monocytes and dendritic cell subsets in patients com- pared to the controls. RNA sequencing gene expression data indicated higher transcript levels for activation, scavenger receptors, and apoptosis markers while levels were lower for histones, T cell receptors, variable, and constant immunoglobulin chain in podoconiosis patients compared to healthy controls. Our finding provides evidence that podoconiosis is associated with high levels of immune activation and inflammation with over-expression of genes within the pro-inflammatory axis. This offers further support to a working hypothesis of podoconiosis as soil particle-driven, HLA-associated disease of immuno- pathogenic aetiology. Podoconiosis is a form of lymphoedema that causes progressive in podoconiosis is triggered by an unidentified volcanic clay soil painful swelling of the legs1. It is a neglected tropical disease that component causing lymphatic fibrosis that leads to swelling and affects poor communities living in remote highland regions in nodular changes in the foot and lower leg2. The nature of the soil endemic countries. Available evidence suggests that inflammation trigger is not precisely known but epidemiological and geological 1 Brighton and Sussex Centre for Global Health Research, Department of Global Health and Infection, Brighton and Sussex Medical School, Brighton, UK. 2 Armauer Hansen Research Institute, Addis Ababa, Ethiopia. 3Department of Medical Laboratory Science, College of Health Sciences, Addis Ababa University, Addis Ababa, Ethiopia. 4Department of Biochemistry and Biomedicine, School of Life Sciences, University of Sussex, Brighton, UK. 5School of Public Health, Addis Ababa University, Addis Ababa, Ethiopia. 6Department of Infectious Disease, Imperial College London, London, UK. 7Department of Immunology and Inflammation, Imperial College London, London, UK. e-mail: [email protected] Nature Communications | (2024)15:2020 1 Article https://doi.org/10.1038/s41467-024-46347-z studies suggest certain climatic conditions are required that con- immunophenotyping and RNA expression studies, moving toward tribute for the formation of the soil particles. These include annual improved understanding of the pathogenesis of podoconiosis. rainfall of above 1500 mm, an altitude above 1500 m and surface temperature of 19–21 °C3. Such topographic features are abundant Results globally, but the lack of affordable protective footwear to prevent Study subjects, clinical, and socio-demography data contact with the soil (and hence disease) limits the geographical A total of 64 podoconiosis patients and 49 healthy controls were distribution of podoconiosis to low-income countries in Africa, Asia enrolled in the study. The majority of the patients (59 of 64, 92.2%) and Central and South America2. were in stage 2 of the disease clinically based on the Tekola staging It is estimated that there are about four million cases of podo- system17, had bilateral disease (58, 90.6%), were males (36, 55.7%), and coniosis globally and the highest prevalence of the disease is observed their mean age was 47.8 ± 12 years. The majority of healthy controls in Cameroon and Ethiopia with prevalence of 8.08% and 7.45%, were males (26 of 49, 53.1%), their mean age was 34.4 ± 8 years, and respectively4. A recent study in 2017 used epidemiological data and almost all the participants were farmers (Supplementary Table 1). modeling techniques to estimate that around 1.5 million cases of podoconiosis in Ethiopia and the prevalence of podoconiosis in the Immunophenotyping of peripheral blood mononuclear cells area where the current study was conducted (Gojam Zone, Amhara T cells. Surface expression of markers of activation (HLA-DR, CD38), region) is 3.5%5,6. memory (CD62L) and proliferation (Ki-67) was analyzed on CD4 and Affected people experience progressive swelling and debilitating CD8 T cells from 56 podoconiosis cases and 44 healthy control subjects pain associated with intermittent episodes of acute adenolymphangi- using a gating strategy as defined in Supplementary Fig. 2. The average tis. These episodes hamper normal day-to-day and agricultural activ- percentage of CD4 and CD8 T cells expressing HLA-DR was significantly ities which patients’ livelihoods depend on, leading to further higher in podoconiosis patients compared to healthy controls (P < 0.001 impoverishment and negative psychosocial impacts6,7. Familial clus- with median of 10.7% vs 7.1% for HLA-DR on CD4 and 23.4% vs 15.8% for tering of the disease suggests genetic factors play a role in the HLA-DR expression on CD8 cells, respectively). In contrast, the expres- pathogenesis of the disease. Segregation analysis of multigenerational sion of CD62L on these T-cell subsets was significantly lower in podo- podoconiosis-affected families in Wolaita, southern Ethiopia, showed coniosis patients compared to healthy controls (CD4CD62L, P < 0.0001, an estimated sibling recurrence risk ratio (λs) and heritability of 5.07 with median value of 53.6% vs 63.1%, respectively; CD8CD62L, P = 0.001, and 0.63, respectively8. These findings led to the undertaking of with median value of 22.8% vs 36.8%, respectively). There were no dif- genome-wide association study (GWAS) in the Wolaita population9 and ferences between patients and healthy controls in CD38 and Ki-67 a second larger GWAS in three ethnic populations in Ethiopia10. Both expression on either CD4 or CD8 T-cell subsets (Fig. 1). these studies reported that variation in HLA class II genes (DRB1, DQA1, Furthermore, we confirmed the increase in activation marker by and DQB1) was significantly associated with susceptibility to podoco- comparing median MFI of each activation marker among markers niosis. The Occam’s razor explanation for HLA class II disease asso- ungated. For example, the MFI of HLA-DR on CD4 T cells was 131 in ciations is generally a central role of CD4 T cells. podoconsiosis patients versus 87 in healthy controls (P < 0.0001). Taken together with the epidemiological observations, these Similarly, the MFI of HLA-DR on CD8 T cells was significantly different findings suggest that a mineral or other exogenous compound present (348 versus 242, respectively, P = 0.0004). We observed the same in soil triggers an HLA-mediated immune response in susceptible concordance comparing CD62L in the study groups, CD62L MFI on individuals that targets the lymphatic system. Earlier studies by Dr CD4 T cells was 475 and 1539 in patients and controls, respectively Ernest Price suggested minerals absorbed through the skin are taken (P < 0.001); CD62L MFI on CD8 T Cells was significantly lower in up by macrophages and transported to lymph nodes to initiate an patients compared to controls (98.6 versus 187, respectively, inflammatory response11. Price also observed that silicate particles P = 0.0019). caused subendothelial edema, endolymphangitis, collagenisation and