Plasma sphingolipid and glycerophospholipid shifts during early treatment in zinc-supplemented adults with HIV–tuberculosis co-infection: a paired exploratory lipidomics study

Plasma sphingolipid and glycerophospholipid shifts during early treatment in zinc-supplemented adults with HIV–tuberculosis co-infection: a paired exploratory lipidomics study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Plasma sphingolipid and glycerophospholipid shifts during early treatment in zinc-supplemented adults with HIV–tuberculosis co-infection: a paired exploratory lipidomics study Chao Chen, ShiYun Chen, YingYing Zhang, Man Rao, HaiTao Zhang, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9309714/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Background: Zinc deficiency is frequent in HIV and may impact sphingolipid metabolism via the zinc-dependent secretory acid sphingomyelinase. To date, associations of early treatment with zinc-supplemented adults with HIV/tuberculosis (TB) co-infection with plasma lipidomic alterations have not been explored. Methods: Paired fasting plasma samples were collected at baseline and ~4 weeks from 14 antiretroviral therapy (ART)-naive adults with HIV/TB co-infection initiating oral zinc gluconate (40 mg/day elemental zinc), standard anti-TB therapy (from baseline), and ART (from week 2). Untargeted UHPLC-MS/MS lipidomics detected 1,230 lipid species, and targeted metabolomics quantified 42 energy-related metabolites. Candidate lipids were prioritized by paired Wilcoxon p-values, within-subject effect sizes, signal completeness, and bootstrap confidence intervals. Results: At follow-up, plasma zinc concentrations were higher (11.32 ± 1.39 to 13.06 ± 1.78 μmol/L; p = 0.001), and CD4+ T-cell counts had doubled (95.8 ± 75.7 to 191.6 ± 96.0 cells/μL; p = 0.0004). No lipid met the global Benjamini–Hochberg correction (all q > 0.30). From the 81 nominally significant species (p < 0.05), four lipids with complete observed data, largest effect sizes (r = 0.65–0.76), and bootstrap confidence intervals (CIs) that excluded 1.0 were mechanistically prioritized: Cer d18:1/26:0 (fold change [FC] 1.31), SM 40:5 (FC 1.44), SM 37:1 (FC 1.17), and PG 16:1-22:1 (FC 1.35), all increased at follow-up. Lactate was the only energy metabolite with nominal significance (p = 0.009). Conclusion: These hypothesis-generating results highlight four candidate lipids with consistent, within-subject shifts in response to early HIV/TB treatment. The effects of zinc supplementation, anti-TB therapy, ART, and immune reconstitution cannot be distinguished. zinc supplementation HIV-tuberculosis co-infection lipidomics sphingolipid metabolism phosphatidylglycerol exploratory analysis Figures Figure 1 Figure 2 Figure 3 1. Introduction Tuberculosis (TB) is the world’s deadliest infectious disease with 1.25 million deaths and 10.8 million new cases in 2023 [ 1 ]. A substantial proportion of these new TB cases occur among people living with HIV, with the latter being at highest risk. Paradoxical clinical worsening and metabolic instability can occur during the first weeks of treatment, when CD4 + T-cell recovery from antiretroviral therapy (ART) begins [ 2 ]. How early immune reconstitution impacts lipid, glucose, and protein metabolism remains poorly characterized, particularly when nutritional supplementation is also provided. Zinc deficiency is common among people with HIV [ 3 , 4 ], particularly in resource-constrained settings where TB is most prevalent. Low zinc status has been linked with adverse clinical outcomes, including more rapid disease progression and more opportunistic infections. Prior work has also suggested that long-term zinc supplementation may decrease the risk of immunological failure [ 5 ]. In the setting of TB, zinc co-supplementation during anti-TB treatment has been associated with modest improvements in sputum conversion [ 6 , 7 ]. The impact of zinc supplementation on the host lipidome during early immune reconstitution in HIV/TB co-infection has not yet been studied. One mechanistic connection between zinc and lipid metabolism is the zinc-dependent secretory acid sphingomyelinase (S-aSMase) pathway, which cleaves sphingomyelin to ceramide and phosphocholine. Activity of this pathway