Radioiodinated DPA-713 Imaging Correlates with Bactericidal Activity of Tuberculosis Treatments in Mice

Radioiodinated DPA-713 Imaging Correlates with Bactericidal Activityof Tuberculosis Treatments in Mice

Alvaro A. Ordonez,a,b,c Supriya Pokkali,a,b,c Vincent P. DeMarco,b,c Mariah Klunk,a,b,c Ronnie C. Mease,d Catherine A. Foss,a,d

Martin G. Pomper,a,d Sanjay K. Jaina,b,c

Center for Infection and Inflammation Imaging Research,a Center for Tuberculosis Research,b Department of Pediatrics,c and Russell H. Morgan Department of Radiologyand Radiological Sciences,d Johns Hopkins University School of Medicine, Baltimore, Maryland, USA

Current tools for monitoring response to tuberculosis treatments have several limitations. Noninvasive biomarkers could accel-erate tuberculosis drug development and clinical studies, but to date little progress has been made in developing new imagingtechnologies for this application. In this study, we developed pulmonary single-photon emission computed tomography(SPECT) using radioiodinated DPA-713 to serially monitor the activity of tuberculosis treatments in live mice, which developnecrotic granulomas and cavitary lesions. C3HeB/FeJ mice were aerosol infected with Mycobacterium tuberculosis and adminis-tered either a standard or a highly active bedaquiline-containing drug regimen. Serial 125I-DPA-713 SPECT imaging was com-pared with 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET) and standard microbiology. Ex vivo studieswere performed to characterize and correlate DPA-713 imaging with cellular and cytokine responses. Pulmonary 125I-DPA-713SPECT, but not 18F-FDG PET, was able to correctly identify the bactericidal activities of the two tuberculosis treatments as earlyas 4 weeks after the start of treatment (P < 0.03). DPA-713 readily penetrated the fibrotic rims of necrotic and cavitary lesions. Atime-dependent decrease in both tumor necrosis factor alpha (TNF-!) and interferon gamma (IFN-") levels was observed withtreatments, with 125I-DPA-713 SPECT correlating best with tissue TNF-! levels (# $ 0.94; P < 0.01). 124I-DPA-713 was also eval-uated as a PET probe and demonstrated a 4.0-fold-higher signal intensity in the infected tuberculous lesions than uninfectedcontrols (P $ 0.03). These studies provide proof of concept for application of a novel noninvasive imaging biomarker to monitortuberculosis treatments, with the potential for application for humans.

Recognizing that tuberculosis (TB) is still a leading cause ofhuman death from a curable disease, the international health

community has set an ambitious target to eliminate TB by 2050.However, using mathematical modeling, Dye and Williams at theWorld Health Organization have shown that while most TB pa-tients can be cured with current drug treatments, the 2050 targetcannot be achieved with current tools and requires a combinationof new diagnostics, shorter TB drug treatments, and new vaccines(1). However, current tools for evaluating TB therapeutics haveseveral limitations. Conventional preclinical studies are limited toanalysis of serial postmortem samples using microbiologic meth-ods that take 3 to 4 weeks for results. Moreover, different groups ofanimals are sacrificed over several time points during the study,and therefore, assessments of disease in the same animal can neverbe made. Similar limitations exist for monitoring TB treatments inhumans. The standard 8-week sputum culture conversion is notavailable in real time, taking several weeks for results. Even thoughnucleic acid amplification tests such as GeneXpert provide resultsrapidly (2), both sputum culture and GeneXpert are subject tosampling bias and provide information only about the lesionscommunicating with the airways. Noncommunicating pulmo-nary or extrapulmonary lesions are often never assessed. Similarly,assessment for relapse can require monitoring hundreds of pa-tients for up to 2 years after treatment completion. With increas-ing rates of multidrug-resistant, extensively drug-resistant, andtotally drug-resistant TB (3, 4), it is imperative to develop evenbetter tools to monitor treatment responses and predict relapse.

Noninvasive imaging provides rapid, three-dimensional viewsof the whole body, as well as the ability to monitor disease in thesame individual. Real-time, longitudinal assessments can alsoprovide new insights into the pathophysiology of disease, which

may be difficult to assess with current technologies. Computedtomography (CT) and 18F-fluorodeoxyglucose (18F-FDG) posi-tron emission tomography (PET) are being increasingly used tomonitor TB (5–8), in both preclinical and clinical settings. How-ever, both CT and 18F-FDG PET lack specificity, and 18F-FDG istaken up by all glycolytically active tissues (9–11). Since activatedmacrophages are key components of TB-associated inflamma-tion, macrophage-avid tracers could serve as more specific imag-ing agents.

The translocator protein (TSPO) is an 18-kDa trans-mito-chondrial membrane channel utilized for transport of cholesteroland other endogenous ligands (12). TSPO expression is high inseveral tissues and in activated immune cells such as microglia andmacrophages (13). We have previously demonstrated that radio-iodinated DPA-713, a low-molecular-weight pyrazolopyrimidineligand for TSPO, specifically accumulates in activated phagocyticcells in Mycobacterium tuberculosis-induced inflammatory lesions

Received 27 August 2014 Returned for modification 30 September 2014Accepted 27 October 2014

Accepted manuscript posted online 17 November 2014

Citation Ordonez AA, Pokkali S, DeMarco VP, Klunk M, Mease RC, Foss CA, PomperMG, Jain SK. 2015. Radioiodinated DPA-713 imaging correlates with bactericidalactivity of tuberculosis treatments in mice. Antimicrob Agents Chemother59:642–649. doi:10.1128/AAC.04180-14.

Address correspondence to Sanjay K. Jain, [email protected].

Supplemental material for this article may be found at http://dx.doi.org/10.1128/AAC.04180-14.

