Insight into blood proteinase-inhibitor system and pathogenesis of renal tuberculosis induced by phylogenomically different Mycobacterium tuberculosis strains in rabbit model

Abstract

Background

This study aimed to evaluate the impact of different Mycobacterium tuberculosis strains on the blood proteinase-inhibitor system and structural changes in the renal parenchyma during the pathogenesis of renal tuberculosis in a rabbit model.

Methods

Renal tuberculosis was modeled on 60 male Soviet Chinchilla rabbits. The susceptible virulent strain M. tuberculosis H37Rv (Euro-American lineage, group 1) and the low-lethal multidrug-resistant strain 5582 (Beijing Central Asian/Russian cluster; group 2) were injected into the cortex of the lower pole of the left kidney. Blood levels of biomarkers and enzymes were measured at baseline (pre-infection), and 2.5 and 22 weeks after infection. Morphological changes in nephron structures were assessed using 26 indicators at 22 weeks. Whole genome sequencing of M. tuberculosis DNA was performed on the DNBSEQ-G50 MGI platform.

Results

At 2.5 weeks, group 1 exhibited a significant increase in matrix metalloproteinases (MMP)-1/9 and cystatin C compared to group 2 (p = 0.02). After 22 weeks, group 1 showed elevated levels of MMP-9 and ceruloplasmin, alongside reduced levels of tissue inhibitor of metalloproteinases-1 (TIMP-1), cystatin C, and albumin (p = 0.02). Group 1 demonstrated a larger area of specific inflammation and less severe fibrotic changes compared to group 2 (p = 0.02). Genome of clinical strain 5582 harboured 55 frameshift and 8 stop codon mutations some of which were in genes known to be involved in intracellular survival and pathogenesis.

Conclusions

In quantitative terms, the structural changes observed in the kidneys of rabbits were inversely related to the virulence of the strains. Specifically, the more virulent strain (H37Rv) induces less pronounced structural changes. Renal tuberculosis induced by H37Rv is characterized by a pronounced imbalance in the MMP/TIMP-1 system, marked by increased MMP-1 and − 9 levels and decreased TIMP-1 levels in the blood. This imbalance is associated with structural kidney damage, including specific and paraspecific changes typical of an immediate hypersensitivity reaction. In contrast, infection with Beijing 5582 maintained a relative balance in the MMP/inhibitor system, with a significant increase in cystatin C and moderately pronounced productive changes in the renal parenchyma, consistent with a delayed hypersensitivity reaction.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12882-025-04411-w.

Background

Renal tuberculosis (RT) is a renal infection caused by Mycobacterium tuberculosis, with the kidney affected in approximately 10% of extrapulmonary tuberculosis cases [1]. As a global public health concern, renal pathology affects over 750 million people worldwide [2–4]. RT is a local manifestation of tuberculosis, primarily of hematogenous origin, beginning with tubercular lesions in the cortical layer of the renal parenchyma and periglomerular zone [3]. Inflammatory kidney diseases are associated with dysregulation of extracellular and intracellular proteolysis. Progressive renal injury often manifests as tubulointerstitial disease, characterized by tubular atrophy, hypoxia, peritubular capillary damage, and interstitial fibrosis [5]. Emerging evidence suggests that various classes of proteases contribute to structural changes in the nephron during inflammation and fibrosis, playing a critical role in the pathogenesis of kidney diseases regardless of etiology [6, 7].

Matrix metalloproteinases (MMPs), a family of zinc-dependent endopeptidases, are capable of degrading a wide range of extracellular matrix components and other protein substrates. This family includes interstitial collagenases (MMP-1, -8, and − 13), gelatinases (MMP-2 and − 9), stromelysins (MMP-3, -7, and − 10), and others. MMP expression is regulated by cytokines and growth factors, while their activity is modulated by tissue inhibitors of metalloproteinases (TIMPs) and serine proteinases such as neutrophil elastase [8]. MMPs are involved in numerous physiological and pathological processes, including inflammation, wound healing, and fibrosis, where tissue remodeling plays a central role. In kidney diseases, MMPs contribute to pathophysiological processes such as degradation of the tubular basement membrane, matrix deposition, vascular calcification, and fibrosis [9].

