Non-Invasive Detection of Tuberculosis Biomarkers from Skin Using UHPLC–IMS-HRMS

LCGC International spoke with Madelien Wooding from the Department of Chemistry at the University of Pretoria, South Africa, about her team’s work exploring the skin surface as a novel, non-invasive matrix for detecting tuberculosis (TB) biomarkers. The study highlights the potential of skin-based diagnostics to offer safe, cost-effective, and field-friendly solutions for TB screening, particularly in resource-limited settings , using ultrahigh-pressure liquid chromatography coupled with ion mobility and high-resolution mass spectrometry (UHPLC–IMS-HRMS).

University of Pretoria campus in evening light with Jacarandas © Stoffel - stock.adobe.com.

Q: What inspired you to explore skin as a matrix for tuberculosis (TB) biomarker detection (1), given the dominance of breath, sputum, and blood in existing research?

The primary motivation was the unmet need for a non-invasive, cost-effective, and non-infectious diagnostic method that bypasses the biohazard risks of breath or sputum. While blood and sputum dominate TB diagnostics, they often require skilled personnel and pose biosafety challenges. Skin, by contrast, is a non-infectious matrix and accessible for sampling. Prior work by Turner et al. and our own group (2–3) had shown that many volatile compounds in breath are also present on skin. This inspired us to investigate whether semi-volatile and non-volatile metabolites relevant to TB could also be detected on the skin surface.

Q: What were the key considerations in selecting ultrahigh-pressure liquid chromatography (UHPLC) as the primary separation technique in your analysis?

UHPLC was chosen due to its ability to handle complex biological mixtures with high resolution and speed. The skin surface presents a diverse chemical matrix ranging from highly polar to non-polar compounds. UHPLC, particularly when paired with ion mobility and high-resolution mass spectrometry (HRMS), allows for sensitive, reproducible, and high-throughput separation. This technique’s robustness was critical for handling the narrow elution windows required for high-precision metabolomics.

Q: How did you optimize the chromatographic conditions—such as column selection, gradient, and mobile phase composition—to handle the diverse chemical properties of skin-emitted metabolites?

We employed a reverse-phase gradient on a 2.1 × 150 mm, 1.8-µm column, suited for polar and semi-polar analytes. A stepwise linear gradient (3% to 100% acetonitrile with 0.1% formic acid over 14 min) allowed optimal resolution. Column temperature (40 °C), flow rate (0.4 mL/min), and injection volume (5 µL) were fine-tuned to maintain peak integrity and reproducibility. This configuration offered balanced retention for both hydrophilic and hydrophobic compounds.

Q: Can you discuss how ion mobility spectrometry (IMS) complements your chromatographic separation?

IMS added a powerful orthogonal dimension of separation. It resolved coeluting isobaric compounds by drift time and provided collisional cross-section (CCS) values, enhancing compound annotation. This helped distinguish closely eluting biomarkers such as umbelliprenin and ramiprilat. IMS also improved data-independent acquisition (DIA) by aligning precursor and fragment ions, increasing confidence in metabolite identification, and enabling more accurate multivariate modeling.

Q: How did you determine the 14 biomarkers identified in your study, and what criteria were used to select them for further analysis?

We used a supervised orthogonal partial least squares discriminant analysis (OPLS-DA) model and receiver operating characteristic (ROC) curve analysis on data from electrospray ionization (ESI+) mode. From over 14,000 features, 463 were statistically significant (p < 0.05). Variables of Importance to Projection (VIP > 1) and fold-change thresholds (>1.5) helped shortlist candidates. The most predictive features were annotated using CCS prediction (CCSondemand), HMDB, and ChemSpider (all open-access software). Finally, 14 features were selected based on their strong area under the curve (AUC) values (>0.9), indicating excellent diagnostic potential.

Q: Para-aminobenzoic acid (PABA) showed good diagnostic performance—what is known about its biological relevance to TB?

