Tissue specific diversification, virulence and immune response to Mycobacterium bovis BCG in a patient with an IFN-γ R1 deficiency

Figures & data

Table 1. Patient treatment and timeline of chronic BCG-osis at NIH

Age (y)	Culture site	Smear	Culture	Imaging-brain	Imaging-lung	Antibiotics	Cytokine adjuvants
3	Scalp	Many	Positive	Multiple enhancing lesions	Pleural mass	Clari, INH, RIF	None
3	Lung	Many	Positive	Multiple enhancing lesions	Pleural mass	Clari, INH, RIF	None
3.2	Scalp	Negative	Positive	Stable	Stable	Clari, INH, RIF, ETH, AMI, Levo	IFNγ
3.3	Scalp	Few	Negative	Stable	Increased size	Azi, INH, RIF, ETH,AMI	IFNγ
3.3	Lung	Moderate	Liquid media only	Stable	Increased size	Azi, INH, RIF, ETH,AMI	IFNγ
3.7	Scalp	Negative	Negative	Stable	Increased size	Azi, INH, ETH, Levo, LZD, Clofaz, Capreo	IFNγ and IFNα
3.8	Scalp	Negative	Negative	Stable	Stable	Azi, INH, ETH, Levo, LZD, Clofaz, Capreo	IFNγ and IFNα
4.0	Lung	Negative	Negative	Stable	Increased size	Azi, INH, ETH, Levo, LZD, Clofaz	IFNγ and IFNα

Clari, clarithromycin; INH, isoniazid; RIF, rifampin; ETH, ethambutol; AMI, amikacin; Levo, levofloxacin; Azi, azithromycin; LZD, linezolid; Clofaz, clofazimine; Capreo, capreomycin.

Figure 1. Evaluation of mycobacterial Isolates: (a) Difference in colony morphology of Mycobacterium bovis isolates within the BCG vaccinated patient with IFNγR1 deficiency compared to the vaccine Mycobacterium bovis BCG isolate. Colonies of the BCG strains at 3 weeks growth in Middlebrook 7H11 plate (100× magnification, Bar = 0.5 inch). (b) Quantification of colony-forming units (CFUs); from the lysate of peripheral mononuclear cells at d 2 of infection by the 3 strains. Data, mean ± SE (n = 6, **p < 0.01; *p < 0.05; two-tailed paired Student’s t-test). (c) Monocyte-derived macrophages were infected with GFP-transfected strain; at MOI of 2 for 2 d. Slides were made with DAPI staining for nucleic acid content. Representative infected cell of each strain, with also gray color as visible light showing bacteria inside of the limits of the cell (bar = 5 μm), with field showing more cells on smaller scale (bar = 100 μm). (d) Core genome phylogenetic analysis of 14 genomes; 3 BCG strains (BCG-vaccine, BCG-brain, and BCG-lung), 5 representative genomes of M. tuberculosis (H37Rv, CDC1551, F11, KZN 1435, and KZN 4207), and 6 M. bovis isolates (BAA_935, Tokyo 172, BCG1 Russia, BCG Pasteur 1173P2, BCG Danish 1331 Belgium, and BCG Denmark)

Table 2. Antibiotic MIC values of vaccine, lung, and brain Mycobacterium bovis BCG strains at 14 d. Values in bold represent MIC values associated with mutations as discussed in the text

MIC (µg/ml)	BCG-Vaccine	BCG-Brain	BCG-Lung
Streptomycin	0.39	0.19	0.19
Kanamycin	1.5	0.78	0.78
Amikacin	0.19	0.19	0.09
Ethambutol	1.5	1.5	1.5
Isoniazid	0.19	0.04	0.09
Rifampin	<0.001	>=1.2	0.14
Rifapentine	<0.001	>=0.1	0.06
Capreomycin	0.78	0.39	0.39
Levofloxacin	0.19	0.19	0.19
Moxifloxacin	0.09	0.04	0.04
Ciprofloxacin	0.31	0.31	0.15
Clofazimine	0.19	0.78	3.12
Ethionamide	25	12.5	50
Linezolid	0.39	0.78	0.39
Clarithromycin	<0.02	0.19	0.04
Azithromycin	0.78	1.5	0.39

Table 3. Identification of mutations in the clinical isolates that have M. tuberculosis H37Rv homologs