The spatial expression of MMPs and TIMPs in the kidney is complex and not yet fully understood. MMPs are primarily localized in glomeruli (podocytes), tubular epithelial cells, and collecting ducts [10, 11]. The expression level of MMP-9 correlates with the severity of renal failure, and the MMP-9/TIMP-1 ratio serves as a diagnostic marker for acute kidney injury. MMP-9 also enhances inflammatory processes by exerting chemotactic effects on dendritic cells [12]. However, the role of MMP-9 in renal fibrosis progression remains controversial. While MMPs are implicated in all phases of renal fibrosis—from mononuclear cell infiltration to proliferation and scar formation [13]—some studies report increased MMP-9 activity in early-stage chronic renal failure but decreased activity in later stages [14]. Additionally, increased MMP-9 gene expression has been linked to epithelial-mesenchymal transition in renal tubular cells, leading to interstitial fibrosis [15]. Conversely, other studies associate fibrotic changes with elevated expression of ICAM-1, collagen types III and IV, transforming growth factor-beta 1 (TGF-β1), reduced gelatinase activity, and increased TIMP-1 levels [16, 17].

Cystatin C, a low-molecular-weight protein secreted by all nucleated cells, acts as an endogenous inhibitor of cysteine proteinases and lysosomal cathepsins, thereby mitigating the destructive effects of proteases on the extracellular matrix. Its levels fluctuate in response to inflammation, immune disorders, and malignancies [18]. Clinically, cystatin C is a reliable serum marker of renal failure (glomerular filtration rate), independent of age, gender, or muscle mass. Its concentration increases approximately twofold in various renal dysfunctions [19]. Unlike creatinine, cystatin C is freely filtered by glomeruli, reabsorbed, and catabolized by proximal tubular cells, making it a more stable marker of renal function [20, 21].

Despite the critical role of proteases in renal pathology, studies evaluating the proteinase-inhibitor system in RT remain limited. In the previous rabbit model study, we demonstrated that RT caused by a multidrug-resistant (MDR) M. tuberculosis Beijing strain disrupts the balance of the MMP/inhibitor system in the bloodstream [22]. A key area of interest is the correlation between laboratory markers of collagen metabolism, renal function, and morphological changes in nephron structures, which may vary depending on the biological properties of the pathogen. Interpretation of renal dysfunction through MMP changes in the blood is supported by Hsu et al. [23], who suggested that circulating levels of MMPs likely reflect their expression in the whole body as well as in the kidney. Also, Lods et al. [24] showed that serum MMP-2 and − 9 levels in patients with glomerulonephritis were consistent with their expression in renal biopsy tissue.

The pathobiological features of M. tuberculosis strains, particularly their genomic characteristics, significantly influence disease development at the individual level and shape the epidemic process at the population level in humans. Strains from different phylogenetic lineages and genotypes exhibit varying virulence in animal and macrophage models, as well as differing capacities to acquire antibiotic resistance. In Russia, the M. tuberculosis population is dominated by strains of the Beijing genetic family (40–80%) [25–28]. Strains of the Central Asian/Russian subtype of the Beijing genotype are the most common branch in Russia and are relatively less frequently MDR compared to strains of the Russian successful clonal cluster Beijing B0/W148. Beijing genotype strains differ in their virulence, and some of them are either more or less virulent compared to the reference strain H37Rv [29]. Mouse model studies have shown that H37Rv is more virulent and lethal than LAM strains of Lineage 4, while drug-sensitive LAM strains are more virulent and lethal than MDR LAM strains of different spoligotypes [30]. Interestingly, no correlation was observed between MDR status and virulence/lethality in mouse models: two MDR strains of the ancient Beijing sublineage differed markedly in virulence, with H37Rv occupying an intermediate position [29].

In this study, we employed a rabbit model to evaluate changes in the proteinase-inhibitor system and structural alterations in the renal parenchyma during the pathogenesis of RT induced by phenotypically and genotypically distinct M. tuberculosis strains.

Materials and methods

Bacterial strains

M. tuberculosis strains used in this study were obtained from the collection of the St. Petersburg Research Institute of Phthisiopulmonology. The drug-sensitive reference strain H37Rv was acquired in 2013 from the Federal Scientific Centre for Expertise of Medical Products, Moscow, Russia. Originally, it was obtained from the Institute of Hygiene and Epidemiology in Prague in 1976.