is dependent on the exogenous addition of zinc [ 8 – 10 ]. Ceramide is a bioactive second messenger involved in many aspects of cellular function, including programmed cell death and inflammation [ 11 ]. Consequently, zinc availability could plausibly influence sphingomyelin/ceramide balance and thereby perturb phospholipid composition. To our knowledge, evidence on whether oral zinc supplementation is associated with changes in plasma sphingolipid–glycerophospholipid profiles during early treatment in HIV/TB co-infection is lacking. Lipid metabolism is already perturbed in HIV infection, and the initiation of ART further reconfigures the plasma lipidome. ART initiation, including integrase inhibitor-based regimens, has been associated with changes in body composition and lipid metabolism in treatment-naive patients [ 12 ]. At the molecular level, Chaudhary et al. reported that different ART regimens induced divergent responses in ceramide, sphingomyelin, and phosphatidylethanolamine species [ 13 ]. These studies suggest that the plasma lipidome is responsive to ART in HIV. In contrast, the lipidomic effects of nutritional supplementation, including zinc, have been largely overlooked, particularly during the first weeks of treatment when immune reconstitution, anti-TB drug effects, and nutritional repletion all converge. Accordingly, paired plasma samples from 14 ART-naive adults with HIV/TB co-infection who received zinc supplementation during the first month of treatment were analyzed in this study. In this cohort, zinc supplementation and standard anti-TB therapy were initiated at baseline, whereas ART was started 2 weeks later. The primary objective was to identify within-subject lipid shifts that were sufficiently consistent and analytically robust to warrant targeted future investigation. Given the small exploratory sample size and pilot design of the present study, candidate lipids were prioritized using a combined assessment of effect sizes, confidence intervals, imputation resistance, and transparent reporting of global false-discovery rates. 2. Materials and Methods 2.1 Study Design and Ethical Approval This prospective paired single-arm study was conducted at Shenzhen Third People’s Hospital between 2023 and 2024 as a substudy nested within a larger randomized controlled trial (NCT05847715) of zinc supplementation in HIV/TB co-infected adults. The protocol was approved by the Institutional Ethics Committee of Shenzhen Third People’s Hospital (approval number: 2023-045-02). Written informed consent was obtained from all participants. The study was conducted in accordance with the Declaration of Helsinki. 2.2 Study Population Adults between 18 and 70 years of age with confirmed HIV infection and newly diagnosed active pulmonary TB were eligible. All participants were ART-naive at recruitment. Standard anti-TB therapy was initiated at enrollment, whereas ART was started at the scheduled 2-week follow-up visit. Inclusion criteria were: (1) confirmed HIV-1 infection by Western blot or nucleic acid testing; (2) active pulmonary TB diagnosed by sputum smear, culture, or GeneXpert MTB/RIF; (3) no prior ART exposure; and (4) willingness to adhere to the supplementation protocol. Patients with hepatic insufficiency (ALT > 5× upper limit of normal), renal failure (eGFR < 30 mL/min/1.73 m²), concurrent malignancy, pregnancy, or recent use of zinc-containing preparations were excluded. 2.3 Intervention and Sample Collection Participants began oral zinc gluconate (40 mg elemental zinc per day) and standard first-line anti-TB therapy (isoniazid, rifampin, ethambutol, and pyrazinamide; HRZE) on the day of enrollment, after baseline blood sampling. An intermediate follow-up visit was conducted approximately 2 weeks later, at which antiretroviral therapy (ART) was initiated. The ART regimen consisted of lamivudine, tenofovir disoproxil fumarate, and dolutegravir, with dolutegravir dosed twice daily because of concomitant rifampin use [ 25 ]. No blood sample from the 2-week visit was included in the present paired lipidomics analysis. The second fasting blood sample for the present study was collected approximately 4 weeks after enrollment. Thus, the follow-up sample reflected approximately 4 weeks of zinc supplementation and anti-TB therapy, as well as approximately 2 weeks of ART exposure (Supplementary Fig. S1 ). Plasma was separated within 2 hours, aliquoted, and stored at − 80°C. No untreated or non-zinc comparator arm was included in the present lipidomics substudy. Zinc supplementation adherence was assessed by pill counts at each visit; all 14 participants maintained full adherence. 