Copyright © 2015, American Society for Microbiology. All Rights Reserved.

doi:10.1128/AAC.04180-14

642 aac.asm.org January 2015 Volume 59 Number 1Antimicrobial Agents and Chemotherapy

on Decem

ber 30, 2014 by WELC

H M

EDIC

AL LIBRAR

Y - John Hopkins U

http://aac.asm.org/

Dow

nloaded from

and is more specific than 18F-FDG PET in mice (14). In this study,we utilized serial 125I-DPA-713 single-photon emission computedtomography (SPECT) to monitor the activity of a novel, bedaquiline-containing TB drug regimen being developed against multidrug-re-sistant (MDR) TB (15) in live mice that develop necrotic and hypoxicTB lesions (16, 17). Ex vivo studies were performed to characterizeand correlate DPA-713 imaging with cellular and cytokine responsesin different TB lesions, including cavities.

MATERIALS AND METHODSAll protocols were approved by the Johns Hopkins Biosafety, RadiationSafety, and Animal Care and Use Committees.

Animal infection and treatments. Four- to six-week-old femaleC3HeB/FeJ (Jackson Laboratory, Bar Harbor, ME) mice were aerosol in-fected with frozen stocks of M. tuberculosis H37Rv, using the Middlebrookinhalation exposure system (Glas-Col, Terre Haute, IN). Six weeks afterinfection, mice were randomly divided into treatment groups and orallyadministered (5 days per week) either the standard RHZ (rifampin, 10mg/kg; isoniazid, 10 mg/kg; pyrazinamide, 150 mg/kg) or a highly activeJZC (bedaquiline, 25 mg/kg; pyrazinamide, 150 mg/kg; clofazimine, 20mg/kg) regimen for 12 and 8 weeks, respectively (Fig. 1). All TB drugswere obtained from Sigma (St. Louis, MO) except bedaquiline (!98%chemical purity), which was purchased from Adooq Bioscience LLC (Ir-vine, CA). Mice were sacrificed to determine the bacillary burden 1 dayafter infection and at each time point (14). Both sides of the lungs werehomogenized in 1 ml of phosphate-buffered saline, serially diluted, andplated onto Middlebrook 7H11 selective plates (Becton Dickinson, Frank-lin Lakes, NJ) to determine CFU per ml. Five mice were used for eachgroup and at each time point, except at the 4-week time point for theuntreated group, where only 4 mice were used. A separate group of iden-tically infected mice were used for imaging studies. A cohort of mice (n "13) receiving the standard (RHZ) treatment were also followed for 16weeks after cessation of treatment and sacrificed to determine the pulmo-nary bacterial burden (entire lungs) at the final time point.

Imaging. 125I-DPA-713 was synthesized in-house as described previ-ously with !99% radiochemical purity and specific activities rangingfrom 66,600 to 77,700 GBq/mmol (18). Live M. tuberculosis-infected an-imals were imaged within a sealed biocontainment bed (Minerve, Ester-nay, France) modified in-house to comply with biosafety level 3 (BSL-3)containment (5, 19). Filters (0.22 #m; GE Whatman, Pittsburgh, PA)were used at both the inlet and the outlet to contain the bacteria within thedevice. Animals were anesthetized and sealed inside the biocontainmentdevice in the BSL-3 facility. A standard small-animal anesthesia machinewas used to deliver an isoflurane (Henry Schein, Melville, NY)-oxygenmixture during transport and imaging. Prior to imaging, mice were fastedfor 8 h. Each animal was weighed, injected intravenously with 7.3 MBq of18F-FDG and 37 MBq of 125I-DPA-713 simultaneously via tail vein, andimaged 45 min and 24 h after the intravenous injection, using a Mosaic HP

PET (Philips, Bothell, WA) and NanoSPECT/CT (Bioscan, Washington,DC) small-animal imager, respectively (PET, SPECT, and CT imaging ofeach animal at each time point). The same group of mice was imaged overseveral time points (outlined in Fig. 1). To prevent “cross talk” betweenPET and SPECT, 18F-FDG PET was performed 45 min after injection onthe first day. Due to its short half-life (109 min), 18F decays completely by24 h, when the 125I-DPA-713 SPECT was performed. Also, the primaryenergies of 125I (maximum of 35 keV) are much lower than those of 18F(all PET tracers emit at 511 keV, resulting from electron-positron annihi-lation); photons from 125I are excluded from the PET scan. Five mice wereimaged for each group and at each time point except at week 0, when fourmice were imaged (untreated group). Images were reconstructed andcoregistered with computed tomography (CT) images using AMIDE 1.0.4(http://amide.sourceforge.net). Standardized uptake values (SUV) werecomputed as described previously (14). Briefly, by using the coregisteredCT images as a reference, spherical (9-mm3 volume) regions of interest(ROIs) were drawn around three randomly selected pulmonary lesions,making sure not to overlap the surrounding PET-active bone marrow orheart, creating 15 ROIs per group for each time point. These ROIs wereapplied to the coregistered SPECT and PET images. Since each sphericalROI contains both full and partial voxels depending on the orientation,correction for partial volume effects was applied to all imaging data usingAMIDE. This is considered standard for image analyses. The SUV data arepresented as a percentage of the signal noted at the initial time point (startof treatment).

Histology and immunofluorescence. DPA-713-IRDye680LT was ad-ministered to M. tuberculosis-infected mice (8 weeks after infection; no TBtreatment) intravenously 24 h prior to euthanasia, and lungs were pro-cessed as described previously (14). Paraffin-embedded formalin-fixedtissue sections were probed using fluorescent antibodies (see Table S1 inthe supplemental material). A Nikon 80i upright epifluorescence micro-scope (Nikon Instruments, Melville, NY) equipped with a Nikon DS-Qi1Mc dark-field charge-coupled device (CCD) camera and Nikon Inten-silight C-HGFI lamp was used. All images were recorded and processedusing Nikon Imaging software elements.