H37Rv gene product name H37Rv locus tag	Function H37Rv locus tag	Gene product name BCG Danish 1331	BCG Danish 1331 locus tag and SNP position in genome	BCG-vaccine	BCG-brain	BCG-lung
Acetyl-CoA acetyltransferase fadA2 Rv0243	Lipid degradation	Acetyl-CoA C-acetyltransferase	FCU26_RS01340 291,101	BIS44_3264 Wild type	BIT18_1027 568A>C Thr190Pro	BIT17_0787 Wild type
Acyl-CoA dehydrogenase Rv0244c (probable fadE5)	Lipid degradation	Butyryl-CoA dehydrogenase	FCU26_RS01345 293,963	BIS44_3263 Wild Type	BIT18_1028 35 T > G Val12Gly	BIT17_0786 Wild type
Diacylglycerol O-acyltransferase Rv3087	Mycolic acid synthesis, virulence, lipid metabolism	Wax ester/triacylglycerol synthase family O-acyltransferase	FCU26_RS16610 3,387,640	BIS44_3892 Wild type	BIT18_3976 Wild type	BIT17_3018 413A>G His138Arg
RNA polymerase, beta subunit, rpoB Rv0667	RNA synthesis, Rifampin resistance	DNA-directed RNA polymerase subunit beta	FCU26_RS03640 Brain SNP 762,861 Lung SNP 762,912	BIS44_1904 Wild type	BIT18_3441 1304A>T Asp435Val	BIT17_4071 1355 T > C Leu452Pro
Heat-shock protein transcriptional repressor HrcA Rv2374c	Virulence, detoxification, adaptation	Heat-inducible transcriptional repressor	FCU26_RS12720 2,599,334	BIS44_2080 Wild Type	BIT18_1210 Wild Type	BIT17_4383 881 G > T Gly294Val
Probable conserved transmembrane transport protein MmpL5 Rv0676c	Fatty acid transport	Siderophore RND transporter MmpL5	FCU26_RS03690 777,504	BIS44_1915 Wild type	BIT18_3430 2728A>C Ile910Leu	BIT17_4082 Wild type
Hypothetical protein Rv0678	Putative Regulatory function, Clofazimine resistance	Hypothetical protein	FCU26_RS03700 781,024	BIS44_1917 Wild type	BIT18_3428 Wild type	BIT17_4084 BIT17_4085 (284 T > A) Leu95* Point mutation (amber)
Amidophosphoribosyltransferase purF Rv0808	De novo purine biosynthesis	Amidophosphoribosyltransferase	FCU26_RS04420 904,098	BIS44_0866 Wild type	BIT18_0945 Wild type	BIT17_1246 1235 G > A Arg412His
Hypothetical protein Rv0877	Unknown	DUF3027 domain-containing protein	FCU26_RS04830 977,277	BIS44_2274 Wild type	BIT18_0488 BIT18_0489 Frameshift	BIT17_3190 Wild type
Hypothetical protein Rv2164c	Unknown	Hypothetical protein	FCU26_RS11595 2,387,417	BIS44_0261 Wild type	BIT18_0263 Wild type	BIT17_3438 695A>G Glu232Gly
Hypothetical protein Rv2267c	Unknown	Sulfotransferase	FCU26_RS12145 2,500,583	BIS44_2709 779A>G His260Arg	BIT18_3241 779A>G His260Arg	BIT17_4495 779A>G His260Arg
PPE family protein PPE54 Rv3343c	Unknown	PPE family protein	FCU26_RS18410 3,745,609	BIS44_1763 1301 C > A Ser434Tyr	BIT18_2846 1301 C > A Ser434Tyr	BIT17_2415 1301 C > A Ser434Tyr
Acyl-CoA dehydrogenase fadE34 Rv3573c	Lipid degradation	Acyl-CoA dehydrogenase	FCU26_RS19620 4,016,308	BIS44_3539 Wild type	BIT18_0278 1513 T > G Trp505Gly	BIT17_0253 Wild type
Hypoxanthine phosphoribosyltransferase hpt Rv3624c	Salvage of purines	Hypoxanthine phosphoribosyltransferase	FCU26_RS19850 4,059,295	BIS44_1560 Wild type	BIT18_0654 347 T > G Val116Gly	BIT17_3703 Wild type

Figure 2. Mycobacterium bovis lipid profiles. TLC images of BCG-vaccine and clinical isolates from the BCG vaccinated patient with IFNγR1 deficiency are shown