Two M. tuberculosis strains were used in this study: drug-sensitive reference strain H37Rv and MDR clinical strain Beijing 5582. Strain H37Rv belongs to the Euro-American phylogenetic lineage (Lineage 4, sublineage 4.9). H37Rv is a virulent strain, exhibiting greater virulence than some Lineage 4 strains (e.g., LAM) but also demonstrating higher virulence and lethality than certain Beijing strains, as previously shown in mouse model studies [29–31]. Clinical strain 5582 was isolated in 2012 from a purulent specimen of a patient with tuberculous spondylitis in Buryatia, Russia. Strain 5582 was chosen for its phylogenetic and phenotypic representativeness. Strain 5582 belongs to the Central Asian/Russian subtype of the Beijing genotype, specifically clade A [32–34] or the L2.2.M4.9.2 sublineage according to the SNP barcode system [35]. Previous studies in a C57BL/6 mouse model of experimental tuberculosis demonstrated that strain 5582 (designated as CladeA strain in that study) exhibited the lowest lethality [31].

WGS and bioinformatics

Genomic DNA from M. tuberculosis was isolated using a standard protocol involving SDS, proteinase K, and cetyltrimethylammonium bromide for cell lysis. Whole genome sequencing (WGS) was performed on the DNBSEQ-G50 platform (MGI, China).

WGS fastq files were mapped to the reference genome H37Rv (NC_000962.3) using the SAM-TB online tool (https://samtb.uni-medica.com/index) and additionally checked with the Geneious R package (Biomatters, New Zealand). The WGS data (fastq files), were deposited in the NCBI Sequence Read Archive (SRA) under accession numbers SRR32263914 and SRR18572457.

Animal model

All procedures involving laboratory animals adhered to relevant biosafety requirements and were conducted in accordance with Russian national guidelines (“Rules for Working with Laboratory Rodents and Rabbits,” 2016). All experimental protocols were approved by the Independent Ethics Committee of the St. Petersburg Research Institute of Phthisiopulmonology (Protocol No. 80, dated June 23, 2021) and complied with the Resolution of the State Sanitary Doctor of the Russian Federation No. 4 (dated January 28, 2021), “On Approval of Sanitary Rules and Regulations SanPiN 3.3686-21: Sanitary and Epidemiological Requirements for the Prevention of Infectious Diseases”. The research was conducted in accordance with the Declaration of Helsinki.

The study utilized 60 male Soviet Chinchilla rabbits weighing 3.54 ± 0.26 kg, obtained from the Federal State Unitary Enterprise “Rappolovo Laboratory Animal Nursery.” The animals were kept in a certified vivarium at the St. Petersburg Research Institute of Physiopulmonology, from January 24, 2022, to June 27, 2022. Each rabbit was individually housed with food and water consumption ad libitum.

The study modeled renal tuberculosis in two experiments conducted over 22 weeks, following a previously established protocol [36]. In experimental tuberculosis studies using rabbits, sample sizes typically range from 4 to 18 animals per group. The specific number can vary depending on the experimental design, such as the number of treatment groups or time points for analysis [37–39]. In this view, our study with two groups (30 animals each) had sufficiently large sample size.

Thirty rabbits were assigned to each of two groups: Group 1 received infection with the H37Rv strain, while Group 2 was infected with the 5582 strain. A bacterial suspension (10⁶ microbial cells in 0.2 mL saline) was injected into the cortical layer of the left kidney’s lower pole via fine-needle puncture. Prior to infection, anesthesia was induced using an intramuscular injection of Zoletil [zolazepam + tiletamine (Virbac, France)] at 25 mg/kg body weight, combined with a 2% Xyla solution (Xylazinum, Interchemie Werken “de Adelaar” BV, the Netherlands; 1.0–1.5 mL). Post-infection antibiotic prophylaxis consisted of cefazolin (50,000 U/kg, 1.5 mL intramuscularly) administered over five days.

Renal tuberculosis progression was evaluated 18 days post-infection. First, a delayed-type hypersensitivity test was performed using the recombinant tuberculosis allergen Diaskintest^® (Generium, Moscow, Russia). The test was administered intradermally (0.1 mL of 2 µg/mL saline solution) on the back, corresponding to the kidney projection. Subsequently, an abdominal CT scan was conducted using a Toshiba One Aquilion tomograph. For contrast imaging, Ultravist-370 was injected as a bolus into the marginal ear vein at 1 mL/min (total volume: 4 mL). Sequential abdominal images were captured within 30 s, enabling perfusion map generation.