Flow cytometry. DPA-713-IRDye680LT and brefeldin A (Sigma)were administered to a separate group of M. tuberculosis-infected mice (8weeks after infection; no TB treatment) intravenously 24 and 6 h prior toeuthanasia, respectively, as described previously (14, 20). Lungs were har-vested and processed as described previously (21). Briefly, cells (2 $ 106

cells/well) were incubated with cell phenotype-specific markers (see TableS1 in the supplemental material) in the presence of anti-FcRII/III anti-body (1.3 #g/ml) at 4°C for 1 h. Thereafter, cells were washed, fixed, andpermeabilized for intracellular cytokine staining. Samples were analyzedon a LSR-II instrument (BD, San Jose, CA), and data were analyzed usingFACSDiva v5.0.1 software (BD). The gating strategy for these analyses isshown in Fig. S1 in the supplemental material. A minimum of 10,000events were acquired within each cell-specific gate. Samples from fivedifferent mice were utilized for these assays.

Cytokine estimation. Whole-lung homogenates from mice receivingTB treatments were frozen at %80°C. Samples were thawed and filtersterilized to remove tissue and bacterial debris. Tumor necrosis factoralpha (TNF-&) and interferon gamma (IFN-') were assayed by using anenzyme-linked immunosorbent assay (ELISA) kit (R & D Systems, Min-neapolis, MN). Protein content was estimated by using the Bradford as-say. The cytokine concentrations are expressed as pg per mg of total pro-tein. Samples from four different mice were utilized for each group and ateach time point.

124I-DPA-713 PET imaging. Clinical-grade 124I-DPA-713 was syn-thesized using current good manufacturing practices (cGMP) under aresearch contract (3D Imaging, Maumelle, AR). Each mouse was weighed,injected intravenously with 13.5 MBq of 124I-DPA-713 (!95% radio-chemical purity; specific activity, 96,089 GBq/mmol), and imaged 24 hlater using the Mosaic HP PET. Five mice were imaged for each group.Image analyses were performed as described above.

FIG 1 Timeline and experimental scheme. Six weeks after aerosol infectionwith Mycobacterium tuberculosis, mice were randomly divided into treatmentgroups. Different animal cohorts received either the standard (RHZ) or a highlyactive bedaquiline-containing regimen (JZC). Animals receiving standard treat-ment were also followed for 16 weeks after the completion of treatment to monitorfor relapse. Mice were serially imaged at weeks 0, 4, 8, and 12 to assess the bacteri-cidal activity. A separate cohort of 13 animals which received 12 weeks of standard(RHZ) treatment were followed for the development of relapse and imaged atweeks 18 and 28 (6 and 16 weeks after cessation of TB treatment). The pulmonarybacterial burdens from a separate cohort of similarly infected and treated micewere determined using standard microbiology at each time point.

Radioiodo-DPA-713 Imaging To Monitor TB Treatments

January 2015 Volume 59 Number 1 aac.asm.org 643Antimicrobial Agents and Chemotherapy

on Decem

ber 30, 2014 by WELC

H M

EDIC

AL LIBRAR

Y - John Hopkins U

http://aac.asm.org/

Dow

nloaded from

Statistical analysis. Statistical comparisons were performed using atwo-tailed Mann-Whitney U or Kruskal-Wallis (for multiple compari-sons) test and Spearman’s rank correlation (GraphPad Software Inc., LaJolla, CA). P values of (0.05 were considered statistically significant.

RESULTSPulmonary M. tuberculosis bacillary burdens 1 day and 6 weeks(time of initiation of TB treatments) after infection were 2.00 )0.13 and 8.05 ) 0.02 log10 CFU/ml, respectively.

125I-DPA-713 SPECT imaging correlates with bactericidalactivity of TB treatments. We evaluated serial 125I-DPA-713SPECT imaging to measure the response to a novel, bedaquiline-containing TB drug regimen (JZC). 125I-DPA-713 SPECT wascompared with 18F-FDG PET imaging and standard microbiol-ogy. M. tuberculosis-infected mice were serially imaged, generat-ing 102 image sets (SPECT, PET, and CT). In agreement with the

FIG 2 125I-DPA-713 SPECT imaging correlates with bactericidal activity ofTB treatments. Four-to-six-week-old female C3HeB/FeJ mice were aerosolinfected with M. tuberculosis. Mice were sacrificed to determine the bacillaryburden of whole lungs 1 day after infection and at each time point. A separategroup of identically infected mice were used for imaging studies. Six weeksafter infection, mice were randomly divided into treatment groups and orallyadministered (five times per week) either the standard RHZ or a highly activeJZC regimen for 12 and 8 weeks, respectively. (A) Consistent with the higherbactericidal activity of bedaquiline-containing regimens, 8 weeks of treatmentwith JZC cleared the infection in the majority of mice, versus 12 weeks requiredto achieve the same bacterial killing with the standard (RHZ) treatment. (B)Pulmonary 18F-FDG PET imaging correlated with the pulmonary bacterialburden (Spearman’s * " 0.78; P ( 0.01) but was unable to correctly identifythe bactericidal activities of the two TB treatments (P ! 0.49). (C) Pulmonary125I-DPA-713 SPECT imaging correlated well with the pulmonary bacterialburden (Spearman’s * " 0.92; P ( 0.01) and also correctly identified the

bactericidal activities of the two TB regimens as early as 4 weeks after the startof treatment (P ( 0.03). CFU data are presented on a logarithmic scale asmeans and standard deviations. The SUV are presented as percentages of thesignal noted at the initial time point (start of treatment) on a linear scale,expressed as medians and interquartile ranges.

FIG 3 Correlation between imaging and bacterial burden during relapse. M.tuberculosis-infected mice were serially imaged 16 weeks after completion ofstandard (RHZ) TB treatment to monitor relapse. The correlation between thechange in the standardized uptake value ratio (SUVR) between 6 and 16 weeksafter cessation of TB treatment (SUVR16week/6week) and postmortem pulmo-nary bacterial burdens at the final imaging time point is shown. Each datapoint represents an independent animal. A significant correlation was foundbetween 125I-DPA-713 SPECT results and bacterial burden (Spearman’s * "0.79; P ( 0.01) (A) but not between 18F-FDG PET imaging and bacterialburden (Spearman’s * " 0.34; P " 0.26) (B).

