Research ArticleTuberculosis

Urine lipoarabinomannan glycan in HIV-negative patients with pulmonary tuberculosis correlates with disease severity

See allHide authors and affiliations

Science Translational Medicine  13 Dec 2017:
Vol. 9, Issue 420, eaal2807
DOI: 10.1126/scitranslmed.aal2807
  • Fig. 1 Nanocages that were covalently functionalized with copper complex dye Reactive Blue 221 sequestered and concentrated lipoarabinomannan from urine.

    (A) Schematic depicting high internal/external surface area ratio and binding capacity of nanocages. Affinity ligands covalently immobilized in the inner volume establish high-affinity noncovalent interaction with tuberculosis (TB) antigens. (B) Schematic showing the concentration factor given by the volumetric ratio between the initial urine volume and the final testing volume. Structures within the urine sample are nanocages. (C) Molecular structure of lipoarabinomannan (LAM) (right) and affinity probe Reactive Blue 221 (RB221) {cuprate(4-),[2-[[[[3-[[4-chloro-6-[ethyl[4-[[2-(sulfooxy)ethyl]sulfonyl]phenyl]amino]-1,3,5-triazin-2-yl]amino]-2-hydroxy-5-sulfophenyl]azo]phenylmethyl]azo]-4-sulfobenzoato(6-)]-,tetrahydrogen} (left). (D) Western blot, glycan staining, and image analysis of protein macroarray assay of LAM. C, LAM control (50 ng); IS, initial solution (50 ng of LAM spiked in 50 μl of human urine); S, supernatant; E, eluate from the nanocages; P, nanocages; AU, arbitrary units. Mean and SD, n = 3 replicates.

  • Fig. 2 LAM antigen was detected in the urine of HIV-negative/TB-positive patients using RB221 nanocages for diseased and control patients listed in Table 1.

    (A) Image of a quantitative immunomacroarray for LAM detection, incorporating (B) a dilution curve in every membrane. Neg, negative; BKG, background. (C) Example immunomacroarray comparing urine samples from a set of true-positive and known TB-negative samples using nanocage preprocessing. (D) Bar plot of the intensities of LAM determined via immunomacroarray and ImageJ analysis from urine samples from healthy TB-negative, TB-negative diseased, and TB-positive patients shown in Table 1 (mean ± SD, n = 4 patient replicates).

  • Fig. 3 Urinary LAM concentration predicted pulmonary TB and correlated to mycobacterial burden and weight loss.

    (A) Box plot of the intensities of LAM in the urine of HIV-negative/TB-positive patients versus controls collected in endemic areas (Wilcoxon signed-rank test). (B) Box plot of the intensities of LAM in the urine of HIV-negative/TB-positive patients stratified on the basis of the auramine staining (low amount of microorganism, scores 0 and 1; high amount of microorganism, scores 2 and 3; Wilcoxon signed-rank test; n = 42). (C) Receiver operating characteristic analysis of the LAM intensity data. AUC, area under the curve. (D) Ordinal regression analysis shows statically significant correlation between the concentration of urinary LAM and the loss of body mass (P = 0.038, n = 37).

  • Fig. 4 Nanocages captured multiple TB-related analytes.

    (A) SDS–polyacrylamide gel electrophoresis (PAGE) analysis; chemical bait incorporated in the nanocages (NP1, blue 3G-A; NP2, pigment red 177; NP3, disperse yellow 3). P, nanocage eluate. (B) Affinity probes (affinity probe 1, pigment red 177; affinity probe 2, blue 3G-A; affinity probe 3, trypan blue). (C) Nanocages effectively captured TB-related analytes from human urine (Western blot). U, negative control; C, recombinant protein (positive control, 75 ng). (D) Nanocage detection of TB antigen ESAT6 in the urine of untreated HIV-negative/TB-positive patients (Western blot).

  • Fig. 5 Magnetic hydrogel nanocages.