2.4 Clinical and Biochemical Assessments Fasting blood was collected for routine clinical parameters at each time point: complete blood count, hepatic function (ALT, AST, GGT, albumin, prealbumin), renal function (uric acid, eGFR), fasting glucose, ferritin, and a lipid panel (total cholesterol, triglycerides, HDL-C, LDL-C, apolipoprotein A1, apolipoprotein B, lipoprotein(a), small dense LDL). Plasma zinc was measured by inductively coupled plasma mass spectrometry. CD4 + T-cell counts were determined by flow cytometry, and interleukin-6 (IL-6) by chemiluminescent immunoassay. 2.5 Untargeted Lipidomics Analysis Lipids were extracted from 100 µL plasma using methyl tert-butyl ether/methanol (Matyash protocol) [ 14 ]. Chromatographic separation used a Waters ACQUITY UPLC CSH C18 column (2.1 × 100 mm, 1.7 µm) at 55°C with acetonitrile/water (60:40) and isopropanol/acetonitrile (90:10) mobile phases, both containing 10 mM ammonium formate and 0.1% formic acid. Data were acquired in positive and negative ESI modes on a Thermo Scientific Q Exactive Plus mass spectrometer. QC samples were prepared by pooling equal aliquots from all study samples and injected regularly throughout the run. Pearson correlation coefficients among QC replicates were > 0.99. Lipid species were identified by matching accurate mass and MS/MS spectra against HMDB [ 15 ], LIPID MAPS [ 16 ], and an in-house spectral library. The extraction of metabolite ion peaks and the identification of metabolites are carried out using the quantitative software Skyline [v.21.1.0.146]. This will result in a data matrix containing information such as the identification results and quantitative results of the metabolites. Subsequently, further information analysis and processing will be conducted on this table. for peak detection, alignment, and feature extraction. A total of 1,230 species spanning 10 subclasses passed basic quality control. Data preprocessing: Peak intensities were normalized to internal standards. Among the 1,230 detected species, 696 features had missingness ≤ 20% across the 28 observations; the remaining 534 features had missingness > 20%. Missing values were imputed using k-nearest neighbors for features judged to be missing at random and by the minimum observed value after normalization for values judged to be below the detection limit. For exploratory hypothesis screening, paired Wilcoxon tests were applied to all 1,230 features to avoid excluding potentially informative low-abundance species at the testing stage. However, candidate prioritization and biological interpretation were restricted to features with observed non-imputed intensities in all 14 paired samples (i.e., complete-data features), as described in Section 2.7 . Features with missingness > 20% were retained in unsupervised (PCA) and exploratory supervised (OPLS-DA) multivariate overviews, where they contribute to the overall variance structure. No batch correction was applied because all samples were analyzed in a single analytical batch. Because low-abundance features can generate unstable fold-change estimates after imputation, we prespecified a sensitivity analysis to distinguish robust observed signals from imputation-sensitive signals in the downstream interpretation. 2.6 Targeted Energy Metabolomics Eighty MRM transitions covering central carbon and energy metabolites were monitored using a Waters ACQUITY H-ClassD UPLC coupled to an AB Sciex QTRAP 6500 + mass spectrometer in scheduled multiple reaction monitoring mode. Separation used a BEH Amide column (2.1 × 100 mm, 1.7 µm) at 60°C with gradient elution in both positive and negative ESI modes. Peak integration and quantification were performed using MultiQuant software (version 4.0, SCIEX). Fifty-one metabolites were quantifiable above the lower limit of quantification, and 42 of these passed quality filters for paired statistical testing (≥ 5 complete pairs with both time points above the detection limit). 