obliteration of the lymphatic lumen12. More recent studies demon- Monocytes. The expression of HLA-DR, CD40, CD86, and CD36 was strated patients had thickened dermal collagen, reduced elastic fibers, analyzed on monocytes from 43 podoconiosis patients and 34 healthy dilated and often sclerotic blood vessels, with a moderate lympho- controls. Monocytes were first gated into classical, intermediate and plasmacytic infiltrate which also contained mast cells, and scattered non-classical monocyte subsets based on CD14 and CD16 expression macrophages, but few neutrophils or eosinophils13. (see Supplementary Fig. 3). There were no statistically significant dif- Although class II HLA molecules typically present “foreign” anti- ferences in the distribution of the three monocyte subsets between gens of pathogen origin, they can also be involved in T-cell-mediated podoconiosis patients and healthy controls (Supplementary Fig. 4). autoimmune or hypersensitivity disease through direct recognition of Expression of the activation markers CD40 and CD86 was sig- self-antigens, modified epitopes, or through molecular mimicry14. An nificantly higher on classical monocytes from podoconiosis patients example of possible relevance to podoconiosis is the pathogenesis of compared to healthy controls, with median values of 35.6% vs 25.5% for berylliosis in which HLA-DP gene products are implicated, either CD40 and 13.7% vs 7.4% for CD86 (P = 0.03, P = 0.001, respectively). through presentation of a self-peptide modified by beryllium, or There were no differences between the two groups in expression of through direct binding of beryllium to the HLA-peptide binding CD40 or CD86 on non-classical and intermediate monocyte subsets, groove, thus triggering an inflammatory CD4 T-cell response15. Soil although there was a trend toward higher expression in podoconiosis from podoconiosis-endemic regions is rich in a number of elements patients. CD36 expression in the intermediate monocyte subpopula- including beryllium, one of six elements shown to be statistically tion was significantly higher in podoconiosis patients than in healthy linked to podoconiosis16. Beyond the genetic association with the class controls with median value of 56.7% vs 33.2% (P < 0.0001), as shown II HLA region, little is known about the immune response in podoco- in Fig. 2. niosis, reflecting the neglected status of the condition and the com- munities it affects. However, understanding the mechanisms of Dendritic cells. Expression of HLA-DR, CD40 and CD86 on dendritic disease could accelerate the development of effective treatments and cells (DCs) was analyzed in 43 podoconiosis patients and 34 healthy early diagnostic or screening tests to identify susceptible individuals controls after they were sorted into the three DC subsets, myeloid that could help eliminate the condition. As the soil trigger and an (mDC), plasmacytoid (pDC) and cross-presenting (cp-DC) based on experiment model on podoconiosis is lacking, this study aimed to their CD11c, CD123 and CD141 expression respectively (see Supple- characterize innate and adaptive immunity in podoconiosis through mentary Graph 5). Nature Communications | (2024)15:2020 2 Article https://doi.org/10.1038/s41467-024-46347-z Fig. 1 | Median expression of CD62L, HLA-DR, CD38, and Ki-67 markers on CD4 (A–D) and CD8 (E–H) T cells, respectively, in 56 podoconiosis patients (Podo) and and CD8 T cells from podoconiosis patients and healthy controls. The figure 44 healthy controls (HC). *P values were derived using the Mann–Whitney U test of shows box plots depicting median, interquartile range, minimum and maximum two-sided independent t test. Source data that is used to generate this graph is values of subset frequencies positive for CD62L, HLA-DR, CD38, and Ki-67 on CD4 provided as a “Source Data” file Figs. 1–3. There was no significant difference in the distribution of the three immune response, immunoglobulin domain, and TCR alpha and beta DC subsets between podoconiosis patients and healthy controls domains. The clusters with the highest enrichment score are presented (Fig. 3A–C). Evaluation of the activation biomarkers within the three in Fig. 5A, B for the down- and upregulated GO categories separately. subsets showed that the expression of CD40 was significantly higher in all three DC subsets, in particular on mDCs, in podoconiosis patients Protein–protein interaction network compared to healthy controls (median value of 8.4% vs 3.7%, respec- Enrichment categories with a score of more than 1.3 and a P value of tively, P = 0.003). There was no statistically significant difference 1.5 and P < 0.05. The F test was used for deriving P values and adjust- healthy control samples. The log2-fold change values are plotted on the x axis and ments were made for multiple comparison. Source data that is used to generate this compared with the negative log10 P values on the y axis. Blue dots represent graph is provided as “Source Data” file Figs. 4 and 5. HLA molecule, thereby altering the repertoire of self-antigens that tuberculosis29 and HIV30, or autoimmune origin such as rheumatoid binds to the HLA pocket. This eventually drive the binding of new self- arthritis31 and systemic lupus erythematosus32. However, there is little peptides that activate polyclonal CD8 T-cell response27. Given the role direct evidence that such activated cells are antigen-specific. In fact, in of soil exposure, it is possible that an exogenous element or mineral studies of HIV in particular it would appear that the vast majority of particles incorporating an immunogenic element could similarly activated T cells are not HIV-specific33, raising the question as to interact with a particular HLA molecule in podoconiosis. This could be whether expression of activation markers in chronic diseases neces- further investigated through analysis of the T-cell receptor repertoires sarily reflects a presumed antigen induction. in podoconiosis patients. It is widely accepted that HLA-DR expression represents a state of T-cell expression of HLA-DR along with other activation markers cellular activation. But there are some discrepancies on its expression such as CD38, CD69 and Ki-67, is known to be induced after activation and functional role on human T cells in the literature. One study by antigens or mitogens28. For example, increased expression of acti- showed increased expression of HLA-DR on T cells enhanced memory vation markers is commonly seen in acute infections, and other pool generation and activation of cytotoxic CD8 T cells for anti-tumor chronic inflammatory conditions of either infectious origin such as responses through T-cell–T-cell synapse formation and IFN-γ Nature Communications | (2024)15:2020 6 Article https://doi.org/10.1038/s41467-024-46347-z Fig. 5 | Functional enrichment and protein–protein network analysis for the enriched pathway. Here nodes