Biochemical methods

Blood serum from the rabbits was analyzed to determine the concentrations of key components of the proteinase/inhibitor system. The levels of various biomarkers and enzymes were measured in the blood at baseline (before infection), after 2.5 and 22 weeks. NE activity was measured as described previously [40]. Matrix metalloproteinases MMP-1 (SEAO97Rb) and MMP-9 (SEA556Rb) and tissue inhibitor of metalloproteinases-1 (TIMP-1, SEA 552Rb), Cystatin C (SEA896Rb) were quantified using solid-phase enzyme immunoassay kits (Cloud-Clone Corp, USA), while albumin, creatinine, and urea concentrations were measured using a SynchronCX5 PRO biochemical analyzer (Beckman Coulter, USA). Ceruloplasmin (CP) levels were assessed using the Ravin method, and adenosine deaminase (ADA) activity was measured to evaluate purine metabolism [41].

Histological methods

After 22 weeks of experiment, the animals were euthanized by an overdose of sodium thiopental 250 mg (Sintez, Kurgan, Russia) and pipecuronium bromide 1 mg (Veropharm, Moscow, Russia) injected into the marginal ear vein. We followed the established protocol that we also employed in our previous study [36, 42].

Kidney tissues were fixed in 10% neutral formalin, embedded in paraffin, and sectioned for staining. Sections were stained with hematoxylin and eosin (H&E) for general histology, Masson’s trichrome to assess connective tissue maturity, and Ziehl-Neelsen staining to detect acid-fast bacteria. Digital images of the stained sections were acquired using a Panoramic scanner and analyzed using Orbit Image Analysis and PanoramicViewer version 3.64 software (National Institutes of Health, USA). Morphometric parameters evaluated included the width of the renal cortex and medulla, the area of specific inflammation, inflammatory infiltration, and fibrosis, as well as structural parameters of renal corpuscles and the renal vascular system.

Statistical analysis was performed after confirming normal distribution using the Shapiro-Wilk test. Data are presented as medians with lower and upper quartiles (Me, Q1; Q3). Differences between groups were assessed using the Mann-Whitney U-test, and correlations were analyzed using Spearman’s method.

Results

M. tuberculosis strains: phylogenomic analysis

No drug resistance mutations were detected in the H37Rv strain, as expected. In contrast, the genome of strain 5582 contained four high-confidence resistance mutations: rpoB S450L, katG S315T, rrs C517T, and pncA T135P, which confer resistance to rifampicin, isoniazid, streptomycin, and pyrazinamide, respectively. A сompensatory mutation was detected in rpoC Glu1092Asp.

Strain 5582 differed from the reference genome H37Rv (NC_000962.3) in 1315 positions, as determined using the SAM-TB resource, whereas hypervariable PE/PPE loci were excluded.

The genome of strain H37Rv used in this study is not identical to the reference genome H37Rv (NC_000962.3), which is commonly used for genomic comparisons and mutation mapping. The H37Rv strain was originally isolated in the early 20th century as the H37 strain from a patient in the United States [43]. Over time, laboratory-derived variants accumulated specific mutations during their propagation in different laboratories. Not surprisingly, strain H37Rv used in our study differed slightly, in 41 genomic positions, from the reference genome H37Rv.

Rabbit model experiment and analysis of biochemical markers

The rabbit model experiment lasted 22 weeks, whereas the levels of various biomarkers and enzymes were measured in the blood before infection with M. tuberculosis strains, after 2.5 and 22 weeks. Renal tuberculosis (RT) in rabbits was confirmed 2.5 weeks post-infection using Diaskintest^® (erythema diameter: 2.30 [1.50; 2.70] mm in group 1 and 2.8 [2.20; 2.95] mm in group 2), the presence of M. tuberculosis growth in urine cultures, and ultrasound examination of the kidneys.

By 2.5 weeks post-infection, rabbits in Group 1 (H37Rv-infected) showed a marked rise in peripheral blood levels of matrix metalloproteinases (MMP-1, MMP-9) and cystatin C compared to baseline (Table 1). In contrast, tissue inhibitor of metalloproteinases-1 (TIMP-1) concentrations remained unchanged in both groups. Chronic inflammation triggers neutrophil activation, leading to the release of neutrophil serine proteases such as NE [44]. NE, which degrades extracellular matrix components (e.g., fibrin, fibronectin, and elastin), exhibited a significant increase in activity—37% in Group 1 and 55% in Group 2 (Table 1).

Table 1.