2.7 Statistical Analysis All statistical analyses were performed in R (version 4.3.3; R Foundation for Statistical Computing, Vienna, Austria). Multivariate analyses were conducted using the ropls package (version 1.34.0, Bioconductor 3.18) [ 17 ]. Imputation of missing values (described in Section 2.5 ) was performed using the impute package (version 1.76.0, Bioconductor; k = 10). Figures were generated using the matplotlib library (version 3.8, Python 3.11) from precomputed R outputs. Given the exploratory pilot design and the small paired sample size (n = 14), the study was not powered to confirm findings across 1,230 lipid features. Instead, we adopted a robustness-based candidate prioritization framework rather than a fold-change-threshold screening framework. For each of the 1,230 lipid species, paired differences between baseline and follow-up were assessed using the exact Wilcoxon signed-rank test (wilcox.test with exact = TRUE, stats package, base R). All reported p-values in the final manuscript were derived from the same statistical implementation, and all figures and tables were generated from a single precomputed master results table to ensure internal consistency. This broad exploratory screen included high-missingness features whose fold-change estimates may be unstable; downstream candidate prioritization was restricted to complete-data features as described below. Global Benjamini–Hochberg (BH) adjusted q-values were calculated and are reported transparently, but were not used as the sole basis for candidate nomination in this pilot dataset [ 18 ]. To describe signal strength beyond p-values, we calculated the Wilcoxon effect size (r = |Z|/√N), with r ≥ 0.5 considered large [ 19 ]. Fold change was summarized as the median of within-pair fold changes (follow-up/baseline for each participant) and accompanied by bootstrap 95% confidence intervals (10,000 resamples with replacement, seed = 42; boot package, version 1.3–28.1). Primary candidate lipids were prioritized using five criteria considered jointly: (1) nominal paired p < 0.05; (2) large within-subject effect size (r ≥ 0.5); (3) observed non-imputed intensities across all 14 paired samples; (4) bootstrap 95% confidence intervals for fold change that excluded 1.0; and (5) directional consistency with the a priori sphingolipid-related biological hypothesis (upregulated species). Complete data lipids that met criteria 1–4 but were not prioritized under criterion 5 are reported transparently in Supplementary Table S1 . Lipids with repeated values at or below the imputation threshold (< 1 × 10⁻⁴) were classified as imputation-sensitive exploratory signals and were not used for mechanistic interpretation. PCA was used for an unsupervised overview. For exploratory supervised analysis, OPLS-DA was performed on log10-transformed intensities with mean centering (ropls 1.34.0; 1 predictive and 1 orthogonal component). Model validity was assessed using a 200-fold permutation test (seed = 42). Because cross-validated predictive performance was limited, VIP scores were not used as mandatory criteria for feature selection. Class-level BH correction within lipid subclasses was performed as a supplementary analysis. KEGG mapping and pathway annotation were used descriptively when available [ 20 ]. Exploratory Spearman correlations among nominally changing lipids were calculated; however, only six species met the threshold (|r| > 0.8 and p < 0.05), yielding insufficient nodes for a meaningful network visualization, and this analysis was therefore not presented as a standalone figure. Routine clinical parameters were compared using paired Wilcoxon signed-rank tests (two-sided p < 0.05). For the targeted metabolomics panel, paired Wilcoxon tests and BH-adjusted q-values were calculated separately from the lipidomics dataset. Only metabolites with at least five complete paired observations above the detection limit were included. 