represent proteins with different colors represent- top upregulated and downregulated genes in podoconiosis patients relative to ing the different submitted gene lists and additional proteins with interaction with healthy controls. Most enriched categories for downregulated (A) and upregu- the submitted gene lists. The lines represent known and predicted interactions lated (B) genes in podoconiosis patients. The y axis represents enrichment scores between proteins, with the thickness of the line indicating the strength of inter- as –log10 (P value) and the x axis indicates the gene ontology (GO) category enri- action. The F test was used for deriving P values and adjustments were made for ched in the pathway analysis. Gene names from the most enriched GO category multiple comparison. Source data that is used to generate this graph is provided as were submitted to the STRING database for the downregulated (C) and upregulated “Source Data” file Figs. 4 and 5. genes (D) separately and a protein network interaction was generated for the given secretion34. The higher HLA-DR expression on fresh PBMC T cells in monocytes was higher in podoconiosis patients. The three DC subsets podoconiosis patients could be due to bystander activation from the had significantly higher levels of CD40 expression while expression of pro-inflammatory cytokine milieu. Consistent with this possibility, we CD86 was higher only in myeloid DCs in podoconiosis patients com- observed higher levels of TNF-α and IL-1β in unstimulated culture wells pared to the healthy controls. The classical monocytes and myeloid and IL-1β mRNA in the peripheral blood of podoconiosis patients DCs are the main subsets in each lineage which are involved in pro- (manuscript in preparation). duction of inflammatory cytokines and reactive oxygen species during The expression of CD62L in podoconiosis patients was sig- infection or recognition of ligands by their pattern recognition nificantly lower compared to healthy controls further highlighting the receptors38,39. In other chronic inflammatory HLA-associated diseases activated state of T cells in podoconiosis patients. Reduced CD62L such as multiple sclerosis40 and coeliac disease41, monocytes have been expression is a feature of effector memory T cells and following CD62L shown to have increased expression of CD40, CD86 and HLA-DR in shedding, memory T cells re-enter the circulation from lymph nodes ex vivo stimulation assays compared to healthy controls. Similarly, where they can exert their effector function35. The recruitment of immunohistochemistry analysis in Crohn’s disease and ulcerative T cells to inflamed sites and their effector function is greatly enhanced colitis patients showed the levels of activation markers like CD83 and after shedding of CD62L and acquisition of selectins such as CD62E CD86 expressed on DCs, primarily mDCs, were significantly higher in and CD62P to attach to the endothelial cells wall36. Immunohis- inflamed mucosa from patients compared with non-inflamed mucosa tochemistry analysis of Price’s archival lymph node samples showed and healthy controls42,43. It has been suggested that mature DCs with that CD4 T lymphocytes were the predominantly infiltrating cells from higher levels of activation marker expression are likely to play a key podoconiosis patient biopsy samples, suggesting these T cells could role in mediating inflammation at the site of pathology43. Dendritic play a role in development of the disease37. cells are strategically and abundantly located under the skin sub- There was no significant difference in the distribution of the three mucosa and they could be the first immune cells to be exposed to monocytes subsets (classical, intermediate and non-classical) between potential foreign antigens or particles that get access through the skin podoconiosis patients and healthy controls. However, the expression in podoconiosis44. In the current study we have only studied cells from of the co-stimulatory molecules CD86 and CD40 among classical peripheral blood, targeted analysis of the local microenvironment Nature Communications | (2024)15:2020 7 Article https://doi.org/10.1038/s41467-024-46347-z Table 1 | List of significantly upregulated genes from the top enriched clusters based on their gene ontology category, in peripheral blood from 19 podoconiosis patients and 15 healthy controls GO category Upregulated gene Gene description Fc Podo vs HC P value Immune response CD1A CD1a molecule 1.7 0.03 CD80 CD80 molecule 1.2 0.07 CD86 CD86 molecule 1.3 0.06 HLA-DQB1 HLA-DQ beta molecule 1.8 0.3 WAKMAR Wound and keratinocyte-associated lnRNA 1.57 0.028 Cell membrane and signaling OLR1 Oxidized low-density lipoprotein 1.74 0.03 CD163L1 CD163-like molecule 1 1.57 0.02 MSR1 Macrophage scavenger receptor 1 1.6 0.05 ADIPOR1 Adiponectin receptor 1.54 0.04 Cell projection ITGA1 Integrin subunit alpha 1 1.3 0.04 FFAR4 Free fatty acid receptor 4 1.6 0.03 Extracellular region CTSB Cathepsin B 1.7 0.01 MPO Myeloperoxidase 1.7 0.04 EREG Epiregulin 2.5 0.006 BNIP3L BCL2 interacting protein 3 like 1.57 0.04 GO gene ontology, Fc fold change, HC healthy controls, vs versus. Gene names are italicized. Table 2 | List of significantly downregulated genes from the top enriched clusters based on their gene ontology category, in peripheral blood from 19 podoconiosis patients and 15 healthy controls GO category Downregulated gene Gene description Fc Podo vs HC P value Histones H3C12 H3 clustered histone 12 3.1 0.0002 H2AC12 H2A clustered histone 12 3 0.0002 H2BC14 H2B clustered histone 14 2.8 0.001 H3C8 H3 clustered histone 8 3.1 0.0002 Cell division cycle CDC6 Cell division cycle 6 2.3 0.0001 CDCA5 Cell division associated 5 2.6 0.003 Immunoglobulin domain IGLV3–21 Immunoglobulin lambda variable 3–21 4.9 0.005 IGLV1–47 Immunoglobulin lambda variable 1–47 3.5 0.007 IGLC3 Immunoglobulin lambda constant 3 3 0.006 IGLC2 Immunoglobulin lambda constant 2 2 0.03 T-cell receptor TRAV26-1 T-cell receptor alpha variable 26-1 1.57 0.03 TRAV10 T-cell receptor alpha variable 10 1.74 0.04 TRBC1 T-cell receptor beta constant 1 1.75 0.03 GO gene ontology, Fc fold change, HC healthy controls, vs versus. Gene names are italicized. could reveal more insight into which cells and subsets are mainly The upregulation of HLA-DQB1 in this study is of interest due to localized in the inflamed areas. the documented association between this gene (as well as HLA- The RNA-sequencing result showed 108 genes were significantly DRB1) and podoconiosis9,10. Variation in the HLA-DQB1 gene and its upregulated in podoconiosis and functional enrichment analysis for upregulation have been linked with susceptibility to different solid these genes showed the cell membrane and immune response clusters organ and tissue fibrosis45,46. A recent study in 2022 by Zhou et al. had the highest enrichment score in the pathway analysis. The upre- which