Dynamic changes in markers of the blood proteinase/inhibitor system during the progression of renal tuberculosis, me, Q1 - Q3

Analyzed parameters	Before infection (baseline)	Time-points after infection
		2.5 weeks		22 weeks
		Group 1 (H37Rv)	Group 2 (Beijing 5582)	Group 1 (H37Rv)	Group 2 (Beijing 5582)
MMP-1, ng/ml	1.52 [0.97;3.40]	2.63 [1.20;3.94] *р=0.022	1.70 [1.00;2.64]	1.40 [1.20;2.86]	1.29 [0.89 ;2.94]
MMP-9, ng/ml	230.70 [164.30;341.56]	398.25 [323.74;516.30] *р=0.022	294.77 [170.59;431.54]	693.38 [548.25;805.29] *р=0.022	519.61 [444.32;561.65] *р=0.011
TIMP-1, ng/ml	23.59 [12.66;33.13]	29.52 [11.89;33.90 ]	25.21 [13.32;32.06 ]	2.97 [1.89;7.56] *р=0.022	18.97 [13.05;30.13 ]
Neutrophil elastase, U	237.20 [172.00; 293.40]	326.00 [195.60 ;355.00 ] *р=0.022	369,50 [239.10;521.60] * р=0.041	376.00 [347.70;405.70] *р=0.022	282.50 [195.60;360.70]
Cystatin С, ng/ml	1197.91 [849.02;2115.09]	1954.01 [1222.54;3398.00 ] *р=0.022	1135.37 [731.74;1878.21]	2139.72 [1875.24;2600.51]	3282.41 [1726.61;5774.33] *р=0.022

Note. ММР-1,9 – matrix metalloproteinase, ТИМП-1 – tissue ingibitor-1

* - significant difference compared to baseline (Mann-Whitney U test)

Under conditions of inflammation and tissue damage, extracellular adenosine levels rise, serving as an anti-inflammatory purine regulator [45]. Changes in adenosine levels were assessed by measuring adenosine deaminase (ADA) activity. Both groups showed a significant increase in ADA activity, alongside elevated levels of ceruloplasmin, another marker of inflammation. No changes in creatinine or urea concentrations were observed in either group compared to baseline levels (Table 2).

Table 2.

Dynamic changes in markers of kidney function and inflammatory response during the progression of renal tuberculosis (Me [Q1; Q3])

Analyzed parameters	Before infection (baseline)	Time-points after infection
		2.5 weeks		22 weeks
		Group 1 (H37Rv)	Group 2 (Beijing 5582)	Group 1 (H37Rv)	Group 2 (Beijing 5582)
Creatinine, µmole/L	81.00 [69.00;87.00]	79.00 [69.00;85.00]	82.00 [71.50;88.00]	68 [62.00;72.00]	87.30 [75.00;98.00]
Urea, µmole/L	6.67 [5.17;7.44]	5.89 [5.70;6.60]	5.15 [4.85; 5.45]	5.16 [5.11;6.31 ]	6.50 [4.80;6.60]
Albumin, g/L	46.00 [44.00;51.00]	45.00 [43.00;47.00 ]	47.00 [45.00;49.00]	43.00 [40.00;43.00]	47.00 [45.00;48.00]
Ceruloplasmin, g/L	0.24 [0.22;0.35]	0.41 [0.34;0.46] *р=0.022	0.42 [0.36 ;0.53] *р=0.022	0.25 [0.22 ;0.31]	0.47 [0.19;0.58]
Adenosine deaminase, U/L	7.20 [4.4;10.6]	21.20 [20.10;23.90] *р=0.021	24.10 [16.05 ;30.65] *р=0.021	14.3 [12.80;17.50] *р=0.021	26.20 [16.80;29.60] *р=0.021

* significant difference compared to baseline (Mann-Whitney U test)

Thus, group 1 (H37Rv) was characterized by a pronounced imbalance in the MMP/inhibitor system, marked by increased proteolytic activity (elevated MMP-1 and MMP-9 levels) and a rise in the renal function marker cystatin C, accompanied by a decrease in animal body weight. Both groups exhibited similar increases in ADA, NE, and ceruloplasmin levels, with no changes in creatinine or urea concentrations.

After 22 weeks post-infection, an imbalance in the MMP/inhibitor system was observed in both groups, though it was more pronounced in group 1 (p = 0.02) (Table 1). A three-fold increase of MMP-9 and a five-fold decrease of TIMP-1 versus a two-fold increase of MMP-9 and no significant changes in TIMP-1 were observed in groups 1 and 2, respectively. The concentration of cystatin C in group 2 (strain 5582) exceeded that in group 1. No significant differences in MMP-1 levels were observed between the two groups. Group 1 was characterized by higher levels of NE (p = 0.04) and ceruloplasmin (p = 0.04), as well as a lower level of albumin (p = 0.02) and reduced body weight, which reached 2965 (2900; 3100) g compared to 3956 (3904; 4130) g in group 2 (Table 2). Both groups exhibited similar trends in increased ADA activity, while urea and creatinine concentrations remained within baseline values (Table 2).

