Introduction

Leishmaniasis is a disease caused by a group of protozoan parasites belonging to the genus Leishmania, transmitted to mammals via the bite of infected Phlebotomine sandflies. In humans, leishmaniasis has three major clinical manifestations: cutaneous leishmaniasis (CL), mucosal leishmaniasis (ML), and visceral leishmaniasis. It is distributed widely in tropical and subtropical regions, affecting 98 countries [1]. ML corresponds to lesions with significant tissue destruction occurring mainly on the oropharyngeal and/or nasal mucosa, that appears as a systemic complication in a proportion of patients either months or years—usually one to five years—after cutaneous lesions have healed [2], predominantly in the South American region [3]. ML seriously affects patients’ quality of life due to treatment failure, relapse, and slight or severe facial deformity [2, 4].

Leishmania species belonging to the Leishmania (Viannia) subgenus, such as L. (V.) braziliensis, L. (V.) guyanensis, and L. (V.) panamensis, are predominant in South America and associated with the occurrence of ML. In CL cases produced by L. (V.) braziliensis, pentavalent antimonial treatment failure is higher when compared with cases from other members of the L. (V.) subgenus [3, 5–7]. In addition to the parasite species, studies have found host factors such as the patient’s age and sex and immune response gene polymorphism to be involved in the appearance of ML [8–14]. It is widely accepted that the host’s immune response is of great relevance in the pathophysiology of ML [15, 16].

LRV is a cytoplasmatic double-strand RNA virus associated with some Leishmania spp. strains [17, 18]. Experiments involving mice have shown that the infection with L. (V.) guyanensis carrying Leishmania RNA virus type 1 (LRV1) promotes inflammation and subverts the host immune response [19]. Mice infected with L. (V.) guyanensis LRV1–negative but that were simultaneously infected with exogenous dsRNA viruses, such as lymphocytic choriomeningitis virus (LCMV), showed immunological and clinical profiles equivalent to those presented by isogenic mice infected with L. (V.) guyanensis LRV1–positive [20]. Comparison of in vitro infections of human macrophages by LRV1–negative with infections by LRV–positive strains of L. (V.) braziliensis revealed that the cytokines’ expression profile in the latter case showed Th2 predominance [21].

Therefore, it is clear that the presence of LRV1 does not go unnoticed by the host immune system. However, the clinical relevance of human infections with LRV1-carrying Leishmania strains is unknown. Two studies consistently suggest that LRV1 is associated with therapeutic failure in CL [22, 23]^, but findings regarding the role of LRV1 in the occurrence of ML are inconclusive [24–27].

The present work corresponds with a retrospective case-control study involving 103 Colombian patients suffering from either ML or CL to assess the association between ML and LRV1. The main findings from this study support this association but additionally emphasize the protective role of a complete treatment for the cutaneous leishmaniasis event in the occurrence of ML.

Methods

Results

One hundred and three patients’ clinical records and biopsies were examined; 70 corresponded to controls and 33 to cases. The male-to-female ratio was 2.81 (76:27) for the whole group, 2.88 (52:18) for controls, and 2.66 (24:9) for cases. The variables associated with the age of the patients at the time of diagnosis of the events (ML or CL) did not show a normal distribution. The median age at which patients in the control group were diagnosed with CL was 23.5 years (95% CI:22–27.72). Notably, seven (21.2%) patients from the cases’ group did not report a previous cutaneous lesion; the median age at which the remaining 26 cases were diagnosed with CL was 29.24 years (95% CI: 23.45–35.54). In the cases group, the median age at which ML was diagnosed was 38 years (95% CI: 33–45).

In the case group, the received treatment for the CL event was: 12 patients received meglumine antimoniate, one received miltefosine, one received ketoconazole, seven did not receive treatment because they did not have cutaneous leishmaniasis, and 12 patients were not treated with any medication. Of the treated patients in the cases group, 12 showed skin lesion healing, and 2 (14.3%) experienced therapeutic failure; one was treated with meglumine antimoniate, and one was treated with miltefosine. In the control group, 66 patients were treated with meglumine antimoniate, two were treated with ketoconazole, and in two patients, the treatment regimen was not stated in their clinical records. Among the controls, 44 showed skin lesion healing, while 21 (30.1%) patients experienced therapeutic failure; from those, 20 were treated with meglumine antimoniate and 1 with ketoconazole. There was no information regarding the treatment outcome in three patients from the control group treated with meglumine antimoniate. No statistical differences were found between cases and controls in CL therapeutic failure to meglumine antimoniate (p = 0.211).