Ordonez et al.

644 aac.asm.org January 2015 Volume 59 Number 1Antimicrobial Agents and Chemotherapy

on Decem

ber 30, 2014 by WELC

H M

EDIC

AL LIBRAR

Y - John Hopkins U

http://aac.asm.org/

Dow

nloaded from

previously reported higher bactericidal activity of bedaquiline-containing regimens (22), 8 weeks of treatment with JZC clearedthe infection in the majority of mice, versus 12 weeks required toachieve the same bacterial killing with the standard (RHZ) treat-ment (Fig. 2A). Pulmonary 125I-DPA-713 SPECT activity corre-lated well with the decrease in the pulmonary bacterial burden(Spearman’s * " 0.92; P ( 0.01). Although 18F-FDG PET couldrapidly distinguish treated versus untreated animals (P ( 0.01), itwas unable to correctly identify the bactericidal activities of thetwo TB treatments (P ! 0.49) (Fig. 2B). In contrast, pulmonary125I-DPA-713 SPECT activity not only was able to rapidly distin-guish the treated versus untreated groups (P ( 0.01) but alsocorrectly identified the bactericidal activities of the two TB regi-mens as early as 4 weeks after the start of treatment (P ( 0.03)(Fig. 2C).

A cohort of 13 M. tuberculosis-infected mice were imaged 6 and16 weeks after completion of standard (RHZ) TB treatment toassess relapse. We hypothesized that for an ideal imaging bio-marker, an increase in the mean pulmonary activity during relapsewould correlate with the pulmonary bacterial burden. A total of 78image sets (SPECT, PET, and CT) were analyzed to measure thechange in the SUV between 6 and 16 weeks after cessation of TB

treatment (SUVR16week/6week) and compared with postmortempulmonary CFU at the final time point (Fig. 3). A significant correla-tion was found between the interval increase in 125I-DPA-713 SPECTactivity for each animal with the pulmonary bacterial burden at re-lapse (Spearman’s * " 0.79; P ( 0.01). However, interval increase in18F-FDG PET activity did not correlate well with the pulmonary bac-terial burden at relapse (Spearman’s * " 0.34; P " 0.26).

125I-DPA-713 SPECT imaging of TB lesions. Since pulmonaryTB is characterized by multiple different pathologies, we evaluatedthe ability of DPA-713 to detect and penetrate different TB lesions.Figure 4A to C show CT, 18F-FDG PET/CT, and 125I-DPA-713SPECT/CT images from a mouse with pulmonary lesions 8 weeksafter infection. Both 18F-FDG PET and 125I-DPA-713 SPECT signalscolocalized with the TB lesions seen on CT. Postmortem histopathol-ogy demonstrated the necrotic granuloma (Fig. 4D), with numerousintracellular and extracellular bacilli (Fig. 4E) and fibrosis around therim (Fig. 4F). Immunofluorescence analyses confirmed that intrave-nously injected DPA-713 penetrated into the cellular rim and colo-calized with inflammatory cells (Fig. 4G to J).

One month after the completion of RHZ treatment, pulmo-nary cavitary lesions were observed on CT imaging in 61% (11 of18) of mice. As before, both 18F-FDG PET and 125I-DPA-713

FIG 4 Imaging of necrotic pulmonary TB lesions. The transverse (left), coronal (middle), and sagittal (right) views from an M. tuberculosis-infected mouse (8weeks after infection; no TB treatment) demonstrating a necrotic TB lesion (cross-hairs) are shown. (A) The TB lesion is visible as a radiodense area on the CTimages. (B) The 18F-FDG PET signal localizes at the site of the TB lesion. (C) The 125I-DPA-713 SPECT signal also colocalizes with the TB lesion. H, heart. (D)Postmortem histopathology (hematoxylin and eosin staining [H&E]) demonstrates a necrotic granuloma with a cellular rim. (E) Both intracellular andextracellular bacilli are seen on acid-fast stain. (F) Masson’s trichrome stain demonstrates collagen deposition (blue) at the fibrotic rim of the necrotic granuloma.Immunofluorescence demonstrates signal from DPA-713-IRDye680LT (G), CD11b+ macrophages (H), and Hoechst stain for nuclei (I). (J) Overlay of allchannels shows that DPA-713-IRDye680LT penetrates the fibrotic lesion and colocalizes with the CD11b+ signal.

Radioiodo-DPA-713 Imaging To Monitor TB Treatments

January 2015 Volume 59 Number 1 aac.asm.org 645Antimicrobial Agents and Chemotherapy

on Decem

ber 30, 2014 by WELC

H M

EDIC

AL LIBRAR

Y - John Hopkins U

http://aac.asm.org/

Dow

nloaded from

SPECT signals colocalized with the rim of the cavitary lesion (Fig.5A to C). Postmortem histopathology demonstrated the cavity(Fig. 5D and E), with numerous intracellular and extracellularbacilli (Fig. 5F) and prominent fibrosis around the rim (Fig. 5Gand H). Picrosirius red staining (23, 24) also suggested remodel-ing of the extracellular matrix (Fig. 5I). Immunofluorescenceanalysis confirmed that intravenously injected DPA-713 pene-trated into the cellular rim and colocalized with inflammatorycells (Fig. 5J to O).

Optical imaging in live animals followed by subsequent post-mortem ex vivo whole-lung imaging also confirmed localization ofDPA-713 with the TB lesions (see Fig. S2 in the supplementalmaterial).

