In this work we carried out exploratory research regarding the presence of microorganisms resistant to meta(oid)s and antibiotics in patients with CKD. We also performed an analysis of total DNA of stool samples to find resistance determinants in each patient, depending on CKD stage.

To analyze resistance/tolerance of metal(loid) and resistance to antibiotics in bacteria, samples from patients with CKD were analyzed in YCFAm culture medium using the MIC calculated from E. coli isolated from fecal samples (Additional file 7: Table S3). Our results exhibited colony growth in the presence of arsenic and lead under aerobic conditions (Additional file 1: Figure S1). No colonies were found on plates containing YCFAm medium in the presence of Hg and Cd (data not shown). Colonies harvested from each plate were used to perform an antibiotic resistance analysis.

Bacterial isolates were analyzed using the Kirby Bauer method in the presence of paper disks with antibiotics and characterized according to CLSI parameters for resistant, intermediate, or sensitive bacteria (Fig. 2A and B) in the presence or absence of ½ MIC of metal(oid)s, a sublethal concentration. For healthy subjects, a decreased appearance of gentamicin resistance in the presence of metal(oid)s, ceftazidime, and ciprofloxacin was observed compared with healthy subjects, which indicates a significant increase in resistance to antibiotics in the presence of metal(loid)s. For CKD3 patients, no significant difference in antibiotic resistance was observed compared with the healthy controls and ½ MIC of E. coli for metal(loid)s (arsenic and mercury both). However, higher numbers of microorganisms resistant to antibiotics ampicillin, cefazolin, and ciprofloxacin were shown in plates containing ½ MIC of E. coli metal(loid)s. No differences between groups (CKD3 versus healthy controls) and toxicant presence were observed when growth curves of stool samples in the presence of ½ MIC of metal(loid)s (Additional file 2: Figure S2).

Susceptibility to antibiotics in bacteria isolated from stool samples of healthy patients and those with stage 3 chronic kidney disease. Inhibitory zones of bacteria isolated from stool samples were plated on YCFAm agar using the Kirby Bauer method [24]. Paper disks containing gentamicin (10 µg), ampicillin (10 µg), ciprofloxacin (5 µg), cefazolin (30 µg), ceftazidime (30 µg), and gentamicin (120 µg) were placed on previously seeded bacteria on agar plates and incubated at 37 °C for 24 h. Growth inhibition zones were measured and compared with the standardized diameter described by CLSI [25]. ½ MIC of arsenic and lead of E. coli was used to evaluate the effect of metal(loid)s on antibiotic resistance, n = 12. A healthy subjects, B CKD Stage 3 subjects

To identify microorganisms in fecal samples that showed co-resistance to antibiotics and metal(loid)s, sequences of isolated colonies were analyzed by 16S amplicon sequencing. All isolates of healthy subjects and CKD subjects from each disease stage (3, 4, and 5) were compared after antibiotic treatments in the presence of ½ MIC of E. coli of arsenic or lead. For CKD3 patients and lead treatment, Firmicutes and Proteobacteria were mainly found (Fig. 3A). Genus analysis showed Pseudomonas spp., Janibacter spp., Escherichia/Shigella spp., and Bacillus spp.; resistant bacteria to gentamicin, cefazolin, ceftazidime, and ampicillin were also found. For CKD3 patients and arsenic treatment Pseudomonas spp., Escherichia spp./Shigella spp., Bacillus spp. were observed; and Enterococcus spp. showed resistance to ampicillin, ciprofloxacin, and gentamicin. Furthermore, some colonies of interest exhibited resistance to all tested antibiotics (general resistance). On the other hand, for healthy controls and lead treatment, we observed mainly Bacillus spp. and Pseudomonas spp. and resistance to ampicillin, cefazolin, gentamicin, and colonies with a general resistance. Healthy controls and arsenic treatment exhibited Bacillus spp., Escherichia spp./Shigella spp. and Pseudomonas spp. (Fig. 3A–C). It was impossible to detect microorganisms of Phylum Bacteroidetes, mainly because of other selection phenomena such as the culture medium used (YCFAm) and aerobic conditions. Parameters of microorganism abundance in samples grown in the synthetic and selective media were performed, Chao1, Shannon, and Simpson indices were calculated [S2A for metal(loid)s and S2B for antibiotics]. Alpha diversity indexes were associated with the selection pressures presented in healthy patients and CKD3, although, these parameters were not analyzed in depth.