integrated RNA-Seq with GWAS data indicated a SNP gulated immune response genes like CD80, CD86, CD1A, and HLA-DQB1 (rs9273410) in the HLA-DQB1 region was significantly associated correlated with the PBMC surface immunophenotyping data where with increased susceptibility to silicosis47. Silica particles have been classical and myeloid DC also exhibited higher expression of these identified in macrophages and lymph nodes in people living in markers. Further review of the ‘immune response’ GO cluster and podoconiosis-endemic regions11 although evidence directly impli- genes within this category (CD1A, CD80, CD86, and HLA-DQB1) cating silica in its pathogenesis is currently lacking. However, none revealed related pathways which were in the same biological process of these studies, including the current study, elucidated the associated with numerous autoimmune and infectious diseases. Many potential epitope or the mechanism by which HLA-DQB1 contribute of these diseases themselves are caused by complex interactions towards silicosis or the development of fibrosis. Hence, further between environmental and host genetic factors, including HLA gene studies to elucidate the function and regulation of this region in variants. such pathologies are warranted. Nature Communications | (2024)15:2020 8 Article https://doi.org/10.1038/s41467-024-46347-z Variation in HLA genes has been associated with a number of ITGA1, and BNIP3L which directly or indirectly shape the inflammation infectious diseases, raising the possibility that infection may play a role and the extracellular matrix. However, further research to identify the in podoconiosis. However, the available evidence is more consistent soil trigger in podoconiosis is still required. with a noncommunicable etiology and there is no evidence of direct The significantly downregulated genes in podoconiosis patients involvement of a pathogenic organism. Recent studies which analyzed were associated with the GO categories such as; nucleosome, histone, the skin microbiome of podoconiosis patients reported the presence cell division, immune globulin domain and TCR receptor domains. Of of distinct bacterial species48,49. Superficial skin infections leading to these enrichment clusters the histones had the highest enrichment acute dermatolymphangioadenitis attacks could exacerbate the lym- score in the pathway analysis followed by cell division clusters. His- phoedema and cause local inflammation but this tends to occur later in tones are the major structural component of the nucleosome. Mod- the disease, once lymphoedema is established49. Therefore, it is con- ification of histones through deacetylation or methylation is a ceivable that skin microbiome changes may also contribute to disease fundamental aspect of gene regulation by altering access to the progression and the increase in activation markers in podoconiosis binding of transcription factors or polymerases. Histone modification patients, but are unlikely to be involved in initial pathogenesis. None of plays a major role in overall cell division and transcription of genes60. A the study subjects had acute dermatolymphangioadenitis at the time review by Kerstin Klein and Steffen Gay indicated DNA methylation of enrollment. and posttranslational histone modifications in particular in synovial Genes which were significantly upregulated in the cell membrane fibroblasts play a role in the development of rheumatoid arthritis61. It is and signaling GO clusters were mainly from the scavenger receptor not clear how this modification might play a role in podoconiosis, family including OLR1, CD163L1, and MSR1. These receptors are hence future studies are needed to elucidate the role of epigenetic expressed on different cells such as endothelial cells, monocytes and modification in podoconiosis. macrophages. In line with this higher level of CD36 was observed in The downregulation, in particular in the TCR receptor genes was intermediate monocytes subsets of podoconiosis patients from the expected given that PBMC immune-phenotyping showed a relatively PBMC immunophenotyping in the current study. Similar to the higher level of expression of activation markers in podoconiosis CD36 scavenger receptor, these receptors have a broad range of patients. TCR downregulation following persistent TCR ligation is one ligands including oxidized lipoproteins, heat shock proteins, asbestos, mechanism of limiting hyperactive immune responses after pathogen and silica50,51. control62. It is not clear if this downregulation was clone-specific or due Phagocytosis of silica by macrophages leads to impaired lysoso- to a bystander immune exhaustion effect due to the chronic nature of mal degradation because of the particulate nature of the silica. This the disease. Targeted analysis of the TCR repertoire to characterize pathway of silica absorption in macrophages derived from murine cell T-cell clonality or TCR usage would be interesting to investigate in line was linked with induction of inflammatory cytokines, reactive more detail in podoconiosis. oxygen specious, apoptosis and fibrosis50,52. It was suggested CD36 In conclusion the high level of activation markers on T cells, contributes to sterile inflammation via internalization of components classical monocytes and mDCs suggest that these subsets could play a like oxidized lipoproteins and cholesterol crystals by assembling the central role in priming and driving the immune response in podoco- NLRP3 inflammasome complex and secretion of pro-inflammatory niosis patients, keeping in mind that the latter two subsets are the main cytokines53. Podoconiosis patients may have a higher rate of silica producers of inflammatory cytokines. Moreover, upregulated levels of binding and retention compared to healthy controls which could be immune activation, inflammatory enzymes and scavenger receptor mediated by higher expression of these scavenger receptors. Of transcripts suggest persistent inflammation and impaired adipose course, once internalized, differences in responses to the minerals tissue metabolism could be taking place in the lower legs of podoco- between patients and healthy controls could also play a role in sus- niosis patients. This could progressively lead to development of a ceptibility to developing the disease (it has already been shown in the fibrotic microenvironment and impaired lymphatic function. In the current study that there were substantial immunological, lysosomal current study we have made progress in describing and understanding enzyme and protein differences between the two groups). Price’s ele- the immune response in podoconiosis, yet specific driving pathways mental analysis indicated some differences in the ratio of aluminum to and the causative agent(s) remain to be elucidated. Replication and silica and birefringent particles (uncoated particles) found in lymph