Correlation with cellular and cytokine responses. To confirmand further characterize cellular binding, multicolor flow-cyto-metric analyses were performed. Overall, 40.2%, 17.7%, 24.9%,and 2.7% of DPA-713+ cells were CD11b+, CD11c+, Gr-1+, andTCR,+, respectively (Fig. 6A). Moreover, 78.2 ) 6.0% of DPA-713+ cells expressed CD68, with 94.4 ) 1.2%, 94.7 ) 1.0%, and91.6 ) 1.5% of CD11b+, CD11c+ and Gr-1+ cells, respectively,coexpressing CD68. Intracellular expression of TNF-& and IFN-'was analyzed within the DPA-713+ CD68+ cell subsets. Figure 6Band C show histograms for TNF-& and IFN-', respectively.TNF-& was expressed in 17.2 ) 3.2%, 31.6 ) 1.2%, and 12.2 )2.4% of CD11b+, CD11c+, and Gr-1+ cells, respectively. Simi-

larly, IFN-' was expressed in 22.1 ) 2.9%, 31.2 ) 1.6%, and20.3 ) 2.1% of CD11b+, CD11c+, and Gr-1+ cells, respectively.Finally, we also determined cytokine levels in whole-lung homog-enates from animals receiving TB treatments (Fig. 6D and E).There was a time-dependent decrease in both TNF-& and IFN-'levels during TB treatment. 125I-DPA-713 SPECT imaging corre-lated best with tissue TNF-& levels (Spearman’s * " 0.94; P (0.01) (see Fig. S3A in the supplemental material).

124I-DPA-713 PET imaging. Although 125I-DPA-713 SPECTimaging is excellent for preclinical assessments in mice, 125I emitslow-energy photons and lacks the tissue penetration needed forhuman imaging. However, iodo-DPA-713 can be labeled withpositron-emitting (124I) radioisotopes, ideal for PET imaging inhumans. We therefore evaluated the uptake and distribution of124I-DPA-713 in mice (8 weeks after infection; no TB treatment).124I-DPA-713 PET demonstrated excellent localization of the ra-diotracer with TB lesions in M. tuberculosis-infected mice (Fig. 7).124I-DPA-713 PET produced 4.0-fold-higher signal intensity inthe infected TB lesions compared to healthy lungs (uninfectedcontrols) (P " 0.03). Three-dimensional coregistered 124I-DPA-713 PET/CT movies demonstrated discrete areas of 124I-DPA-713PET signal in the infected lung tissues, with only minimal back-ground signal in the uninfected animal (see Movies S1 and S2 inthe supplemental material).

FIG 5 Imaging of cavitary pulmonary TB lesions. The transverse (left), coronal (middle), and sagittal (right) views from an M. tuberculosis-infected mousedemonstrating a cavitary TB lesion (cross-hairs) are shown. (A) The cavitary lesion is visible as a radiolucency on the CT images. (B) The 18F-FDG PET signallocalizes at the rim of the cavity. (C) The 125I-DPA-713 SPECT signal also colocalizes with the rim of the cavitary lesion. H, heart. (D and E) Postmortemhistopathology (H&E) demonstrates a cavity with a cellular rim. (F) Both intracellular and extracellular bacilli are seen on acid-fast stain. (G and H) Collagendeposition is noted on Masson’s trichrome (blue [G]) and reticulin (pink [H]) stains at the fibrotic rim of the cavity. (I) Picrosirius red stain observed undercircularly polarized light microscopy demonstrates both mature (red) and immature (yellow; birefringence) collagen fibers surrounding the cavity. Immuno-fluorescence demonstrates signal from DPA-713-IRDye680LT (K), CD11b (L) CD3 (M) and Hoechst stain for nuclei (N). (O) Overlay of the DPA-713-IRDye680LT and CD11b channels demostrates colocalization of the signals. (J) Overlay of all channels demonstrating that DPA-713-IRDye680LT penetrates thefibrotic lesion and colocalizes with the CD11b+ signal.

Ordonez et al.

646 aac.asm.org January 2015 Volume 59 Number 1Antimicrobial Agents and Chemotherapy

on Decem

ber 30, 2014 by WELC

H M

EDIC

AL LIBRAR

Y - John Hopkins U

http://aac.asm.org/

Dow

nloaded from

DISCUSSIONMolecular imaging of TB pathology can provide valuable real-time information, which could allow early prediction of TB treat-ment responses and the development of relapse (25). While cur-rent tomographic imaging techniques such as CT and 18F-FDGPET can provide information on treatment responses by measur-

ing the change in host immune responses that occur during treat-ment (5–8), they lack specificity, and 18F-FDG is taken up by awide range of inflammatory and noninflammatory cells. Since ac-tivated macrophages are key components of TB-associated in-flammation, we hypothesized that real imaging with a macro-phage-avid tracer such as radioiodo-DPA-713 would serve as a

FIG 7 Pulmonary 124I-DPA-713 PET imaging. The transverse (left), coronal (middle), and sagittal (right) views from an M. tuberculosis-infected mouse (8 weeksafter infection; no TB treatment) demonstrating a TB lesion (cross-hairs) are shown. (A and B) CT and 124I-DPA-713 PET demonstrate colocalization of the PETsignal with the TB lesion observed on the CT image. (C) The pulmonary 124I-DPA-713 PET signal from M. tuberculosis-infected mice is significantly higher thanthe signals from uninfected (control) animals (P " 0.03). Standardized uptake values (SUV) are represented as box plots, where central bars represent medians,the edges of the boxes represent quartiles, and whiskers show the upper and lower limits of the range. At least four mice were imaged for each group.

FIG 6 Correlation with cellular and cytokine responses. To confirm and characterize the phenotype of DPA-713+ cells, multicolor flow-cytometric analyses wereperformed in whole-lung homogenates from M. tuberculosis-infected animals (8 weeks postinfection). (A) Pie chart depicting the percentages of CD11b+

(macrophages), CD11b+ Gr-1+ (inflammatory macrophages), Gr-1+ (granulocytes), CD11c+ (dendritic cells), TCR-,+ (T cells), and other cells within theparent DPA-713+ gate. (B and C) Respective histograms showing the proportion of DPA-713+ CD68+ cells with intracellular expression of TNF-& and IFN-'by cell subset. (D and E) Tissue cytokine levels in whole-lung homogenates from animals receiving TB treatments for TNF-& and IFN-'. Cytokine data arepresented on a linear scale as means and standard errors of the means. Five biological samples were used for each analyses.