Normally, proteases and inhibitors represent a well-balanced system, whereas under pathological conditions, the protease-inhibitor balance shifts towards an increase in the activity of proteolytic enzymes. Thus, in group 1 (H37Rv), the development of a specific inflammatory process at both study periods was characterized by more pronounced proteolysis due to an uncompensated increase in MMP-9 levels. Elevated MMP-9 in the kidneys may exacerbate inflammation by generating collagen fragments that attract neutrophils and stimulate the release of proteinases [9]. In contrast, group 2 (Beijing 5582) exhibited a less pronounced imbalance in the MMP/inhibitor system but showed a significant increase in the cysteine proteinase inhibitor cystatin C by the end of the study period.

The progression of renal tuberculosis in both groups was accompanied by a comparable increase in NE activity. These findings suggest that the features of extracellular matrix remodeling in renal tuberculosis are influenced by the pathobiological properties of the infecting M. tuberculosis strains.

Histological examination

Histological examination of kidney biopsies 22 weeks post-infection revealed the development of a specific inflammatory process in 100% of cases in both groups. Lesions were primarily localized in the deep layers of the medulla and the renal pelvis area. The area of damage in group 1 (infected with strain H37Rv) significantly exceeded that in group 2 (infected with strain 5582) (Table 3).

Table 3.

Histological assessment of the functional structures of the kidneys in renal tuberculosis model 22 weeks after infection (Me [Q1; Q3])

Analyzed parameters	Group 1 (H37Rv)	Group 2 (Beijing 5582)
Specific inflammation area, mm2	40.60 [39.00;42.00]	30.6 [19.60;39.70] *р=0.042
Proportion of interstitial inflammation in the cortex, %	7.10 [6.80;7.20]	1.60 [1.00;2.70] *р=0.042
Proportion of interstitial inflammation in the medulla, %	5.50 [4.90;5.70]	6.10 [2.70;12.00]
Collagen area in the cortex, %	6.70 [4.90;8.00]	25.00 [5.00;31.00] *р=0.042
Collagen area in the medulla, %	27.80 [25.00;31.00]	45.00 [33.00; 54.00] *р=0.041
Parietal leaf of the Shumlyansky-Bowman capsule, µm	3.90 [3.60;4.20]	4.10 [3.60;4.50]
Area of the glomerulus with capsule, µm2	10,525 [90720;11040]	10,854 [9197;12030]
Area of the urinary space, µm2	2629 [2300;2700]	2594 [1726;2858 ]

* significant difference between two groups

The groups differed in the ratio of nonspecific to specific components of inflammation (Table 3). Group 1 exhibited more pronounced diffuse interstitial inflammation, characterized by predominantly lymphocytic infiltration of the stroma, which was 4.5 times greater than in group 2. This suggests a longer and more intense development of nonspecific inflammation in group 1. Consistent with this, group 1 showed less intense and less mature fibrotic changes in the renal cortex outside the zones of specific damage (3.7 times less than group 2), indicating delayed and less pronounced fibroplastic activation of interstitial cells in the perifocal zones.

Thus, group 1 (strain H37Rv) demonstrated a less specific but more pronounced inflammatory reaction, which was localized and associated with minimal changes in the surrounding renal cortex. This pattern may help preserve many of the organ’s functions at this stage. In contrast, group 2 (strain 5582) was characterized by smaller areas of necrosis but more specific cellular reactions, reflecting a competent immune response. This group also exhibited more pronounced protective compensatory reactions extending beyond the zones of specific damage, potentially affecting the functional capacity of the kidney.

The groups did not differ in the nature of structural changes in the renal corpuscle (Table 3). Both groups showed an increase in the thickness of the parietal layer of the Bowman’s capsule, an expansion of the urinary space, and a reduction in the area of the renal corpuscle. Despite similar glomerular cellularity (52.30 [51.00; 53.00] n in group 1 and 50 [49; 50] n in group 2), the diameter of glomerular capillaries differed significantly (9.80 [8.90; 10.00] µm in group 1 vs. 6.10 [5.60; 6.60] µm in group 2). This suggests that the increase in the area of the renal corpuscle in group 2 is due to interstitial changes, which may impair the kidney’s filtration capacity.