validation of these findings observed in podoconiosis patients by node samples from patients relative to healthy controls11. corroborating with data from tissue biopsy analysis could ultimately Higher levels of transcripts from the genes encoding the BCL lead to potential treatment options within these pathways, which interacting protein 3 like protein (BNIP3L) which is involved in activa- could ameliorate the disease symptoms and progression. tion and assembly of the apoptosome protein complex domains54, cathepsin B (CTSB) which is involved in lysosomal protein degradation, Methods processing and presentation as well as extracellular matrix Ethical approval and informed consent degradation55, and myeloperoxidase (MPO) which is involved in reac- A rapid ethical appraisal was conducted using in-depth interviews and tive oxygen species production56,57 were observed in podoconiosis focus group discussion prior to sample collection to tailor the patients compared to healthy controls. Similarly, higher levels of informed consent process to the local socioeconomic norms and to transcripts from the genes encoding epiregulin (EREG), wound and address the participants’ concerns before undertaking the research63. keratinocytes migration associated lnRNA (WAKMAR) which are Ethics approval was obtained from the AHRI/ALERT (Protocol No. PO- involved in tissue and wound healing and integrin subunit alpha 1 3818) and Ethiopian National Science and Technology (Ref N0. AH/ (ITGA1) which is a cellular adhesion molecule for collagen and 00229/001241/21) ethics review committees in Ethiopia and the BSMS laminin58,59 were observed in podoconiosis patients compared to Research Governance and Ethics Committee in the UK (Ref. No. ER/ healthy controls. The binding of collagen and laminin in the extra- BSMS9DJB/1). The study participants gave written consent before cellular matrix through ITGA1 could contribute in the pathology of sample collection. Clinical and biological data obtained from study podoconiosis through enhancing the fibrosis. Although the role of participants were kept secure and confidential as per the data sharing silica in the immunopathogenesis of podoconiosis is still unclear, it and protection agreement between AHRI and BSMS partners. remains possible that silica (and/or other soil mineral) induces inflammation leading to the release of reactive oxygen species and Study population and design apoptosis of cells52,53,58 which could lead to pathology through dis- A community-based case-control study design was employed to enroll ruption of lysosomes and the upregulation of genes like MPO, CTSB, podoconiosis patients and healthy control individuals from northeast Nature Communications | (2024)15:2020 9 Article https://doi.org/10.1038/s41467-024-46347-z Ethiopia. The study was conducted in Bahir Dar and neighboring dis- cells were collected for each analysis. All data were analyzed using the tricts which is a podoconiosis-endemic area. Podoconiosis cases were FlowJo software (Mac Version 9.6, TreeStar Inc, USA). selected from two health centers in the Bahir Dar district. The cases were already diagnosed and staged by trained nurses and health offi- Statistical analysis cers using a validated clinical algorithm64 and the Tekola staging The non-parametric Mann–Whitney U test was applied to assess the system17 and were registered in the health center log books. We difference in the expression of different surface and intracellular bio- deliberately selected podoconiosis cases which were stage two and markers and the median MFI because in general distributions were not three. Stage 1 podoconiosis patients were excluded due to diagnostic observed to be normal for most of the activation markers. However, we difficulties as the clinical features of early podoconiosis could be also compared marker expressoin using mean MFI values and got very confused with a range of other conditions such as heart or liver failure similar P values between podoconiosis patients and healthy controls. that also lead to ankle swelling. Late stages (stages 4 and 5) were also GraphPad Prism V-6 (GraphPad Software Inc., CA, USA) was used to excluded as in these advanced stages of disease fibrosis predominates calculate significance levels using two-sided tests where P value of less and any active inflammation involved in the pathogenesis of podoco- than 0.05 was considered statistically significant. niosis may have “burnt out”. Unrelated sex and age-matched healthy controls who were over the age of 18 years old, had lived in the same RNA extraction, library preparation and sequencing study area as the patients at least for 10 years, who had no family RNA was extracted from samples collected in PAXgene tubes using history of podoconiosis and had never worn shoes consistently, and MagMAXTM bead-based extraction method (Life Technologies, Catalog were therefore exposed to the soil without developing podoconiosis, no. 4451894). The extracted RNA was checked for quantity and quality were selected. Individuals who had underlying chronic disease like using the QubitTM RNA high-sensitivity assay kit (Invitrogen, Catalog diabetes, liver, kidney or cardiac disease, or who were taking medica- no. Q32855) and the dsDNA high-sensitivity assay kit (Invitrogen, tion were excluded from the study. Moreover, individuals who were Catalog no. Q32854). A total 0.5 µg RNA samples from 24 podoconiosis unwell on the day of sample collection (for example, due to inter- and 24 healthy control samples were then used for preparing the current infection), cases who had clinical evidence of secondary skin sequencing libraries. The libraries were prepared using strand-specific infection or acute dermatolymphangioadenitis attack (ADLA), those fast select library preparation kit based on the manufacturer’s with known HIV infection or who tested positive for HIV with screening instruction (Qiagen, Catalog no. 180450). Briefly, RNA was fragmented tests (based on the national testing algorithm) were also excluded.See and ribosomal and globin RNAs removed by incubating it in decreasing Supplementary Table 1 for sociodemographic summary). temperature gradient using QIAseq FastSelect rRNA and globin mRNA removal kit” (Qiagen, Catalog no. 335376). The whole sample from the Specimen collection and laboratory assays fragmentation step was used for first-strand synthesis using reverse Whole blood was collected from a total of 64 podoconiosis patients transcriptase buffer mix. Second strand synthesis was carried out and 49 healthy controls using heparinised tubes for immunological using 2nd strand buffer and enzyme mix. The 5’ phosphorylation during assays and using PAXgene tube for transcriptimics studies. Periph- 2nd strand synthesis allows subsequent strand-specific ligation with eral blood mononuclear cells (PBMCs) were isolated by Ficoll- the Illumina Dual-Index Y-Adapters. The 24-plex