The comparison of the analyzed variables between cases and controls is presented in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Bivariate association between mucosal leishmaniasis, other relevant variables, and Leishmania RNA Virus 1 infection.

https://doi.org/10.1371/journal.pone.0317221.t001

In the bivariate analysis presented in Table 1, the variables significantly differing between cases and controls are the completed treatment for the CL event and carrying a parasite infected with LRV1.

In the cases’ group, lesions were located exclusively on the nasal mucosa in 75.7% of the patients, lesions involving both nasal and oral mucosa were observed in 15.2% of the patients, and lesions involving nasal, oral, and pharyngeal mucosa in 9.4% of the patients. Following the classification system proposed by Lessa for ML [35], most of the mucosal lesions (73%) were stage III and IV.

From the group of cases, in patients who presented the CL antecedent (n = 26), the median onset for ML was three years after the episode of CL (95% CI: 0.5–14 years). The presence of LRV1 was not correlated with the elapsed time between the diagnoses of CL and ML manifestation (Spearman coefficient = -0.024, p = 0.90). A comparison within the cases group was conducted to determine whether there was an association between the presence of LRV1 and the occurrence of ML when there was no clear history of CL. Four out of seven patients who did not present symptoms of CL were positive for LRV1, contrasting with seven out of twenty-six patients who presented CL antecedent. However, this data distribution shows no significant difference between carrying or not the virus and the absence of CL clinical symptoms before developing ML (p = 0.186).

From the whole group of analyzed samples, L. (V.) braziliensis was identified in sixty; L. (V.) panamensis in seventeen; L. (V.) guyanensis in two, and L. (V.) panamensis/guyanensis as complex in one sample. In twenty-three samples, it was not possible to identify the parasite species. L. (V.) braziliensis was present in 63.6% (95% CI,45.2–78.7) of the case samples and 55.7.% (95% 43.6–67.1) of the control samples. L. (V.) panamensis was identified in 18.1% (95% CI,8.0–36.1) of the case and 15.7% (95% CI,8.7–26.5) of the control samples (S1 and S3 Tables) without significant difference in species distribution between cases and controls (Fig 1C–1E).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Distribution of samples according to the geographic origin of the infection, Leishmania species, and LRV1 statutes.

A. Presumptive geographical origin of the acquired infection. Cases are represented as dots, and controls as crosses. Red symbols correspond with samples LRV1(+), and blue ones with LRV1(-). B. Geographic origin distribution of the total group analyzed (top pie), cases (middle left pie), controls (middle right pie), samples LRV1 (+) (bottom left pie), and samples LRV1(-) (bottom right pie). The number on the bottom of each pie corresponds with the n. Panels C.- E. present the distribution of the species identified by hsp70 PCR-RFLP or miniexon PCR in cryopreserved biopsies according to the corresponding patient group and the LRV1 status of the sample. The panel presented in C. corresponds to the species distribution and LRV1 status of the whole group, D. to the cases group, and E. to the control group. Analyses of species distribution between cases/control groups and between LRV1(+) /LRV1(-) were performed using Pearson’s X2 test and Fisher’s exact test, and no significant differences were found. L. panamemsis/guyanensis corresponds to samples that presented an undistinguishable pattern between L. (V.) guyanensis and L. (V.) panamensis. Shadows on the map correspond with the Andes. The map was created using QGIS 3.38.3 (https://qgis.org/license/). The shapefiles were obtained from publicly available and open-access sources: Natural Earth Data - 10m Physical Vectors (https://www.naturalearthdata.com/downloads/10m-physical-vectors/10m-ocean/), DIVA-GIS Data (https://diva-gis.org/data.html) and GADM (https://gadm.org/download_country.html).

https://doi.org/10.1371/journal.pone.0317221.g001

LRV1 was detected in seventeen clinical samples, corresponding to a frequency of 16.5% (95% CI, 10.4–25.1) (S1 and S2 Tables). Six samples, 8.5% (95% CI, 3.8–18.1), corresponded with biopsies from the control group, and eleven, 33.3% (95% CI, 18.8–51.7) corresponded to samples from the cases group. Regarding the occurrence of LRV1 according to Leishmania species, twelve LRV1 positive samples corresponded to L. (V.) braziliensis (cases:7, controls:5), two samples to L. (V.) panamensis (cases:2, controls:0), and three samples to unidentified Leishmania species (cases:2, controls:1) (S2 Table). There was no significant difference in distribution by species between LRV1 negative and positive samples.