PABA is involved in folate biosynthesis and degradation pathways, essential for nucleotide metabolism in both host and pathogen. Its elevation in TB-positive skin samples (up to approx. 4198 ng) suggests altered host-pathogen metabolic interactions. Importantly, it had an AUC of 0.956, specificity of 1, and sensitivity of 0.9, outperforming previous serum-based studies where its AUC was ~0.593. This highlights PABA's promise as a non-invasive TB biomarker when sampled from skin.

Q: Were there any challenges encountered during sample collection and processing using the wearable polydimethylsiloxane (PDMS) sampler?

Yes, maintaining sample integrity during storage (at -18°C) and ensuring consistent skin contact without causing discomfort were initial concerns. PDMS, being non-polar, also required ethanol pretreatment to enhance the uptake of polar compounds. Moreover, we had to balance ethical constraints—especially for untreated TB patients—with analytical needs. A 1-h sampling period was chosen to optimize peak detection without delaying clinical treatment initiation.

Q: Were there any surprising findings in your metabolomic data that may suggest new avenues for TB biomarker research?

Indeed. The identification of 2,6-diethylaniline and umbelliprenin—both previously unreported in TB contexts—was unexpected. Their presence may reflect environmental or dietary exposures, or even self-medication. Another surprising aspect was the absence of kynurenine as a differentiating marker, despite its promise in blood-based studies. These findings open questions about skin-specific metabolic expressions and warrant further investigation into the skin exposome in TB patients.

Q: What are your next steps for this research?

Our immediate goal is to validate these findings in a larger, independent cohort and expand sampling to include latently infected individuals. We aim to refine the diagnostic panel using targeted metabolomics and explore machine learning for predictive modeling. Long-term, we envision a field-deployable skin-based diagnostic tool that could revolutionize TB screening, particularly in low-resource settings.

References

(1) Wooding, M.; Van Pletzen, K.; Naudé, Y. Identifying Skin Surface Chemicals as Potential Tuberculosis Diagnostic Biomarkers Using Ultra Performance Liquid Chromatography-High Resolution Mass Spectrometry. J. Chrom. Open 2025, 7, 100204. DOI: 10.1016/j.jcoa.2025.100204

(2) Turner, C.; B. Parekh, B.; Walton, C.; et al. An Exploratory Comparative Study of Volatile Compounds in Exhaled Breath and Emitted by Skin Using Selected Ion Flow Tube Mass Spectrometry. Rapid Commun. Mass Spectrom 2008, 22, 526–532. DOI: 10.1002/rcm.3432

(3) Makhubela, P.C.; Rohwer, E. R.; Naud´e, Y. Detection of Tuberculosis-Associated Compounds from Human Skin by GCxGC-TOFMS. J. Chromatogr. B 2023, 1231, 123937, DOI: 10.1016/j.jchromb.2023.123937

Madelien Wooding © Image courtesy of interviewee

Madelien Wooding is a researcher in the Department of Chemistry in the Faculty of Natural and Agricultural Sciences at the University of Pretoria, South Africa. She manages the liquid chromatography and high-resolution mass spectrometry facility. At the forefront of her primary research, Wooding uses mass spectrometry methodologies to delve into intricate chemical compositions, enabling the detection and characterization of minute quantities of substances in various matrices. With a keen focus on real-world applications, Wooding's research extends beyond conventional boundaries. In addition to investigating chemical pollutants in water and leaching chemicals from food packaging materials, she ventures into innovative territories, such as probing the chemical landscape of the human skin surface. By employing advanced mass spectrometry techniques, she explores the intricate array of chemicals present on the skin, unveiling potential biomarkers and molecular signatures that hold significant diagnostic value. Her contributions to science have been recognized with several prestigious awards. In 2018, Wooding received the L’Oréal-UNESCO Fellowship for Women in Science. In 2022, she was honoured with The Conversation Africa Award for Communication Excellence for her efforts in making scientific knowledge accessible to the public.