single-use adapter Hypaque (Sigma-Aldrich, Catalog no. GE17-5442-03) density gra- plate was used for the current library preparation. Following adapter dient centrifugation from the heparinised peripheral blood. The ligation the libraries were subjected to PCR amplification cycles. PBMCs were used for immunophenotyping studies of T cells, Finally, the libraries were cleaned based on their size using blue pippin monocytes and dendritic cells by using multicolor flow cytometry. (Sage Science) and sequenced in two runs using the Illumina Next- For the transcriptomics assay 24 samples were selected from each Seq500 NGS platform. study group by matching them based on their age, sex, PBMC number and RNA quantity and quality. PBMC were isolated, stained Quality controls and trimming the sequencing reads and paraformaldehyde fixed at the study site laboratory located on The quality of the sequencing was assessed based on the Phred scale average 15 km from the sample collection site. The fixed samples using FatQC V 0.11.9, and almost all reads have base qualities above were stored at -80oC with freezing media until transportation. Fixed Phred score of 30, which represent an incorrect call rate of 1 in 1000 samples were air-transported within two weeks of processing for bases, giving a base call accuracy of 99.9%. The quality of the reads, flow cytometry acquisition at AHRI in Addis Ababa. The libraries prior to and post-trimming with cutadapt for one representative were also prepared in AHRI and shipped to UK, Leeds University for sample is shown in Supplementary Fig. 5A, B, and the overall quality sequencing. score for all samples generated by MultiQC V1.13. before and after trimming is presented in Supplementary Fig. 6C, D. Flow cytometry The freshly isolated PBMCs were stained with monoclonal antibodies Mapping reads to the human reference genome, sorting and targeting the respective markers for each cell lineage. All antibodies indexing the mapped reads were from Beckton Dickinson (BD) unless specified. T cells (PBMC of Once the reads were controlled for quality and the short inserts ~2 × 105) were stained with CD3-APC-H7, CD4-BV510, CD8-BV421, trimmed, a splice-aware program HISAT2 v2.2.0 was used to align CD38-APC, HLA-DR-PE-Cy7, Ki-67-PerCPCy5.5, CD62L-PE, monocytes the reads to the reference human genome (Ensembl GRCh38). A (PBMC of ~4 × 105) were stained with CD16-FITC, CD14-BV421, CD40- program named SAMtools v0.1.20 was used to compress the aligned PE, CD86-BV510, CD36-PerCP-Cy5.5, HLA-DR-APC-Cy7, dendritic cells reads to a binary format (.bam), index and sort them based on (PBMC of ~4 × 105) were stained with CD11c-BV421, CD123-PerCP-Cy5.5, chromosome order. The sorted.bam file was used by the program CD141-APC, CD40-PE, CD86-BV510, HLA-DR-APC-Cy7, and a lineage featureCounts to count the number of reads that mapped to the cocktail comprised of FITC conjugated antibodies to CD3, CD14, CD16, reference human genome. The alignment rate was very good, CD19, CD20 and CD56. Intracellular staining for Ki-67 was performed whereby ~95% of the reads for all samples mapped to the reference by permeabilising cells with Cytofix/CytopermTM (BD, Catalog no. genome in the current study (Supplementary Tables 3 and 4 pre- 554714). All data were acquired using a FACS Canto II flow cytometer sents summarized output from feature counts including the total (BD Biosciences, San Jose, CA, USA). Unstained and compensation number of reads mapped, primary mapped reads, assigned reads, controls were run during every acquisition. A minimum of 100 000 adapter reads and the overall alignment rate for healthy control and events for T cells and 200 000 events for monocytes and dendritic podoconiosis samples, respectively). Nature Communications | (2024)15:2020 10 Article https://doi.org/10.1038/s41467-024-46347-z Differential gene expression (DGE) 6. Molla, Y. B., Tomczyk, S., Amberbir, T., Tamiru, A. & Davey, G. The counted reads from featureCounts were imported to R (V4.4.3) Podoconiosis in East and West gojam zones, northern Ethiopia. and DGE was performed using edgeR packages. Prior to performing PLoS Negl. Trop. Dis. 6, e1744 (2012). the DGE analysis, the reads were filtered to remove genes below the set 7. Tekola, F. & Yeshanehe, W. E. Podoconiosis: tropical lymphedema of parameter and normalized by assigning a normalization factor for each the lower legs. Center for Research on Genomics and Global Health sample. Reads with more than 5 million counts were included in the National Human Genome Research Institute National Institutes of final DGE analysis. The potential effect of age and sex differences in the Health, USA. https://www.researchgate.net/profile/Fasil-Tekola- DGE was also controlled by fitting these variables into the analysis Ayele/publication/259570324_Podoconiosis_tropical_lymphedema_ model in edgeR. A P value of less 0.05 from the negative log10p was of_the_lower_legs/links/5475c6ae0cf2778985af10e6/Podoconiosis- considered significant. tropical-lymphedema-of-the-lower-legs.pdf (2014). 8. Davey, G. et al. Podoconiosis: a tropical model for gene–environment Biological functions and pathway enrichment analyses interactions? Trans. R. Soc. Trop. Med Hyg. 101, 91–96 (2007). To further explore the biological functions and pathways to which the 9. Tekola, F. et al. HLA class II locus and susceptibility to podoconiosis. differentially expressed genes belong, the list of differentially expres- New Engl. J. Med. 366, 1200–1208 (2012). sed genes was submitted to an online gene ontology (GO) analysis site. 10. Gebresilase, T. et al. Replication of HLA class II locus association The GO and functional enrichment analysis was performed using with susceptibility to podoconiosis in three Ethiopian ethnic groups. Database for Annotation, Visualization and Integrated Discovery Sci. Rep. 11, 1–11 (2021). (DAVID; https://david.ncifcrf.gov/). The list of differentially expressed 11. Price, E. & Henderson, W. The elemental content of lymphatic tis- genes were submitted online to DAVID with their Ensemble gene ID. sues of barefooted people in Ethiopia, with reference to endemic Pathway enrichment analysis was performed by selecting default terms elephantiasis of the lower legs. Trans. R. Soc. Trop. Med Hyg. 72, from the following databases: Kyoto Encyclopedia of Genes and Gen- 132–136 (1978). omes (KEGG), Gene ontology terms in biological processes (BP), cel- 12. Price, E. The pathology of non-filarial elephantiasis of the lower lular component (CC), molecular function (MF) Canonical Pathways legs. Trans. R. Soc. Trop. Med Hyg. 66, 150–190 (1972). (CP), and the Comprehensive Resource of Mammalian protein com- 13. Wendemagegn, E., Tirumalae, R. & Böer‐Auer, A. Histopathological plexes (CORUM). and immunohistochemical features of nodular podoconiosis. J. Cutan. Pathol. 