Information regarding the LRV1 status, parasite species, control follow-up, and geographical origin for each sample is presented in S1 Table and Fig 1. Concerning the geographic origin of the infection, most of the evaluated samples came from the Andean region (52.5%), followed by samples from the Orinoco (22.2%) and Amazon regions (21.2%). Pacific and Caribbean regions, as the infection geographic origin, were found only in 4% of the evaluated sample. In 3.8% of the revised clinical records, there was no information regarding the geographical origin of the infection. The LRV1 distribution was found in samples from all those regions except the Caribbean. Evaluation of the LRV1 positivity according to the region where the sample came from shows that 19% of the samples from the Amazon carried the virus, 18.2% of the samples from the Orinoco, and 15.4% from the Andean. Despite the underrepresentation of the Pacific region, one out of two samples of L. (V.) braziliensis was found to be LRV1 infected (S1 Table).

The bivariate analysis for assessing factors that could show differences between case and control groups identified complete treatment and LRV1 infection as factors with statistically significant differences between groups (Table 1). Based on that analysis, variables with p < 0.20 were selected to be incorporated into the logistic regression analysis (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Multivariate analysis of factors associated with mucosal leishmaniasis.

https://doi.org/10.1371/journal.pone.0317221.t002

The bivariate analysis identified two variables with a potentially significant association with the occurrence of ML. These were complete treatment for the CL event (OR = 0.043; 95% CI, 0.01–0.14) and detectable LRV1 (OR = 5.33; 95% CI, 1.55–19.42). The model in the multivariate analysis showed the presence of LRV1 associated with ML (OR = 6.30; 95% CI,1.52–26.10; p = 0.011) and the fact of receiving a complete treatment for the CL event as a protective factor for ML occurrence (OR = 0.039; 95% CI, 0.01–0.12; p = 0.0001) (Table 2).

Discussion

ML is a disturbing clinical complication associated mainly with infections from New World Leishmania species with an uncertain incidence. There is enough evidence supporting the role of the patient’s immune system in triggering ML [15, 16]. However, insufficient knowledge is yet to guide clinical decisions or well-defined prevention strategies, such as early progression markers from the CL to the ML form of the disease. This case-control study identified an association between the presence of LRV1 in clinical samples and the occurrence of ML (OR = 6.30; 95% CI,1.52–26.10; p = 0.011), and the multivariate analysis showed that providing a completed treatment for CL event decreased the risk of the presentation of ML [OR, 0.039 (95% IC,0.01–0.12); p,0.0001].

The frequency of LRV1 in the whole group of samples was relatively low at 16.5% compared with other Latin-American reports [24–26, 36]. The rate of LRV1 was higher in cases compared to controls, as shown in Table 1. This study also examined the potential link between the parasite species and mucosal leishmaniasis (ML) occurrence. It found that L. (V.) braziliensis was the main species in both groups of patients (cases: 21, controls: 39), followed by L. (V.) panamensis (cases: 6, controls:11). There was no significant difference in species distribution between cases and controls, as shown in Fig 1 and Table 1. Nevertheless, it is essential to note that most patients were infected with parasites from the Leishmania (Viannia) subgenus, which are closely related species [37, 38]. There is a limitation given that it was not possible to identify the Leishmania species in 23 samples (cases:4, controls:19).

A recent study analyzed 56 L. (V.) panamensis isolates from cutaneous lesions in Panama from 2014–2018 and found that 20% were infected with LRV1 [39]. This contrasts with the observation in Colombia, where around 170 isolates of L. (V.) panamensis collected between 1983 and 2021 were found to have none infected with LRV1 [40]. It’s worth noting that none of the referenced studies involved L. (V.) panamensis isolates from ML patients. However, in the analysis presented here, L. (V.) panamensis was found in samples from both groups (cases, 18.1%, and controls, 15.7%), and the only two biopsies infected with L. (V.) panamensis carrying LRV1, belonged to the ML group. Taken together, these observations, along with the fact that most of the samples in this study came from the Amazon, Orinoco, and central and east Andean Colombian regions (Fig 1), suggest that the differential geographic distribution and frequencies of LRV1 could be indicative of different parasite gene flows associated with each environment [41].