We established a relationship between increased MMP concentrations and morphological changes at 22 weeks after infection. In group 1 (H37Rv), a positive correlation was observed between MMP-9 levels and the volume of specific damage (r = 0.58, p = 0.006). In group 2 (strain 5582), the development of inflammation was associated with increased cystatin C levels and a positive correlation with fibroplastic reactions (r = 0.60, p = 0.004).

Discussion

Comparing the progression of renal tuberculosis in a rabbit model caused by different M. tuberculosis strains under identical infection doses and conditions allows to associate the observed biochemical and structural differences with the phenotypic and genotypic characteristics of the pathogen. While H37Rv is a laboratory-maintained virulent strain, clinical strain 5582 belongs to the Beijing genotype. Previous studies in a mouse model of experimental tuberculosis demonstrated that these strains induce similar changes in lung damage indices, but the bacterial load was consistently higher in H37Rv at all time points [31] (Supporting Figure S1). Compared to other multidrug-resistant (MDR) strains of the Beijing genotype, strain 5582 exhibited the lowest lethality in mice [29, 31].

In our study, infection of rabbits with either M. tuberculosis strain resulted in similar structural changes (remodeling) in the kidneys, consistent with membranous glomerulonephritis, albeit with varying severity. Renal tuberculosis induced by H37Rv was characterized by a pronounced imbalance in the MMP/inhibitor system, marked by increased concentrations of metalloproteinase-1 and − 9 and decreased levels of tissue inhibitor of metalloproteinases-1 in the blood. In contrast, renal tuberculosis caused by the Beijing 5582 strain was associated with a relative balance in the MMP/inhibitor system and a significant increase in cystatin C levels in the blood. In both groups, biochemical changes were accompanied by structural changes characteristic of hypersensitivity reaction Type IV.

Regarding the qualitative characteristics of the inflammatory process, infection with the virulent strain H37Rv resulted in glomerulonephritis characterized by hyperergic nonspecific changes, predominantly alterative-infiltrative in nature. In contrast, infection with the low-lethal strain Beijing 5582 led to a predominantly productive reaction with fibrosis. The observed biochemical changes in the blood and structural changes in the kidneys likely reflect different pathogenesis mechanisms in experimental renal tuberculosis caused by genetically and phenotypically different strains H37Rv and Beijing 5582.

Our results corroborate with the published data indicating that the prognosis of a specific infectious process depends not only on the drug resistance profile of the strain but also on its virulence [29, 30, 46–49]. Furthermore, the development of glomerulosclerosis and vascular remodeling in renal tuberculosis, as in nephropathies of other etiologies, involves the participation of various classes of proteinases which play crucial roles in tissue damage, inflammation, and fibrosis [6, 8, 10, 11, 50].

Similar results were reported by Krylova et al. [51] for another low-virulent MDR Beijing genotype strain, 6691, when compared to H37Rv. Phylogenetically, strains 5582 and 6691 belong to evolutionarily distant branches within the Beijing genotype—modern and early ancient, respectively. This highlights the heterogeneity of virulence properties within M. tuberculosis and suggests that low virulence and low immune reactivity traits have emerged independently at different stages of the pathogen’s evolutionary history.

In quantitative terms, the structural changes observed in the kidneys of rabbits were inversely related to the virulence of the strains. Specifically, the more virulent strain (H37Rv) induced less pronounced structural changes. This partly aligns to the published studies although the underlying reasons are elusive and may be different in each specific case.

While M. bovis infection typically causes chronic, progressive disease with cavitary lesions in rabbits, certain M. tuberculosis strains—especially those with higher virulence or specific genetic backgrounds such as the Erdman strain—can present similar, though generally milder, pathological features. Manabe et al. [52] conducted aerosol infection experiments in rabbits using three Euro-American lineage strains of M. tuberculosis: Erdman (Haarlem genotype), H37Rv, and CDC1551. Among these, the Erdman strain demonstrated the highest virulence in rabbits, requiring the fewest inhaled bacilli to form a single tubercle. Infection with Erdman led to coalescing lesions and occasional cavities at lower doses of inhaled bacteria compared to H37Rv. Genetic analysis revealed differences in the RD6 region—a region rich in PPE genes—between H37Rv and Erdman, with this deletion region also varying significantly among M. tuberculosis strains and being absent in many M. bovis strains [53].