42, 173–181 (2015). Construction of protein–protein interaction network 14. Fiorillo, M. T., Paladini, F., Tedeschi, V. & Sorrentino, R. HLA class I or To assess the downstream interaction of the differentially expressed class II and disease association: catch the difference if you can. genes in biological processes, protein–protein interaction (PPI) net- Front. Immunol. 8, 1475 (2017). work analysis was performed using the Search Tool for the Retrieval of 15. Saltini, C. et al. Major histocompatibility locus genetic markers of Interacting Genes/Proteins (STRING) database (http://string-db.org/). beryllium sensitization and disease. Eur. Respir. J. 18, A network was built from our data by submitting gene lists from the 677–684 (2001). highly enriched clusters and the lowes P value for the upregulated and 16. Gislam, H. et al. (eds) Links Between Soil Composition and Podo- downregulated genes separately. Protein interaction networks were coniosis Occurrence and Prevalence in Cameroon (EGU2020: built for the different enriched clusters based on significantly differ- Sharing Geoscience Online, 2020). entially expressed genes from Tables 1 and 2 (see “Results”). 17. Tekola, F., Ayele, Z., Mariam, D. H., Fuller, C. & Davey, G. Develop- ment and testing of a de novo clinical staging system for podoco- Reporting summary niosis (endemic non‐filarial elephantiasis). Trop. Med. Int. Health 13, Further information on research design is available in the Nature 1277–1283 (2008). Portfolio Reporting Summary linked to this article. 18. Mangalam, A. K., Taneja, V. & David, C. S. HLA class II molecules influence susceptibility versus protection in inflammatory diseases Data availability by determining the cytokine profile. J. Immunol. 190, All available data are included in the manuscript and the supplemen- 513–519 (2013). tary file. Data sources used to generate graphs is provided with this 19. Mallal, S. et al. Association between presence of HLA-B* 5701, HLA- paper as ‘Source Data’. The processed RNA-Seq data is also provided as DR7, and HLA-DQ3 and hypersensitivity to HIV-1 reverse-tran- a source data along with the raw RNA-sequencing data which has been scriptase inhibitor abacavir. Lancet 359, 727–732 (2002). deposited in ArrayExpress (https://www.ebi.ac.uk/arrayexpress/) 20. Ooi, J. D. et al. Dominant protection from HLA-linked under the accession number E-MTAB-13860. Source data are provided autoimmunity by antigen-specific regulatory T cells. Nature 545, with this paper. 243–247 (2017). 21. Gabriel, S. B. et al. The structure of haplotype blocks in the human References genome. science 296, 2225–2229 (2002). 1. Price, E. The association of endemic elephantiasis of the lower legs 22. Yin, L. et al. Recognition of self and altered self by T cells in auto- in East Africa with soil derived from volcanic rocks. Trans. R. Soc. immunity and allergy. Protein cell 4, 8–16 (2013). Trop. Med Hyg. 70, 288–295 (1976). 23. Oldstone, M. B. Molecular mimicry and immune‐mediated diseases. 2. Price, E. W. Podoconiosis: Non-filarial Elephantiasis (Oxford Uni- FASEB J. 12, 1255–1265 (1998). versity Press, 1990). 24. Martina, S., Fabiola, F., Federica, G., Chiara, B. & Gioacchino, L. Gian 3. Deribe, K. et al. Spatial distribution of podoconiosis in relation to LdA. Genetic susceptibilty and celiac disease: what role do HLA environmental factors in Ethiopia: a historical review. PLoS ONE 8, haplotypes play? Acta Biomed. 89, 17–21 (2018). e68330 (2013). 25. Lombardi, G. et al. HLA-DP allele-specific T cell responses to ber- 4. Deribe, K., Cano, J., Trueba, M. L., Newport, M. J. & Davey, G. Global yllium account for DP-associated susceptibility to chronic beryllium epidemiology of podoconiosis: a systematic review. PLoS Negl. disease. J. Immunol. 166, 3549–3555 (2001). Trop. Dis. 12, e0006324 (2018). 26. Richeldi, L., Sorrentino, R. & Saltini, C. HLA-DPB1 glutamate 69: 5. Deribe, K. et al. Estimating the number of cases of podoconiosis a genetic marker of beryllium disease. Science 262, 242–244 (1993). in Ethiopia using geostatistical methods. Wellcome Open Res. 27. Illing, P. T. et al. Immune self-reactivity triggered by drug-modified 2, (2017). HLA-peptide repertoire. Nature 486, 554–558 (2012). Nature Communications | (2024)15:2020 11 Article https://doi.org/10.1038/s41467-024-46347-z 28. Caruso, A. et al. Flow cytometric analysis of activation markers on patients in the North West Region of Cameroon. BMC Microbiol. 23, stimulated T cells and their correlation with cell proliferation. 189 (2023). Cytometry: J. Int Soc. Anal. Cytol. 27, 71–76 (1997). 49. Neidhöfer, C. et al. Tropical leg lymphedema caused by podoco- 29. Silveira-Mattos, P. S. et al. Differential expression of activation niosis is associated with increased colonisation by anaerobic bac- markers by Mycobacterium tuberculosis-specific CD4+ T cell dis- teria. Sci. Rep. 13, 13785 (2023). tinguishes extrapulmonary from pulmonary tuberculosis and latent 50. Palecanda, A. & Kobzik, L. Receptors for unopsonized particles: the infection. Clin. Infect. Dis. 71, 1905–1911 (2019). role of alveolar macrophage scavenger receptors. Curr. Mol. Med. 1, 30. Hertoghe, T. et al. T cell activation, apoptosis and cytokine dysre- 589–595 (2001). gulation in the (co)pathogenesis of HIV and pulmonary tuberculosis 51. Murphy, J. E., Tedbury, P. R., Homer-Vanniasinkam, S., Walker, (TB). Clin. Exp. Immunol. 122, 350–357 (2001). J. H. & Ponnambalam, S. Biochemistry and cell biology of 31. Nanki, T. & E Lipsky, P. Cytokine, activation marker, and chemokine mammalian scavenger receptors. Atherosclerosis 182, 1–15 receptor expression by individual CD4+ memory T cells in rheu- (2005). matoid arthritis synovium. Arth. Res. Ther. 2, 1–9 (2000). 52. Chao, S. K., Hamilton, R. F., Pfau, J. C. & Holian, A. Cell surface 32. Zhou, H. et al. Dysregulated T cell activation and aberrant cytokine regulation of silica-induced apoptosis by the SR-A scavenger expression profile in systemic lupus erythematosus. Mediat. receptor in a murine lung macrophage cell line (MH-S). Toxicol. Inflamm. 2019, (2019). Appl. Pharm. 174, 10–16 (2001). 33. Papagno, L. et al. Immune activation and CD8+ T-cell differentiation 53. Sheedy, F. J. et al. CD36 coordinates NLRP3 inflammasome acti- towards senescence in HIV-1 infection. PLoS Biol. 2, e20 (2004). vation by facilitating intracellular nucleation of soluble ligands into 34. Gérard, A. et al. Secondary T cell–T cell synaptic interactions drive particulate ligands in sterile inflammation. Nat. Immunol. 14, the differentiation of protective CD8+ T cells. Nat. Immunol. 14, 812–820 (2013). 356–363 (2013). 54. Imazu, T. et al. Bcl-2/E1B 19 kDa-interacting protein 3-like protein 35. Richard, L. et al. L-selectin (CD62L) expression distinguishes small (Bnip3L) interacts with bcl-2/Bcl-xL and induces apoptosis by resting memory CD4+ T cells that preferentially respond to recall altering mitochondrial membrane permeability. Oncogene 18, antigen. J. Immunol. 