Cantanhêde´s study implied the association between LRV1 and the occurrence of ML in a cohort of 147 patients (CL,109 and ML, 38) [24]; Kariyawasam did not find a link between specific clinical symptoms and the presence of LRV1, but the study did reveal that the relative viral burden of LRV1 was highest in L. (V.) braziliensis isolates associated with ML in a group of American tegumentary leishmaniasis patients [26]; Ito found a high frequency of LRV1 in the mucosa of patients suffering from ML [25], while Pereira found a low frequency of LRV1 in a cohort of CL and ML patients, without LRV1(+) samples between the ML group [27]. This inconclusive evidence could be because ML seems to be a multifactorial disease with regional variations. Still, differences in study designs, including varying lengths of patient follow-up, can produce no comparable results. Based on 103 patients (70 with CL and 33 with ML), this multivariate analysis suggests that contracting an infection with Leishmania spp.-LRV1(+) is independently linked to ML. This result aligns with Cantanhêde’s [24] study, which had a similar patient sample size and was designed to compare CL vs. ML patients. However, the wide confidence interval of the odds ratio (OR) linking LRV1 with ML observed in the present study indicates that a larger sample size is needed to achieve greater precision in this association.

Moreover, the multivariate analysis revealed that completing the treatment for CL significantly reduced the risk of developing ML [OR, 0.039 (95% CI, 0.01–0.12); p < 0.0001]. Although this study did not follow a prospective design to directly assess the impact of completed CL treatment on preventing ML, the results are consistent with previous findings [42, 43]. They support the long-held but unproven assumption that systemic treatment of CL with the appropriate dosage and duration can effectively prevent ML [44, 45]. The systemic versus local treatment of New World CL is a hot topic of discussion. One of the main arguments supporting local therapies is based on reports of viable Leishmania parasites in the original cutaneous scars and the intact nasal mucosa of patients after completing systemic treatment with highly toxic medications [46, 47]. Additionally, even in the present study, it has been observed that some patients adequately treated for CL may still develop ML.

Experimental evidence has shown that Leishmania spp. carrying LRV1 can lead to complications such as relapse and parasite metastasis [19, 20]. The authors of observational studies showing LRV1’s association with therapeutic failure in CL treatment with antimony and pentamidine have suggested that LRV1-mediated changes in the human host immune response potentially affect the efficacy of drugs [22, 23]. It is well established that the host immune system plays a crucial role in the response to treatments [48–50]. Therefore, the mechanisms involving each of the variables associated with ML occurrence are likely rooted in the background of the host immune system, and their exact contributions to the multifactorial occurrence of ML may be challenging to assess.

One of the main limitations of retrospective studies in identifying risk factors is finding appropriate controls and ensuring proper follow-up. To address this, we selected 70 controls from patients diagnosed with CL up to 2010, guaranteeing a minimum follow-up of 8 years after the skin condition diagnosis. Among these controls, 33 were followed up for a median of 12.8 years (ranging from 0.5 to 22.16 years) to ensure the absence of ML symptoms; 32 controls had a less than six-month follow-up period, and five patients were not followed up after completing their CL treatment. Consequently, one limitation of the study was the inability to follow up with all the patients in the control group.

The findings from the present study must be considered in the geographical context of the disease occurrence. The Eastern Mediterranean Region accounts for 80% of the CL cases reported worldwide; however, the incidence of ML in immunocompetent patients is concentrated in the South American Region [1], where most infections are associated with species belonging to the Leishmania (Viannia) subgenus. Consequently, there is an urgent need for a prospective multicentric study that involves LRV1 detection, parasite species identification, and hopefully parasite and viral load estimation in clinical samples from CL patients, with a proper post-treatment follow-up to assess the actual weight of LRV1 as a prognostic marker. This approach would allow a better understanding of the genesis of ML and its geographical determinants. A randomized controlled trial to assess the effectiveness of CL treatment in preventing ML occurrence is not ethically suitable. Therefore, the role of this factor will continue to be a subject of debate until a better understanding of the immunological factors mediating the CL response to treatment is achieved.

Supporting information

Acknowledgments

The authors thank the CDFLLA Staff, particularly Dr. Claudia Colorado, Dr. Carolina Camargo, and Dr. Andres Gonzalez, for collaborating in the search for the medical records and specimens in the CDFLLA biobank; Dr. Luis A. Gamboa for helping with clinical examinations; and Mrs. Emma Lopez for contacting and inviting patients from the control group to participate in the study.