    (A) Schematic of magnetization. (B) Western blot analysis of ESAT6 and CFP10 expression in eluates of centrifugation-separated nanocages (top), in eluates of magnetic-separated nanocages (middle), and in supernatants after magnetic separation of nanocages from urine samples (bottom). Top and middle: Lane 1, positive control (recombinant protein; 10 ng); lanes 2 to 7, two to six eluates from nanocages incubated with 1 ml of urine containing ESAT6 (10, 5, 2.5, 1.2, 0.6, and 0.3 ng/ml) and CFP10 (10, 5, 2.5, 1.2, and 0.6 ng/ml). Bottom: Lane 1, positive control (recombinant protein; 10 ng); lanes 2 to 7, two to six supernatants after nanocage processing of 1 ml of urine containing ESAT6 (10, 5, 2.5, 1.2, 0.6, and 0.3 ng/ml) and CFP10 (10, 5, 2.5, 1.2, and 0.6 ng/ml).

  • Fig. 6 Partially dissolvable nanocages captured antigen for antibody binding in a high-sensitivity sandwich immunoassay.

    (A) Schematic demonstrating nanocage cross-link degradation in an oxidative environment. (B) Change in hydrodynamic diameter after nanocage oxidation [t test, n = 10; mean and SD of nanocage hydrodynamic diameter before (a) and after (b) oxidative degradation]. (C) SDS-PAGE analysis comparing N,N′-(1,2-dihydroxyethylene)bisacrylamide (DHEA) cross-linked nanocages mixed with a solution of monoclonal antibody (Ab) (0.05 mg/ml) with pores open (lanes 2 and 3) and closed (lanes 4 and 5). (D) Immunomacroarray demonstrating that antigen bound to the chemical bait retains its capability to bind to the antibody. a, nanocages deposited on polyvinylidene difluoride (PVDF) membrane after incubation with urine containing ESAT6 (1 ml, 10 ng/ml) and DHEA cross-link degradation; b, nanocages deposited on PVDF membrane after incubation with urine containing ESAT6 (1 ml, 10 ng/ml) in the absence of DHEA cross-link degradation; c, ESAT6 deposited on PVDF membrane (starting amount, 1 ng); d, DHEA nanocages deposited on PVDF membrane after incubation with urine in the absence of ESAT6. (E) Plot of immunoassay signal intensity as a function of bait capture affinity. High-affinity chemical baits achieve >2 log increased sensitivity for antigen capture compared to conventional antibody, as mathematically demonstrated in Supplementary Materials and Methods. (F) Schematic depicting direct, nonelution sandwich immunoassay using partially degradable nanocages. Inset shows an enzyme-linked antibody interacting with TB antigens captured inside the nanocage. (G) Calibration curve of a direct nanocage immunoassay for ESAT6 showing linearity in the 1- to 0.03-ng range. (H) Schematic of a lateral flow immunoassay using one antibody. Nanocages capture and preserve antigen in solution, migrate through the filter membrane, and provide colorimetric detection. (I) Lateral flow immunoassay for ESAT6 detection in urine. Positive signal for 10 ng of ESAT6 in 10 ml of human urine both visually (blue line, a) and with chemiluminescence (black line, b). Negative control urine in the absence of ESAT6 yields no signal (c and d).

  • Table 1 Demographic characteristics of study participants.

    IQR, interquartile range.

    Median age,
    years (IQR)
    Sex, M/F
    TB patients (microbiologically proven)29 (22–37)35/13
    Healthy volunteers26 (22–37)24/15
    Diseased TB-negative controls32 (28–51)9/5
  • Table 2 Clinical characteristics of hospitalized patients (n = 48 microbiologically confirmed TB patients and n = 2 TB-negative patients).

    Urine was collected from patients before therapy. SNAQ, simplified nutritional appetite questionnaire; MODS, microscopic observation broth-drug culture and susceptibility.