(A) PDIMs (Phthiocerol dimycocerosates): [¹⁴C]-propionate radiolabeled lipids from BCG-vaccine, BCG-brain, and BCG-lung were dissolved in chloroform and loaded onto a Silica G60 TLC plate, run in petroleum ether:diethylether (90:10, v/v). The positions of the DIM A and DIM B (arrows) are indicated. (B) Triacyglycerol (TAG): [¹⁴C]-propionate radiolabeled TAGs from BCG-vaccine, BCG-brain, and BCG-lung were dissolved in chloroform and loaded onto a Silica G60 TLC plate. TAGs were separated in the first direction with petroleum ether/ethyl acetate (98:2, v/v, 3 runs) and rotated 90° and separated using petroleum ether:acetone (98:2, v/v, 1 run) in the second direction. TAGs position (arrows) is indicated for the different extracts. (C) Mycolic acids: [¹⁴C]-acetate radiolabeled Mycolic acids from BCG-vaccine, BCG-brain, and BCG-lung were dissolved in dichloromethane and loaded on a silica gel F254 thin-layer chromatography (TLC) plate run in hexane:ethyl acetate (19:1, v/v, two runs). M. bovis BCG-derived isolates have the FAMEs and alpha and keto bands which constitute the MAMEs are indicated with arrows. M. bovis BCG isolates lack the methoxy band which is present in M. tuberculosis isolates. Radiolabeled lipids were visualized using a PhosphorImager system.

Figure 3. Quantification of PDIMs, neutral lipids, and mycolic acids

(A) PDIM profiles were quantified using ImageJ and showed a dramatically reduced amount of the DIM B b and of BCG-lung (**p = 0.004) and BCG-brain (*p = 0.043) compared to BCG-vaccine isolate. There were no differences observed in the DIM A band within the three isolates. The graph shows mean ± standard deviation from two independent experiments (n = 4). (B) Nile red staining of neutral lipids (normalized to the weight of the bacterial pellet) showed the BCG-lung (*p = 0.011) isolate to have a greater amount of neutral lipids compared to the vaccine which was statistically significant while the BCG-brain was lower than the vaccine but statistically insignificant. The graph shows mean ± standard deviation from two independent experiments (n = 6). (C) The mycolic acid alpha MAME band showed more production in the lung isolate compared the brain and vaccine while (D) keto MAME band showed a slight increase in the lung isolate compared to the brain and vaccine, however, these differences were not statistically significant. The values were compared to the BCG-vaccine isolate using one-way ANOVA for multiple comparisons using Bonferroni’s test for statistical significance using the GraphPad Prism software (San Diego, CA). All graphs show mean ± standard deviation from two independent experiments (n = 4).

Figure 4. Diminished apoptosis on infected cells by BCG-lung and BCG-brain strains and higher necrosis by infection with BCG-lung measured by LDH-release

(A1) Cells infected with BCG-brain and BCG-lung strains show markedly reduced apoptosis. dTHP1 cells were not infected or infected with the three different strains of BCG (MOI 2) for 2 d. Protein lysates were subjected to Western blot analysis and proved for PARP and β-actin. Blot shown is representative of three repetitions and (A2) Graph is the cumulative densitometric analysis of the ratio of cleaved PARP over the full-size protein. Data, mean ± SD (n = 3; *p < 0.05; **p < 0.01; ANOVA). (B1) Monocyte-derived macrophages infected at MOI of two with the three strains for 2 d showed higher MCL-1 expression with a trend of higher expression on the lung strain infected cells, blot representative of four repetitions, represented in (B2) Graph, is the cumulative densitometric analysis of MCL-1 levels normalized to actin levels (MCL-1/β-actin). Data, mean ± SD (n = 4; **p < 0.01; *p < 0.05; ANOVA). (C) Necrosis is augmented on dTHP1 cells infected with the BCG-lung strain. The supernatants of dTHP1 cells non-infected or infected with the three strains of BCG at a MOI of 2 were used to quantify the release of LDH from the cells. LDH release is expressed as fold increases when compared to the non-infected condition (arbitrarily defined as “1”). LDH levels were elevated in BCG-lung. Data, mean ± SD (n = 4; *p < 0.05; ANOVA)

Figure 5. Infection of dTHP1 cells with the BCG-lung strain resulted in a higher secretion of cytokines. The supernatants of dTHP1 cells infected with the three strains of BCG at a MOI of 2 were used to quantify cytokines IL-1β, IL-18, IL-6, IL-10, TNF⍺, and GM-CSF secretion (A–F). Cytokine levels were compared to untreated cells and within cells infected with BCG isolates. Data, mean ± SD (n = 4; *p < 0.05 **p < 0.01; ANOVA)

Figure 6. Different alterations in STAT signaling between the lung and brain strains compared to the vaccine strain. Monocyte-derived macrophages infected at MOI of 2 with the three strains for 2 d were probed by western blot for: pSTAT1-Y701, STAT1 total, pSTAT5-Y694, STAT5total, SOCS3, and β-actin (n = 3 for STAT1 immunoblots and n = 4 for STAT5 and SOCS3 immunoblots). (A1), (B1), and (C1) representative blots and (A2), (B2), and (C2) are the cumulative data for the experiments. Data, mean ± SD (n ≥ 3; *p < 0.05, **p < 0.01, ***p < 0.001; ANOVA)