Radioiodo-DPA-713 Imaging To Monitor TB Treatments

January 2015 Volume 59 Number 1 aac.asm.org 647Antimicrobial Agents and Chemotherapy

on Decem

ber 30, 2014 by WELC

H M

EDIC

AL LIBRAR

Y - John Hopkins U

http://aac.asm.org/

Dow

nloaded from

more specific biomarker technique to monitor the response to TBtreatments. Pulmonary 125I-DPA-713 SPECT demonstrated sig-nificant correlation with the bactericidal activities of TB treat-ments and was able to correctly identify the bactericidal activities(as early as 4 weeks) of the standard (RHZ) and a novel, highlyactive bedaquiline-containing TB regimen (JZC). We also foundthat an early increase in 125I-DPA-713 SPECT activity, but not18F-FDG PET activity, correlated significantly with the bacterialburden at relapse. Collectively, these data suggest that iodo-DPA-713 imaging may be more specific for TB-associated inflamma-tion than 18F-FDG PET and could provide a better means of mon-itoring TB treatments.

After deposition of M. tuberculosis in the alveoli via an airborneroute, initial events include bacterial phagocytosis by alveolarmacrophages and other phagocytic cells, bacterial replication, de-velopment of delayed-type hypersensitivity, and the formation ofthe pulmonary granuloma (26). With the emergence of delayed-type hypersensitivity (DTH), infected cells in the interior of thegranuloma are killed, leading to the formation of the “classic” TBgranuloma—an organized structure with central areas of caseousnecrosis surrounded by dense infiltrates of neutrophils, activatedmacrophages, lymphocytes, and fibroblasts. Some lesions also ex-pel their contents into airways to form cavitary lesions with highbacillary burdens (107 to 109), making the patient highly infec-tious (27). Therefore, active pulmonary TB in humans is charac-terized by a complex spectrum of disease with multiple patholo-gies, including granulomas with caseous necrosis and cavitation(28, 29). While standard mouse strains cannot recapitulatenecrotic or cavitary pathologies, we and others have demonstratedthat in addition to developing necrotic and hypoxic lesions,C3HeB/FeJ mice also develop cavitary TB (24, 30, 31). Similar tohuman pathology, where granulomas are surrounded by a fibroticcapsule (32), extensive fibrosis was also noted around chronic TBgranulomas and cavities in C3HeB/FeJ mice. Despite the fibroticbarrier, radiolabeled and fluorescent DPA-713 analogs readilypenetrated the TB lesions, suggesting that radioiodinated DPA-713 could be used to detect and quantify TB-associated inflamma-tory responses in different TB lesions noninvasively.

Since DPA-713 accumulates specifically in activated inflamma-tory cells, we wanted to correlate the imaging findings with the in-flammatory responses. We therefore characterized and correlated125I-DPA-713 SPECT with tissue cellular and cytokine responses inM. tuberculosis-infected lungs. As reported previously, the major-ity of DPA-713+ cells were found to express CD68 (14). Morespecifically, DPA-713 bound to activated (CD68+) antigen-pre-senting cells (macrophages [CD11b+], dendritic cells [CD11c+],and neutrophils [Gr-1+]) but not lymphocytes (TCR,+). A smallproportion (14.2%) of DPA-713+ cells did not stain positive forthese cell markers and could represent activated epithelial cells orpneumocytes (14). We also found that CD68+ neutrophils com-prised a significant proportion (17.7%) of the DPA-713+ cells.Gottfried et al. and Amanzada et al. demonstrated that CD68 isnot specific to macrophages and is a lysosomal protein enriched inactivated macrophages, dendritic cell, granulocytes, and some fi-broblasts (33, 34). We also found that 18% of DPA-713+ cellswere CD11b+ Gr-1+ inflammatory macrophages. Myeloid-de-rived suppressor cells (inflammatory macrophages) are character-ized by the coexpression of CD11b and Gr-1, and a systemic ex-pansion of these cells has been noted previously with TB (35).Also, as expected, functionally active CD68+ and DPA-713+ cells

had high levels of intracellular TNF-& and IFN-' expression, bothof which are key cytokines in TB pathology (36). Finally, we alsodetermined cytokine levels from whole-lung homogenates of an-imals receiving TB treatments. A time-dependent decrease in bothTNF-& and IFN-' levels was observed with TB treatments, and125I-DPA-713 SPECT imaging correlated best with TNF-& levels.It should be noted that while cytokines were determined fromwhole-lung homogenates, DPA-713 SPECT measured the activityin discrete lung lesions, which would presumably also be the pre-dominant sites of cytokine production.

Although 125I-DPA-713 SPECT imaging is excellent for pre-clinical assessments in mice, 125I emits low-energy photons andlacks the tissue penetration needed for human imaging. However,iodo-DPA-713 can be labeled with higher-energy direct ' (123I) orpositron-emitting (124I) radioisotopes, ideal for SPECT or PETimaging in humans, respectively. We therefore utilized 124I-DPA-713 manufactured under cGMP and demonstrated excellent PETsignal-to-noise ratios for pulmonary TB lesions, suggesting that124I-DPA-713 PET is an excellent candidate imaging biomarkertechnique for TB-associated inflammation. 124I-DPA-713 PETcould be utilized for preclinical studies but also has the potentialfor translation to humans, especially in settings such as TB trials,where resources are not as limited but rapid, and accurate bio-markers for monitoring TB treatments and relapse are urgentlyrequired.

ACKNOWLEDGMENTSThis work was supported by R01-HL116316 (S.K.J.), NIH Director’s NewInnovator Award DP2-OD006492 (S.K.J.), subcontract 5P30AI060354-09from Harvard University Center for AIDS Research (S.K.J.), and the K-RITHtravel award (S.K.J.).

We thank Haofan Wang (JHU) for providing DPA-713-IRDye680LT,as well as Jonathan Shahbazian and Lloyd Miller (JHU) for their assistancewith optical imaging.

REFERENCES1. Dye C, Williams BG. 2008. Eliminating human tuberculosis in the twenty-

first century. J R Soc Interface 5:653–662. http://dx.doi.org/10.1098/rsif.2007.1138.