170, 28–32 (2003). 4523–4529 (1999). 36. Yang, S., Liu, F., Wang, Q. J., Rosenberg, S. A. & Morgan, R. A. The 55. Turk, V. et al. Cysteine cathepsins: from structure, function and shedding of CD62L (L-selectin) regulates the acquisition of lytic regulation to new frontiers. Biochim Biophys. Acta 1824, activity in human tumor reactive T lymphocytes. PLoS ONE 6, 68–88 (2012). e22560 (2011). 56. Nauseef, W. M. Myeloperoxidase in human neutrophil host defence. 37. Yardy, A., Williams, A. T. & Davey, G. Podoconiosis pathogenesis: Cell Microbiol. 16, 1146–1155 (2014). renewed use of an historical archive. Trans. R. Soc. Trop. Med. Hyg. 57. Van der Veen, B. S., de Winther, M. P. & Heeringa, P. Myeloperox- 112, 417–418 (2018). idase: molecular mechanisms of action and their relevance to 38. Boyette, L. B. et al. Phenotype, function, and differentiation poten- human health and disease. Antioxid. Redox Signal. 11, tial of human monocyte subsets. PLoS ONE 12, e0176460 (2017). 2899–2937 (2009). 39. Serbina, N. V. et al. Distinct responses of human monocyte subsets 58. Zhang, W.-M. et al. α11β1 integrin recognizes the GFOGER to Aspergillus fumigatus conidia. J. Immunol. 183, sequence in interstitial collagens. J. Biol. Chem. 278, 2678–2687 (2009). 7270–7277 (2003). 40. Chuluundorj, D., Harding, S. A., Abernethy, D. & La Flamme, A. C. 59. Calderwood, D. A., Tuckwell, D. S., Eble, J., Kuhn, K. & Humphriesi, Expansion and preferential activation of the CD14+ CD16+ mono- M. J. The integrin a1 A-domain is a ligand binding site for collagens cyte subset during multiple sclerosis. Immunol. Cell Biol. 92, and laminin. J. Biol. Chem. 272, 12311–12317 (1997). 509–517 (2014). 60. Bannister, A. J. & Kouzarides, T. Regulation of chromatin by histone 41. Cinova, J. et al. Gliadin peptides activate blood monocytes from modifications. Cell Res. 21, 381–395 (2011). patients with celiac disease. J. Clin. Immunol. 27, 201–209 (2007). 61. Klein, K. & Gay, S. Epigenetics in rheumatoid arthritis. Curr. Opin. 42. Despalatović, B. R., Babić, M., Bratanić, A., Tonkić, A. & Vilović, K. Rheumatol. 27, 76–82 (2015). Difference in presence and number of CD83+ dendritic cells in 62. Gallegos, A. M. et al. Control of T cell antigen reactivity via pro- patients with ulcerative colitis and Crohn’s disease. Sci. Rep. 10, grammed TCR downregulation. Nat. Immunol. 17, 379–386 (2016). 1–8 (2020). 63. Negash, M. et al. Rapid ethical appraisal of stakeholder views on 43. Middel, P., Raddatz, D., Gunawan, B., Haller, F. & Radzun, H.-J. research prior to undertaking immunopathogenesis studies on Increased number of mature dendritic cells in Crohn’s disease: podoconiosis in northeast Ethiopia during a period of social evidence for a chemokine mediated retention mechanism. Gut 55, instability. Trans. R. Soc. Trop. Med. Hyg. 115, 1026–1038 (2021). 220–227 (2006). 64. Deribe, K. et al. Developing and validating a clinical algorithm for 44. Haniffa, M., Gunawan, M. & Jardine, L. Human skin dendritic cells in the diagnosis of podoconiosis. Trans. R. Soc. Trop. Med. Hyg. 114, health and disease. J. Dermatol. Sci. 77, 85–92 (2015). 916–925 (2020). 45. Kamdem, S. D., Moyou-Somo, R., Brombacher, F. & Nono, J. K. Host regulators of liver fibrosis during human schistosomiasis. Front. Acknowledgements Immunol. 9, 2781 (2018). The authors thank the Amhara regional health bureaus and the Amhara 46. Fingerlin, T. E. et al. Genome-wide imputation study identifies novel Public Health Institute for their support with participant enrollment, data HLA locus for pulmonary fibrosis and potential role for auto- and sample collection in Ethiopia. The authors like to thank Dawit Hailu at immunity in fibrotic idiopathic interstitial pneumonia. BMC Genet. AHRI for his assistance during library preparation. The authors also thank 17, 1–12 (2016). the study participants who volunteered to participate in this study. This 47. Zhou, Y. et al. Integrating RNA-Seq with GWAS reveals a novel SNP work was supported by the National Institute for Health Research (NIHR) in immune-related HLA-DQB1 gene associated with occupational Global Health Research Unit on NTDs at Brighton and Sussex Medical pulmonary fibrosis risk: a multi-stage study. Front. Immunol. 12, School using Official Development Assistance (ODA) funding [Grant 5822 (2022). number GHR 16/136/29], grant received by M.J.N. and G.D. The views 48. Nkwetta, D. L. et al. Phenotypical characterization, and antibiotics expressed here are those of the author(s) and not necessarily those of susceptibility patterns of skin bacteria found in podoconiosis the NIHR or the Department of Health and Social Care. Nature Communications | (2024)15:2020 12 Article https://doi.org/10.1038/s41467-024-46347-z Author contributions Reprints and permissions information is available at M.N.: sample collection, data analysis, figure production, and writing the http://www.nature.com/reprints first draft of the manuscript. M.C.: study conception and fieldwork supervision. T.G.: sample collection and data analysis. F.A.: sample Publisher’s note Springer Nature remains neutral with regard to jur- collection and data analysis. D.Alc.: data analysis, writing manuscripts, isdictional claims in published maps and institutional affiliations. and interpretation. B.T.: data analysis and interpretation. D.Alt.: data analysis and interpretation. R.B.: data analysis and interpretation. G.D.: Open Access This article is licensed under a Creative Commons study design and manuscript write-up. R.H: study design, data inter- Attribution 4.0 International License, which permits use, sharing, pretation, manuscript writing, and Co-PI in Ethiopia. M.J.N.: study con- adaptation, distribution and reproduction in any medium or format, as ception and design, writing the manuscript, PI for funding, and overall long as you give appropriate credit to the original author(s) and the study. All authors edited the manuscript and approved the final version. source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this Competing interests article are included in the article’s Creative Commons licence, unless The authors declare no competing interests. indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended Additional information use is not permitted by statutory regulation or exceeds the permitted Supplementary information The online version contains supplementary use, you will need to obtain permission directly from the copyright material available at holder. To view a copy of this licence, visit http://creativecommons.org/ https://doi.org/10.1038/s41467-024-46347-z. licenses/by/4.0/. Correspondence and requests for materials should be addressed to © The Author(s) 2024 Mikias Negash. Peer review information Nature Communications thanks Jürgen Bosch and Manuel Ritter for their contribution to the peer review of this work. A peer review file is available. Nature Communications | (2024)15:2020 13