Fig 1.
PET CT features of reactivation tuberculosis during TNF neutralization.
Axial views of the lung are shown from two representative latently infected macaques with granulomas seen before (left panels, top and bottom) and during the course of TNF antibody treatment (middle and right panels, top and bottom). Pre-existing lesions (green arrows) within the same lung lobe can increase in size and FDG avidity or remain the same during the TNF neutralization. New lesions (red arrows) can arise in the lungs in an already involved lobe (top row) and/or in a new lobe or extrapulmonary sites such as the liver (red arrows, bottom row).
Fig 2.
Disease pathology, bacterial burden and bacterial killing after TNF neutralization among reactivated and non-reactivated animals.
(A) Macaques with radiographically defined reactivation have greater gross pathology at necropsy compared to animals that did not reactivate. Each symbol represents an animal. Statistics: Mann-Whitney (B) Thoracic bacterial burden (the sum of Mtb growth from all lung granulomas, other lung pathologies and all mediastinal lymph nodes, [MLN]) is higher in reactivated animals than in non-reactivated animals after TNF neutralization. Each symbol represents an animal. Statistics: Mann-Whitney (C) The number of granulomas among animals that developed reactivation (N = 13 macaques, 114 granulomas) are shown with the proportion of sterile granulomas reported in light green compared to animals that did not reactivate (N = 12 macaques, 100 granulomas). The p-value shown compares the proportion of sterile granulomas between groups of animals (Fisher’s Exact analysis). Up to 10 granulomas (randomized) are represented per animal to reduce bias. (D) Animals with reactivation had a smaller proportion of sterile MLNs compared to animals that did not reactivate. Proportions of sterile MLN are shown in the context of the total number of MLN in each group (reactivated group: N = 13 macaques, 85 MLNs vs non-reactivated group N = 13 macaques, 87 MLNs). The p-value shown compares the proportion of sterile granulomas between groups of animals (Fisher’s Exact analysis). Up tp 7 lymph nodes (randomized) are represented per animal to limit bias. (E) Dynamic granulomas (N = 42) had greater CFU per granuloma than stable (N = 132) and new (N = 23) granulomas after TNF neutralization. Percent sterile granulomas is noted for each group. Statistics: Kruskal-Wallis with Dunn’s multiple comparisons. (F) Dynamic lesions had less bacterial killing compared to stable lesions, as assessed by CFU/CEQ ratio. Statistics: Mann Whitney. The numbers indicated above each figure represents the p-value for the given statistical method.
Fig 3.
PET CT characteristics that distinguish reactivated and non-reactivated animals prior to TNF neutralization.
(A) Similar numbers of lung granulomas are observed by PET CT in animals that reactivated and those that did not. (B) Total lung FDG avidity is higher prior to TNF neutralization in macaques that reactivated compared to those that did not. Macaques with an extrapulmonary lesion site (before TNF neutralization) are depicted are in pink. (C, D) The maximum FDG avidity (SUV) and size of any granuloma within a single monkey before TNF neutralization was greater among reactivator compared to non-reactivator animals. Each symbol represents an animal. (E) The number of extrapulmonary (EP) lesion sites before TNF neutralization was greater in animals that reactivated compared to those that did not. (F) A receiver operator curve (ROC) was calculated (AUC = 0.94) based on the presence of EP lesions and total lung FDG avidity (cut-off of 947.2 SUV) showing high sensitivity (92%) and specificity (92%). Panels A-D: each symbol represents a macaque. Statistics for A-E: Mann-Whitney. The numbers in each figure represents the p-value each analysis.
Fig 4.
Parameters of reactivation risk in latently infected control macaques (without TNF neutralization).
Animals were stratified for high or low risk of reactivation based on cut-offs determined by recursive partitioning (high = 947.2 Total FDG Activity or the presence of extrapulmonary lesions). (A) The maximum CFU per granuloma for an individual animal is greater in high-risk (N = 10) compared to low risk (N = 12) animals. (B) High risk LTBI control animals had higher total lung bacterial burden compared to low risk LTBI animals. (C) A trend toward higher CFU per individual MLN was observed in high-risk compared to low-risk LTBI animals. Up to 7 MLN are represented per animal to limit bias. (D) The proportion of sterile MLN observed among low and high risk LTBI controls is shown in the context of the total number of MLN per group. A greater proportion of sterile MLN was observed in low risk LTBI controls compared to high risk animals. The p-value shown compares the proportion of sterile MLN between groups of animals (Fisher’s Exact analysis). Statistics for A-C: Mann-Whitney.
Fig 5.
Frequencies of Mycobacterium tuberculosis (Mtb) specific T cells producing cytokines from individual granulomas of LTBI controls at high (N = 9) or low risk (N = 7) of reactivation.
T cells from granulomas were stimulated with ESAT-6 and CFP-10 peptides. The frequency of T cells producing only IL-10 (A), only IL-17 (B) or only IL-2 (C) was greater in granulomas from high-risk compared to low risk LTBI animals. No differences were observed in IFN- γ (D) or TNF production (E). Each symbol represents a granuloma. Statistics: Mann-Whitney.