In a rabbit model of tuberculous meningitis, the disease severity caused by clinical M. tuberculosis isolates—CDC1551 (Euro-American lineage), Beijing genotype strains HN878 and W4—and the laboratory strain H37Rv was assessed [54]. Both Beijing strains exhibited greater virulence than CDC1551 and H37Rv controls. Notably, the hypervirulence of HN878 was linked to phenolic glycolipid (PGL) production, a feature shared with Beijing strain W4 but absent in strains of the Euro-American lineage CDC1551 and H37Rv [55]. Compared to CDC1551 infections, central nervous system infections with Beijing strains resulted in higher bacterial loads in cerebrospinal fluid and brain tissue, increased dissemination to other organs, and more severe clinical signs. Rabbits infected with an HN878 mutant deficient in the pks1-15 gene - which is essential for PGL synthesis - exhibited reduced pathological damage and milder clinical symptoms relative to those infected with CDC1551.

Indeed previous tuberculosis research demonstrated that gene-inactivating mutations, including large genomic deletions and frameshift mutations, can significantly influence the pathogenic potential of M. tuberculosis strains, affecting drug resistance, fitness, and interaction with the host. These mutations can lead to altered drug susceptibility, such as increased resistance to isoniazid due to katG gene deletion [56]. Beyond drug resistance, gene disrupion, including phase variation, functional genetic variations in PE/PPE and other genes contribute to the diversification of M. tuberculosis lineages and their interactions with the human host [57–61].

In our study, Beijing strain 5582 was less-virulent compared to the reference virulent strain H37Rv. The genome of strain 5582 harboured 55 frameshift and eight stop codon mutations compared to H37Rv (Supplementary Table S1). As noted above, strain 5582 belongs to the most prevalent in Russia branch of the Beijing genotype - Beijing Central Asian/Russian. Therefore, we further compared its genome with those of Russian strains of different genotypes (project PRJNA305488 in NCBI SRA). This comparison revealed that most of these mutations (51 of 55 frameshift mutations and all eight stop codon mutations) were shared by strain 5582 and other Beijing Central Asian/Russian (CAR) strains. Notably, four of these mutations were in oxidoreductase genes potentially involved in intracellular survival, virulence, and pathogenesis; two were in transcriptional regulators; one was in a PE gene (PE35 [62]); and others were in genes related to virulence and adaptation (mce3F, pks6, mmpS5, mmpL9, vapB17/vapB18, fadE36) (according to mycobrowser.epfl.ch). Hypothetically, these gene/protein-inactivating mutations could correlate with immunogenic and pathogenic properties of strain 5582. These findings corroborate previous results from the C57BL/6 mouse model of experimental tuberculosis, which demonstrated that this strain 5582 elicits a low-reactivity immune response [31].

Examining the correlation between relative pathogenicity, including disease severity in the rabbit model, and the strain genotype may aid in pinpointing specific M. tuberculosis genes that play a critical role in disease progression in humans. Nonetheless, the above considerations are speculative, and the actual impact of the specific gene-inactivating mutations found in the studied Beijing strain requires further investigation through allele replacement experiments and knockout mutant studies.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

RT

Renal tuberculosis

MDR

Multidrug resistance

MMP

Matrix metalloproteinases

TIMP

1–tissue inhibitor of metalloproteinases–1

CR

Creatinine

NE

Neutrophil elastase

ADA

Adenosine deaminase

AL

Albumin

CP

Ceruloplasmin

Author contributions

D.E. and T.V. designed the study and supervised it. D.E., I.M., N.A., N.B., M.Dy., M.Do., A.V., D.P., and T.V. performed experiments and data analysis. D.E., T.V., B.A., and I.M. drafted the manuscript. P.Y. implemented the project, provided resources, and edited the manuscript. All authors have read and approved the final version of this manuscript.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

All data of this study are presented in the article and supplementary material. The datasets generated and/or analyzed during the current study are available in the NCBI Sequence Read Archive repository, under accession numbers SRR32263914 and SRR18572457.

Declarations

Ethics approval and consent to participate

All experimental protocols were approved by the Independent Ethics Committee of the St. Petersburg Research Institute of Phthisiopulmonology (Protocol No. 80, dated June 23, 2021) and complied with the Resolution of the State Sanitary Doctor of the Russian Federation No. 4 (dated January 28, 2021), “On Approval of Sanitary Rules and Regulations SanPiN 3.3686-21: Sanitary and Epidemiological Requirements for the Prevention of Infectious Diseases”. The research was conducted in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

Igor Mokrousov is Senior Editorial Board Member in another BMC Series journal (BMC Microbiology). Other coauthors declare that no conflicts of interest exist.