    Microbiological data
    Auramine sputum smear microscopy result
      0 (%)10
      1 (%)17
      2 (%)9
      3 (%)6
      Paucibacillary (%)7
    MODS
      Positive48
      Negative2
    Mycobacterium tuberculosis isolate sensitive to
      Isoniazid48
      Rifampicin48
    Weight (kg)52.9
    SNAQ composite score13.2
    Self-reported symptoms
      Cough
        Yes (%)38
        No (%)12
      Hemoptysis
        Yes (%)11
        No (%)39
      Fever
        Yes (%)32
        No (%)14
      Fatigue
        Yes (%)26
        No (%)24
  • Table 3 Simple and multiple linear regression analysis.

    Analysis shows that cough and SNAQ scores (20, 36), when combined, were significantly correlated to the concentration of urinary LAM. CI, confidence interval.

    Univariate (95% CI)PMultivariate (95% CI)P
    Cough0.243 (−0.253 to 0.511)0.0750.269 (0.0100 to 0.528)0.042 (n = 50)
    Eating habits0.049 (−0.001 to 0.100)0.0570.0541 (0.00476 to 0.103)0.032 (n = 37)
  • Table 4 SNAQ scoring (20, 36).
    Components of SNAQ scoreLAM > 0.115
    (n = 41)
    LAM < 0.115
    (n = 9)
    My appetite is
      Very poor20
      Poor144
      Average143
      Good91
      Very good21
    When I eat
      I feel full after eating only a few
    mouthfuls
    20
      I feel full after eating about a third
    of a meal
    20
      I feel full after eating more than half
    a meal
    103
      I feel full after eating most of the
    meal
    246
      I hardly ever feel full30
    Food tastes
      Very bad00
      Bad76
      Average241
      Good101
      Very good01
    Normally I eat
      Less than one meal a day00
      One meal a day30
      Two meals a day70
      Three meals a day266
      More than three meals a day53
  • Table 5 Ordinal regression analysis.

    A significant correlation between the urinary LAM concentration and body mass change was observed.

    Low-level LAM
    (n = 6)
    (115 pg/ml)
    Mid-level LAM
    (n = 15)
    (116–319 pg/ml)
    High-level LAM
    (n = 16)
    (>320 pg/ml)
    Mean percent
    weight change
    −7.86 kg−8.63 kg−16.97 kg
    Odds ratio0.933
    (95% CI)(0.0873–0.996)
    P0.038 (n = 37)
  • Table 6 Hydrogel nanocage synthesis: Quantity (in millimoles) of comonomers and total volume of reaction.

    NIPAm, N-isopropylacrylamide; BIS, N,N′-methylenebisacrylamide; DHEA, N,N′-(1,2-dihydroxyethylene)bisacrylamide; AAc, acrylic acid; KPS, potassium persulfate, TEMED, N,N,N′,N′-tetramethylethylenediamine; BAC, N,N′-bis(acryloyl)cystamine; AA, allylamine; NHA, N-(hydroxymethyl)acrylamide.

    NIPAmBISDHEABACAAcAANHATEMEDKPSVolume
    1:NBaAl390.94.50.170.18150
    2:NBiAc422.67.31.02500
    3:NBiDAc6.160.640.640.560.1770
    4:NBiNh422.67.31.02500
  • Table 7 Dye coupling to cages.

    RB221, reactive blue 221; RBB, remazol brilliant blue R; DY3, disperse yellow 3; PR117, pigment red 177; FB28, fluorescent brightener 28.

    RB221Trypan blue (17)RBB (14)DY3 (14)PR177 (14)FB28
    1:NBaAl2:NBiAc, 3:NBiDAc2:NBiAc2:NBiAc2:NBiAc4:NBiNh

Supplementary Materials

  • www.sciencetranslationalmedicine.org/cgi/content/full/9/420/eaal2807/DC1

    Materials and Methods

    Fig. S1. The Kd affinity between RB221 and LAM exceeds that of FB28.

    Fig. S2. Copper dyes outperform copper free dyes such as fast blue B and safranin O.

    Fig. S3. Nanocages dissociate biomarker from interfering substances, in silico mathematical modeling.

    Fig. S4. CS-35 mAb is specific for LAM diluted in human urine, batch verification.

    Fig. S5. Competition assay confirmed the specificity of CS-35 mAb.