Relative abundance of bacteria. Taxonomic profiling of the 16 s rRNA amplicon was performed from colonies that showed antibiotic resistance. Colonies were isolated from healthy and stage 3 CKD subjects and compared in the presence of arsenic, lead, and antibiotics (ampicillin, cefazolin, ceftazidime, ciprofloxacin, and gentamicin) with general or multi-resistance. The abundance of each sample was performed by read count analysis of the phylum (A) and genus (B). C beta-diversity was determined by principal component analysis (PCA) and results sorted by OTUs > 0.2%. Experimental groups are indicated by color and shape, arsenic (green) and lead (orange), healthy subject (triangle) or CKD patients (circle)

We found that the Bacillus spp. in the presence of metal(loid)s. showed multiple resistance to ampicillin, gentamicin, cefazolin, ceftazidime and ciprofloxacin Pseudomonas spp. showed the same behavior in the presence of metal(loid)s, exhibiting the same antibiotic resistance, but less representation.

On the other hand, Enterococcus spp. and Escherichia spp./Shigella spp. exhibited resistance to ampicillin and gentamicin, respectively. We noted that both healthy controls and CKD3 patients showed some bacteria with general resistance to antibiotics, but resistance differed mainly in the type of metal (arsenic for CKD stage 3 and lead for a healthy group). Most studies that have found this co-resistance phenomena in stool samples have been conducted in wastewater systems [37]. The phenomenon of co-selection allows conserving and promoting resistance to antibiotics and resistance/tolerance of metal(loid)s in bacteria, even in the absence of antibiotics through different co-resistance mechanisms, which opens the way to future investigations [17].

In taxonomic profiling, we identified the presence of Janibacter spp., which has been recently reported to be involved in the virulence process related to antibiotic resistance in HGM of celiac and immunocompromised subjects [38], expressing proteins such as β-lactamase and other 7 genes related to metal(loid) resistance/tolerance including arsenic [39]. This led us to perform further analysis to characterize the isolated strain and identify the genes when this gut pathogen was exposed to metal(loid)s and antibiotics.

Beta diversity analysis between control and CKD stage 3 isolated bacteria showed no significant differences when stool samples were cultured under aerobic conditions and when exposed to metalloids and different antibiotics (Fig. 3C). This could be explained by a continuous process of the multi-resistance capacity of bacteria. Interestingly, this result suggests that the early stages of CKD do not influence the development of multi-resistant bacteria. To explore this idea in depth, we performed an analysis to detect resistance markers by qPCR.

It was initially suggested that CKD patients have augmented resistance genes in HGM because of a possible accumulation or transcendent increase of toxicant compounds such as metal(loid)s related to decreased renal filtration rate and the use of antibiotics for disease treatment. To determine the presence of resistance genes and genetic determinants related to metal(loid)s and antibiotics resistance in stool samples of healthy patients and to compare with subjects at different stages of CKD (3, 4, and 5), qPCR was performed (Additional file 6: Table S2). Primers were constructed to amplify regions of selected genes for metal(loid) resistance analysis corresponding to arsenic (arsC, arsA), lead (pbrA), mercury (merA), and cadmium (cadA). Specifically, for cadmium, 2 primers cadA2k and cadA3k were designed because in silico PCR results against MEGARes and CARD databases showed amplicons of different sizes and quantities. To increase coverage, both primers were synthesized.

To choose primers for antibiotic resistance genes, 140 described genes were analyzed [30] and 13 genes that represented each family were selected. For genes such as floR, a gene product resistance to chloramphenicol, a second product was initially detected in melting curves (not shown). To correct this bias, qPCR was performed with an annealing temperature gradient (60, 65, and 70 °C). It was observed that at 65 °C, only a single specific product was obtained. sulll primers of Sulfonamides dihydropteroate synthetase described by the group of Szczepanowski et al. [30] could not be used due to the detection of nonspecific products when analyzing melting curves in our samples (not shown). This could mean nonspecific interactions in other DNA regions, and therefore nonspecific amplification in our samples, producing false-positive results.

