A streamlined CRISPR-based test for tuberculosis detection directly from sputum

Abstract

Mycobacterium tuberculosis (Mtb) is a major threat to global health, and there is an urgent need for affordable, simple tuberculosis (TB) diagnosis in underresourced areas. Here, we combine recombinase polymerase amplification with Cas13a and Cas12a detection to create two parallelized one-pot assays that detect two conserved elements of Mtb (IS6110 and IS1081) and a human DNA internal control. These assays are compatible with lateral flow and can be readily lyophilized. Our final assay showed a limit of detection of 69.0 CFU per milliliter for Mtb H37Rv and 80.5 CFU per milliliter for Mycobacterium bovis BCG in spiked sputum, with no cross-reactivity to diverse bacterial or fungal isolates. Clinical tests on 13 blinded sputum samples revealed 100% (six of six) sensitivity and 100% (seven of seven) specificity compared to culture. SHINE-TB streamlines TB diagnosis from sample to answer by combining amplification and detection while being compatible with lateral flow and lyophilization.

Fig. 1.. CRISPR-Cas12a assay optimization for detecting Mtb.

(A) Schematic of Cas-12a/Cas13a-based CRISPR reaction, where DNA is amplified using target specific primers, recombinases, strand-displacing polymerase, and single-stranded DNA (ssDNA)–binding proteins and directly detected by Cas12a or transcribed to RNA with T7 reverse transcription for detection with Cas13a. (B) Cas12a detecting synthetic DNA for six different crRNA-primer designs. Amplification was allowed to proceed for 30 min before the addition of crRNA and Cas12a. (C) Cas12a detecting synthetic DNA with varying concentrations of Cas12a. Molar concentration alone indicates all components present from the start of the reaction. Molar concentration spiked indicates amplification was allowed to proceed for 30 min before the addition of crRNA and Cas12a. (D) Cas12a detecting synthetic DNA in a one-pot format, as all components are present from the start of the reaction. (E) Performance of Cas12a assay for decreasing concentrations of H37Rv genome. All bars are significant with P_adj < 0.0001 except for 10⁰ that has P = 0.0410. (F) Specificity panel for Cas12a assay against various NTMs and other bacteria/fungi at 1 ng/μl. All bars except for M. tuberculosis are not significant (n.s.). In (B) and (D), error bars: SD based on n = 2 technical replicates. In (C), (E), and (F), error bars: SD based on n = 3 technical replicates. In (E) and (F), one-way analysis of variance (ANOVA) with Dunnett’s test was performed, and P_adj values were calculated compared to the nontarget control (NTC) that contained water. a.u., arbitrary units.

Fig. 2.. CRISPR-Cas13a–based detection of Mtb.

(A) Cas13a detecting synthetic targets for six different crRNA-primer designs against two different targets, IS6110 and IS1081. (B) Comparison of Cas12a to Cas13a detection of IS6110 synthetic target. Cas13a is a single-crRNA assay for direct comparison. (C) Cas13a detecting H37Rv in single-target detection [IS6110 only; P_adj < 0.0001 for all bars or IS1081 only; P_adj (left to right) is <0.001, <0.001, <0.001, <0.01, and <0.05] and dual-target detection (IS6110 and IS1081; P_adj < 0.0001 for all bars) formats. (D) Cross-reactivity panel for Cas13a dual-detection assay against various NTMs and other pathogens at 0.1 ng/μl. All samples are n.s. except for positive controls. In all panels, error bars: SD based on n = 3 technical replicates. In (C) and (D), one-way ANOVA with Dunnett’s test was performed and P_adj values were calculated compared to NTC that contained water. a.u., arbitrary units.

Fig. 3.. Validation of Cas13a dual-detection assay against clinically relevant samples.

(A) Cas13a detecting BCG genomic DNA spiked into sputum:TE matrix and Cas12a detecting internal control. Sp is sputum, and gBCG is genomic BCG. Units are genomes per microliter. Cas13a reporter is FAM, and Cas12a reporter is HEX. (B) Comparison of fresh Cas13a (left) or lyophilized Cas13a (right) assays detecting synthetic internal control (IC) and BCG genomic DNA. Synthetic IC and genomic BCG were mixed at a concentration of 10⁴ and 10³ copies per microliter, respectively. (C) Lateral flow readout of Cas13a detecting dual targets IS6110 + IS1081 (top) and Cas12a detecting internal control (bottom) in BCG spiked into sputum:TE matrix. Appearance of the top line is a positive readout. Faint lines are interpreted as negative. Disappearance of the bottom line is not necessary. (D) Range finding experiment demonstrating the dynamic range of fresh Cas13a dual-detection assay. Cas13a detecting BCG and Cas12a detecting internal control. BCG culture [left; P_adj (left to right) is <0.001, <0.001, <0.001, <0.01, <0.05, n.s., n.s., and n.s.] and BCG spiked into sputum:TE [right; P_adj (left to right) is <0.0001, <0.0001, <0.0001, <0.05, n.s., n.s., and n.s.]. gBCG is genomic BCG and units are genomes per microliter. All other units are colony-forming units (CFU) per milliliter. (E) Testing fresh Cas13a and Cas12a coupled assay against TB-negative clinical sputum samples. Numbers indicate patient number. All samples are n.s. except Sp-17, where P = 0.015. (F) LoD for H37Rv. LoD = 69.0 (51.0 to 86.9). (G) LoD for BCG. LoD = 80.5 (59.4 to 101.6). In (A), (B), (D), and (E), error bars: SD based on n = 3. (C) A single representative sample. (F) n = 20 and (G) n = 21. In (D) and (E), one-way ANOVA with Dunnett’s test was performed, and P_adj values were calculated compared to NTC that contained water. a.u., arbitrary units.

Fig. 4.. TB detection from sputum using fresh SHINE-TB.

(A) Schematic showing workflow of patient sample testing. Sputum samples were collected and tested using GeneXpert at CIDEIM in Colombia. Samples were shipped to Rutgers for sample processing. Processed samples were used for CRISPR assays and dPCR. (B) TB detection in 14 different sputum samples using Cas13a dual detection. For each sample, a line is shown at the mean of n = 3 technical replicates. Shading: gray, controls; green, correct call; yellow, unmatched call for GeneXpert versus CRISPR and culture. LoD is shown as a dotted line at 2900 a.u. Triplicates were averaged for determination of positive and negative. Positive sputum is 500 CFU/ml. Plus signs mean positive detection, and minus signs mean no detection. C means contamination/not determined. ND, not done or data not collected. (C) Tables showing Cas13a dual detection compared to culture (top left), GeneXpert (top right), and dPCR (middle right). Table showing GeneXpert versus culture (bottom right). For further details, see table S5. NA, data not collected. (D) Average Cas13a (FAM) versus Cas12a internal control (HEX) signal for each sample seen in (B). Linear trend was determined for the positive samples with a coefficient of determination (R²) value of 0.7058. a.u., arbitrary units.