2. Boehme CC, Nabeta P, Hillemann D, Nicol MP, Shenai S, Krapp F,Allen J, Tahirli R, Blakemore R, Rustomjee R, Milovic A, Jones M,O’Brien SM, Persing DH, Ruesch-Gerdes S, Gotuzzo E, Rodrigues C,Alland D, Perkins MD. 2010. Rapid molecular detection of tuberculosisand rifampin resistance. N Engl J Med 363:1005–1015. http://dx.doi.org/10.1056/NEJMoa0907847.

3. Centers for Disease Control and Prevention. 2006. Emergence of My-cobacterium tuberculosis with extensive resistance to second-line drugs—worldwide, 2000-2004. MMWR Morb Mortal Wkly Rep 55:301–305. http://dx.doi.org/10.1037/e566482006-002.

4. Velayati AA, Masjedi MR, Farnia P, Tabarsi P, Ghanavi J, Ziazarifi AH,Hoffner SE. 2009. Emergence of new forms of totally drug-resistant tu-berculosis bacilli: super extensively drug-resistant tuberculosis or totallydrug-resistant strains in Iran. Chest 136:420 – 425. http://dx.doi.org/10.1378/chest.08-2427.

5. Davis SL, Nuermberger EL, Um PK, Vidal C, Jedynak B, Pomper MG,Bishai WR, Jain SK. 2009. Noninvasive pulmonary [18F]-2-fluoro-deoxy-D-glucose positron emission tomography correlates with bacteri-cidal activity of tuberculosis drug treatment. Antimicrob Agents Che-mother 53:4879 – 4884. http://dx.doi.org/10.1128/AAC.00789-09.

6. Sathekge M, Maes A, Kgomo M, Stoltz A, Van de Wiele C. 2011. Use of18F-FDG PET to predict response to first-line tuberculostatics in HIV-associated tuberculosis. J Nucl Med 52:880 – 885. http://dx.doi.org/10.2967/jnumed.110.083709.

7. Coleman MT, Maiello P, Tomko J, Frye LJ, Fillmore D, Janssen C, Klein E, LinPL. 2014. Early changes by (18)fluorodeoxyglucose positron emission tomogra-phy coregistered with computed tomography predict outcome after Mycobacte-

Ordonez et al.

648 aac.asm.org January 2015 Volume 59 Number 1Antimicrobial Agents and Chemotherapy

on Decem

ber 30, 2014 by WELC

H M

EDIC

AL LIBRAR

Y - John Hopkins U

http://aac.asm.org/

Dow

nloaded from

rium tuberculosis infection in cynomolgus macaques. Infect Immun 82:2400–2404. http://dx.doi.org/10.1128/IAI.01599-13.

8. Via LE, Schimel D, Weiner DM, Dartois V, Dayao E, Cai Y, Yoon YS,Dreher MR, Kastenmayer RJ, Laymon CM, Carny JE, Flynn JL, Hers-covitch P, Barry CE, 3rd. 2012. Infection dynamics and response tochemotherapy in a rabbit model of tuberculosis using [18F]2-fluoro-deoxy-D-glucose positron emission tomography and computed tomogra-phy. Antimicrob Agents Chemother 56:4391– 4402. http://dx.doi.org/10.1128/AAC.00531-12.

9. Bell GI, Burant CF, Takeda J, Gould GW. 1993. Structure and functionof mammalian facilitative sugar transporters. J Biol Chem 268:19161–19164.

10. Pauwels EK, Ribeiro MJ, Stoot JH, McCready VR, Bourguignon M,Maziere B. 1998. FDG accumulation and tumor biology. Nucl Med Biol25:317–322. http://dx.doi.org/10.1016/S0969-8051(97)00226-6.

11. Zhuang H, Alavi A. 2002. 18-Fluorodeoxyglucose positron emissiontomographic imaging in the detection and monitoring of infection andinflammation. Semin Nucl Med 32:47–59. http://dx.doi.org/10.1053/snuc.2002.29278.

12. Rupprecht R, Papadopoulos V, Rammes G, Baghai TC, Fan J, Akula N,Groyer G, Adams D, Schumacher M. 2010. Translocator protein (18kDa) (TSPO) as a therapeutic target for neurological and psychiatric dis-orders. Nat Rev Drug Discov 9:971–988. http://dx.doi.org/10.1038/nrd3295.

13. Zavala F, Haumont J, Lenfant M. 1984. Interaction of benzodiazepineswith mouse macrophages. Eur J Pharmacol 106:561–566. http://dx.doi.org/10.1016/0014-2999(84)90059-1.

14. Foss CA, Harper JS, Wang H, Pomper MG, Jain SK. 2013. Noninvasivemolecular imaging of tuberculosis-associated inflammation with radioio-dinated DPA-713. J Infect Dis 208:2067–2074. http://dx.doi.org/10.1093/infdis/jit331.

15. Diacon AH, Pym A, Grobusch M, Patientia R, Rustomjee R, Page-ShippL, Pistorius C, Krause R, Bogoshi M, Churchyard G, Venter A, Allen J,Palomino JC, De Marez T, van Heeswijk RP, Lounis N, Meyvisch P,Verbeeck J, Parys W, de Beule K, Andries K, Mc Neeley DF. 2009. Thediarylquinoline TMC207 for multidrug-resistant tuberculosis. N Engl JMed 360:2397–2405. http://dx.doi.org/10.1056/NEJMoa0808427.

16. Harper J, Skerry C, Davis SL, Tasneen R, Weir M, Kramnik I, BishaiWR, Pomper MG, Nuermberger EL, Jain SK. 2012. Mouse model ofnecrotic tuberculosis granulomas develops hypoxic lesions. J Infect Dis205:595– 602. http://dx.doi.org/10.1093/infdis/jir786.