Pie charts (Fig. 4) describe the presence of antibiotic and metal(loid) resistance genes. Interestingly, the healthy control group (Fig. 4A) showed a higher number of resistance markers of metal(loid)s and antibiotics. In the same group, all analyzed genes for resistance to antibiotics and metals were found. The lack of some markers could not be evidenced in this group in comparison with the different stages of CKD. From CKD stage 3 samples (Fig. 4B), the presence of strB, dhfr1, floR, acrB, and arr2 for antibiotic resistance genes and cadA3k, arsC, and cadA2k for metal(loid)s resistance genes were observed. In the same group mefE1, catB4 and qnrB1 were not found. For stage 4 samples (Fig. 4C), the presence of acrB, arr2, qnfrB1, strB, dhfr1, floR, ermB, and tetA antibiotics were found and cadA3k, arsC, cadA2k, and pbrA for metal(loid) resistance genes, ermB was found again in stage 4. Concerning metal resistance genes, arsA and pbrA were not observed in CK3 subjects, however pbrA was identified in stage 4 patients. For stage 5 CKD samples (Fig. 4D), which correspond to subjects who no longer have renal function and require kidney replacement therapies, such as hemodialysis or peritoneodialysis and Firmicutes versus Bacteroidetes ratio tends to decrease because of the uremic state and increased proteolytic metabolism [7, 8], both groups of genes, for resistance to antibiotics and metal(loid)s were diminished compared with healthy subjects. The presence of qnrB1, dhfr1, and floR genes for resistance to antibiotics and cadA2k and merA for metal resistance was observed. We consider that the typical state of intestinal microbiota dysbiosis associated with CKD produces changes in microbiota that are associated with a smaller number of genetic determinants.

Pie charts indicating antibiotic resistance genes (cold colors) and metal(loid)s (warm colors) analyzed by qPCR of DNA obtained from stool samples. A mefE1, arr2, catB4, strB, dhfr1, floR, tetA, ermB, acrB, qnrB1, cadA3k, arsC, arsA, cadA2k, and pbrA genes were detected in healthy controls. B strB, dhfr1, floR, acrB, arr2, cadA3k, cadA2k, and arsC genes were detected in subjects with stage 3 CKD. C acrB, arr2, qnrB1, strB, dhfr1, floR, ermB, tetA, cadA2k, cadA3k, arsC, and pbrA genes were detected in subjects with stage 4 CKD. D qnrB1, floR, dhfr1, merA, and cadA2k genes were detected in subjects with stage 5 CKD. The graph represents the total appearance of resistance genes in 4 patient samples analyzed by group. The total identified appearances were 96 hits for healthy subjects, 14 hits for stage 3 CKD, 23 hits for stage 4 CKD and 5 hits for stage 5 CKD. CKD groups were compared with healthy controls using Two-Way ANOVA with Dunnett’s test for multiple comparisons (two-tailed) P < 0.0001 (****)

The study of the presence of metal(loid) resistance genes and antibiotics from the early stages of the disease is important because, at that stage, there are alterations in HGM composition -or dysbiosis- that could explain the decrease in the appearance of the analyzed genetic markers. Other alterations associated with the progression of disease include intestinal transit alterations, decreased protein absorption, decreased consumption of dietary fiber, and frequent use of antibiotics. These accumulative factors will contribute to the production of systemic inflammation and the bioaccumulation of uremic toxins and presumably metal(loid)s, which are absorbed by the intestine and eliminated by the kidney, these toxicants may play a central role in the physiopathology of CKD [40].