    Fig. S6. Coupling chemistry to covalently incorporate the FB28 dye in the inner volume of the nanocages.

    Fig. S7. LAM binding to RB221 and depletion from supernatant are independent of pH in a 5 to 7 range.

    Fig. S8. RB221 binding to LAM is hindered by the presence of a copper-chelating agent (EDTA).

    Fig. S9. RB221-LAM interaction requires intact diol moieties of LAM as proven by NaIO4 oxidation.

    Fig. S10. Carbohydrate concentration in the LAM reference standard (0.160 mg/ml) was quantified by a linear colorimetric assay.

    Fig. S11. Plot of the 95% CI of the sensitivity and specificity of the ROC analysis reported in Fig. 3C.

    Fig. S12. The RB221 dye is immobilized in the inner volume of the cages and is available for high–molecular weight ligand binding after cross-link degradation and consequent increase of the effective pore size.

    Fig. S13. CS-35 anti-LAM mAb does not cross-react with purified polysaccharides from N. meningitidis and S. pneumoniae.

    Fig. S14. Nanocage capturing followed by CS-35 antibody detection is specific for LAM and does not cross-react with M. tuberculosis lipomannan and arabinogalactan.

    Table S1. Nanocage bait chemistries screened to capture and enrich LAM from human urine.

    Table S2. Medical characteristics of diseased TB-negative controls.

    Table S3. Urinalysis results for all study participants.

  • Supplementary Material for:

    Urine lipoarabinomannan glycan in HIV-negative patients with pulmonary tuberculosis correlates with disease severity

    Luisa Paris, Ruben Magni, Fatima Zaidi, Robyn Araujo, Neal Saini, Michael Harpole, Jorge Coronel, Daniela E. Kirwan, Hannah Steinberg, Robert H. Gilman, Emanuel F. Petricoin, Roberto Nisini, Alessandra Luchini,* Lance Liotta

    *Corresponding author. Email: aluchini{at}gmu.edu

    Published 13 December 2017, Sci. Transl. Med. 9, eaal2807 (2017)
    DOI: 10.1126/scitranslmed.aal2807

    This PDF file includes:

    • Materials and Methods
    • Fig. S1. The Kd affinity between RB221 and LAM exceeds that of FB28.
    • Fig. S2. Copper dyes outperform copper free dyes such as fast blue B and safranin O.
    • Fig. S3. Nanocages dissociate biomarker from interfering substances, in silico mathematical modeling.
    • Fig. S4. CS-35 mAb is specific for LAM diluted in human urine, batch verification.
    • Fig. S5. Competition assay confirmed the specificity of CS-35 mAb.
    • Fig. S6. Coupling chemistry to covalently incorporate the FB28 dye in the inner volume of the nanocages.
    • Fig. S7. LAM binding to RB221 and depletion from supernatant are independent of pH in a 5 to 7 range.
    • Fig. S8. RB221 binding to LAM is hindered by the presence of a copper-chelating agent (EDTA).
    • Fig. S9. RB221-LAM interaction requires intact diol moieties of LAM as proven by NaIO4 oxidation.
    • Fig. S10. Carbohydrate concentration in the LAM reference standard (0.160 mg/ml) was quantified by a linear colorimetric assay.
    • Fig. S11. Plot of the 95% CI of the sensitivity and specificity of the ROC analysis reported in Fig. 3C.
    • Fig. S12. The RB221 dye is immobilized in the inner volume of the cages and is available for high–molecular weight ligand binding after cross-link degradation and consequent increase of the effective pore size.
    • Fig. S13. CS-35 anti-LAM mAb does not cross-react with purified polysaccharides from N. meningitidis and S. pneumoniae.
    • Fig. S14. Nanocage capturing followed by CS-35 antibody detection is specific for LAM and does not cross-react with M. tuberculosis lipomannan and arabinogalactan.
    • Table S1. Nanocage bait chemistries screened to capture and enrich LAM from human urine.
    • Table S2. Medical characteristics of diseased TB-negative controls.
    • Table S3. Urinalysis results for all study participants.

    [Download PDF]

Navigate This Article