17. Pan H, Yan BS, Rojas M, Shebzukhov YV, Zhou H, Kobzik L, HigginsDE, Daly MJ, Bloom BR, Kramnik I. 2005. Ipr1 gene mediates innateimmunity to tuberculosis. Nature 434:767–772. http://dx.doi.org/10.1038/nature03419.

18. Wang H, Pullambhatla M, Guilarte TR, Mease RC, Pomper MG. 2009.Synthesis of [125I]iodoDPA-713: a new probe for imaging inflammation.Biochem Biophys Res Commun 389:80 – 83. http://dx.doi.org/10.1016/j.bbrc.2009.08.102.

19. Davis SL, Be NA, Lamichhane G, Nimmagadda S, Pomper MG, BishaiWR, Jain SK. 2009. Bacterial thymidine kinase as a non-invasive imagingreporter for Mycobacterium tuberculosis in live animals. PLoS One4:e6297. http://dx.doi.org/10.1371/journal.pone.0006297.

20. Liu F, Whitton JL. 2005. Cutting edge: re-evaluating the in vivo cytokineresponses of CD8+ T cells during primary and secondary viral infections.J Immunol 174:5936 –5940. http://dx.doi.org/10.4049/jimmunol.174.10.5936.

21. Bhatt K, Kim A, Kim A, Mathur S, Salgame P. 2013. Equivalent func-tions for B7.1 and B7.2 costimulation in mediating host resistance to My-cobacterium tuberculosis. Cell Immunol 285:69 –75. http://dx.doi.org/10.1016/j.cellimm.2013.09.004.

22. Tasneen R, Li SY, Peloquin CA, Taylor D, Williams KN, Andries K,Mdluli KE, Nuermberger EL. 2011. Sterilizing activity of novel TMC207-and PA-824-containing regimens in a murine model of tuberculosis. An-timicrob Agents Chemother 55:5485–5492. http://dx.doi.org/10.1128/AAC.05293-11.

23. Dayan D, Hiss Y, Hirshberg A, Bubis JJ, Wolman M. 1989. Are thepolarization colors of picrosirius red-stained collagen determined only bythe diameter of the fibers? Histochemistry 93:27–29. http://dx.doi.org/10.1007/BF00266843.

24. Driver ER, Ryan GJ, Hoff DR, Irwin SM, Basaraba RJ, Kramnik I,Lenaerts AJ. 2012. Evaluation of a mouse model of necrotic granulomaformation using C3HeB/FeJ mice for testing of drugs against Mycobacte-rium tuberculosis. Antimicrob Agents Chemother 56:3181–3195. http://dx.doi.org/10.1128/AAC.00217-12.

25. Wallis RS, Kim P, Cole S, Hanna D, Andrade BB, Maeurer M, Schito M,Zumla A. 2013. Tuberculosis biomarkers discovery: developments, needs,and challenges. Lancet Infect Dis 13:362–372. http://dx.doi.org/10.1016/S1473-3099(13)70034-3.

26. Fenton MJ, Vermeulen MW. 1996. Immunopathology of tuberculosis:roles of macrophages and monocytes. Infect Immun 64:683– 690.

27. Canetti G. 1965. Present aspects of bacterial resistance in tuberculosis. AmRev Respir Dis 92:687–703.

28. Robbins SL, Kumar V. 2010. Robbins and Cotran pathologic basis ofdisease, 8th ed. Saunders/Elsevier, Philadelphia, PA.

29. Johnson DH, Via L, Kim P, Laddy D, Lau C-Y, Weinstein EA, Jain S.2014. Nuclear imaging: a powerful novel approach for tuberculosis. NuclMed Biol 41:777–784. http://dx.doi.org/10.1016/j.nucmedbio.2014.08.005.

30. Ordonez AA, Pokkali S, Mease R, Klunk M, Foss CA, Pomper MG, JainSK. 2014. Characterization of iodo-DPA-713 imaging in mice receivingnovel TB treatments, abstr 3002, p 78. In Novel therapeutic approaches totuberculosis. Keystone Symposia Conference, Keystone, CO.

31. Nuermberger E, Jain SK, Tasneen R. 2014. Cavitary TB in C3HeB/FeJmice, abstr 2047, p 75. In Novel therapeutic approaches to tuberculosis.Keystone Symposia Conference, Keystone, CO.

32. Russell DG, Cardona PJ, Kim MJ, Allain S, Altare F. 2009. Foamymacrophages and the progression of the human tuberculosis granuloma.Nat Immunol 10:943–948. http://dx.doi.org/10.1038/ni.1781.

33. Gottfried E, Kunz-Schughart LA, Weber A, Rehli M, Peuker A, MullerA, Kastenberger M, Brockhoff G, Andreesen R, Kreutz M. 2008. Ex-pression of CD68 in non-myeloid cell types. Scand J Immunol 67:453–463. http://dx.doi.org/10.1111/j.1365-3083.2008.02091.x.

34. Amanzada A, Malik IA, Blaschke M, Khan S, Rahman H, Ramadori G,Moriconi F. 2013. Identification of CD68(+) neutrophil granulocytes inin vitro model of acute inflammation and inflammatory bowel disease. IntJ Clin Exp Pathol 6:561–570. http://dx.doi.org/10.1055/s-0033-1352657.

35. Gabrilovich DI, Nagaraj S. 2009. Myeloid-derived suppressor cells asregulators of the immune system. Nat Rev Immunol 9:162–174. http://dx.doi.org/10.1038/nri2506.

36. Hickman SP, Chan J, Salgame P. 2002. Mycobacterium tuberculosisinduces differential cytokine production from dendritic cells and macro-phages with divergent effects on naive T cell polarization. J Immunol168:4636 – 4642. http://dx.doi.org/10.4049/jimmunol.168.9.4636.

Radioiodo-DPA-713 Imaging To Monitor TB Treatments

January 2015 Volume 59 Number 1 aac.asm.org 649Antimicrobial Agents and Chemotherapy

on Decem

ber 30, 2014 by WELC

H M

EDIC

AL LIBRAR

Y - John Hopkins U

http://aac.asm.org/

Dow

nloaded from