A high presence of resistance genes for cadmium and arsenic metabolism were observed in analyzed samples (Fig. 5A). Cadmium (Cd) is a heavy metal, usually present in soils [41]. It is toxic to living organisms, is carcinogenic to humans [41,42,43,44], and lethal [43]. The human body can absorb Cd in small portions through food intake, especially in grains, water, or air [43], where it accumulates and remains for a long time, causing health problems [43,44,45,46,47]. Cd accumulates in the liver and kidneys and has a long biological half-life, 17–30 years in humans. Toxicity involves 2 organ systems, the renal and skeletal, and is largely the consequence of the interactions between Cd and essential metals, particularly calcium [43, 45, 47]. The severity and damage of these metals depend on time, level of exposure, and susceptibility of the person, and which metal is absorbed [43].

heat maps showing representative metal(loid)s resistance and antibiotic genes detected in samples from healthy versus CKD (all stages) subjects. A metal(loid)s resistance genes: cadA3k, arsC, cadA2k, arsA, merA and pbrA. For all cases analyzed, there was a tendency exhibit lower number of resistance genes to metal(loid)s, in particular, CKD patients tended to present a higher amount of cadA3k and then arsC compared to other resistance markers. In healthy patients, there was a tendency to present arsC and then cadA2k. B Antibiotic resistance genes observed were: arr2, catB4, qnrB1, strB, dhfr1, floR, tetA, ermB, acrB, and mefE1. The augmented tendency was observed for the dhfr1 gene in patients with CKD. Healthy patients tended to present arr2 and ermB genes, in slightly higher amounts compared to CKD patients. Groups were compared using Student t test, P < 0.0001 (***)

Mining activities in the second region of Chile represents one of the main commercial and labor activities of the country. This industry emits waste rich in arsenic and other minerals, which affects both the environment and residents of the area [48, 49]. Arsenic is found food, water, and air. The main damage produced by arsenic exposure is cardiovascular, kidney, neurological, respiratory, skin cancer, and reproductive effects [50].

High levels of the dhfr1 gene (dihydrofolate reductase) (Fig. 5B) can be observed on the heat diagram for both the healthy and CKD group, which corresponds to a resistance gene present in some Escherichia coli strains (https://www.uniprot.org/uniprot/Q0MQM2), the main causative microorganism of urinary tract infection (UTI) in healthy and CKD population [51].

To determine whether the observed changes in the content of antibiotic and metal resistance genes were significant between groups, a correlation (Pearson R) of Contingency of Prospective data (chi-square test) was performed and plotted by heat map (Additional file 4: Figure S4A) and principal component analysis PCA (Additional file 4: Figure S4B). Our results showed that healthy subjects (control), CKD3, and stage 4 CKD subjects are considered similar groups with a positive correlation. On the contrary, the stage 5 CKD group did not show a correlation with the previous groups. The results show that the gene content for resistance to antibiotics and metals is substantially lower than in the previous groups, Thus, further studies are needed to determine how late stages of CKD relate to a decrease in resistance markers.

The incidence of UTI in CKD patients increases with disease progression and occurs mainly due to the presence of risk factors. For example, uremia causes alterations in the humoral response, lymphocyte function, macrophages, and polymorphonuclear cells [51,52,53].

The underlying symptom of CKD is sometimes a condition that compromises normal elimination of urine and integrity of the urinary tract or implies its manipulation due to medical procedures (e.g., vesicoureteral reflux, neurogenic bladder, urethral valves, prostatism, bladder catheterization, renal catheterization, complicated lithiasis, and polycystic disease). In other cases, diabetes is an underlying symptom of both CKD and greater susceptibility to the appearance of UTI and its evolution and occurs especially in elderly female patients [51].

This study represents a slightly closer look at the association between the microorganisms present in the intestine and the evolution of a chronic disease. Even though the number of patients included in this study is small, it was possible to show an important modification in the content of microorganisms and resistance markers present in the microbiota. We also demonstrated the possibility of using microorganisms as biomarkers of a person’s health status or biosensors of any toxicant exposure through the study of microbial genetic markers. These genetic markers are indicative of disease progression and/or can indicate the type of metal(loid) or antibiotic to which the patient has been exposed due to genetic determinants in the microorganism of HGM. This description is particularly important in patients suffering an important decrease in one of the most important detoxification systems of the body, the liver.