3. Results 3.1 Participant Characteristics Fourteen HIV-1-infected adults (11 male and 3 female; mean age 41.1 ± 13.0 years; mean BMI 19.34 ± 4.09 kg/m2) completed the paired sampling protocol (all ART-naive at enrollment). The mean ± SD time between baseline and follow-up sampling was 29.6 ± 1.9 days (range 28–34; Table 1 ). Zinc supplementation and standard four-drug anti-TB therapy were begun after the baseline sampling, and ART was initiated at the scheduled 2-week follow-up visit. Plasma zinc increased from 11.32 ± 1.39 µmol/L to 13.06 ± 1.78 µmol/L (p = 0.001). CD4 + T-cell counts rose from 95.8 ± 75.7 to 191.6 ± 96.0 cells/µL (p = 0.0004), consistent with early treatment-related immune recovery. IL-6 declined by 72.2% (30.66 ± 33.09 to 8.52 ± 9.66 pg/mL; p = 0.058). 3.2 Clinical Metabolic Parameters Several other routinely obtained clinical parameters changed in the expected directions during the study period (Table 2 ). HDL-C increased by 51.6% (0.89 ± 0.39 to 1.35 ± 0.51 mmol/L; p = 0.011), apolipoprotein B rose by 36.3% (p = 0.011), and lipoprotein(a) decreased by 50.6% (p = 0.028). Small dense LDL increased by 46.9% (0.49 ± 0.23 to 0.72 ± 0.27 mmol/L; p = 0.026). Albumin increased by 9.2% (p = 0.028) and prealbumin by 34.4% (p = 0.035). Fasting glucose rose modestly (p = 0.035). Ferritin declined from 812.5 ± 514.0 to 401.5 ± 183.2 ng/mL (p = 0.002), and uric acid increased from 374.1 ± 110.8 to 543.2 ± 206.1 µmol/L (p = 0.013). Total cholesterol, triglycerides, LDL-C, apolipoprotein A1, and hepatic transaminases did not change significantly. All changes occurred during the first month of treatment, during which zinc supplementation and anti-TB therapy were started at baseline and ART was introduced 2 weeks later; none can be attributed confidently to a single intervention. 3.3 Multivariate Lipidomic Profiling PCA based on the 1,230 detected lipid species showed substantial overlap between baseline and follow-up samples (Fig. 1 ), indicating that between-subject variation exceeded the overall within-subject shift over time. Exploratory OPLS-DA yielded R²Y = 0.997 and Q² = 0.070 (Fig. 2 A). Permutation testing confirmed that the model’s explained variance was nominally significant (p(R²Y) = 0.045), but predictive performance was not (p(Q²) = 0.405; Fig. 2 B), consistent with overfitting in this high-dimensional, small-sample dataset. Accordingly, the supervised model was interpreted cautiously and used only as an ancillary visualization rather than as a basis for feature selection. 3.4 Prioritization of Differential Lipid Species After global Benjamini–Hochberg correction across 1,230 lipid species tested by paired Wilcoxon signed-rank tests, no feature remained significant (smallest q = 0.30 for DAG 20:1–20:5; all other q > 0.55). We therefore restricted inference to candidate prioritization based on within-subject robustness rather than confirmatory significance testing. In total, 81 lipid species showed nominal paired differences at the uncorrected level (p < 0.05). Of these 81, 51 had complete observed data in all 14 paired samples, whereas 30 had one or more imputed values. Among imputed features, apparent fold changes were often driven by transitions between near-threshold and detectable values and were therefore not considered reliable for biological interpretation. These imputation-sensitive signals are listed in Supplementary Table S1 for transparency. Forty complete-data lipid species met criteria 1–4 (nominal p < 0.05, r ≥ 0.5, non-imputed intensities in all 14 pairs, and bootstrap CI excluding 1.0). Of these, four were prioritized under criterion 5 for mechanistic interpretation because they showed upregulation consistent with the a priori sphingolipid-related hypothesis: Cer d18:1/26:0 (r = 0.73, median FC = 1.31, 95% bootstrap CI 1.16–1.67), SM 40:5 (r = 0.70, FC = 1.44, CI 1.05–1.64), SM 37:1 (r = 0.65, FC = 1.17, CI 1.02–1.36), and PG 16:1–22:1 (r = 0.76, FC = 1.35, CI 1.02–1.66) (Table 3 ). Three of the four prioritized candidates were sphingolipids. The remaining 36 complete-data lipids that met criteria 1–4 but were not prioritized under criterion 5 are reported in Supplementary Table S1 ; these include DAG 20:1–20:5 (r = 0.86, FC = 0.95, direction: decreased), SM 39:2 (r = 0.60, FC = 1.13, direction: increased), and additional PE, PS, and SM species. As a supplementary analysis, within-class Benjamini–Hochberg correction was performed for all lipids within each subclass using paired Wilcoxon p-values. DAG 20:1–20:5 was the only lipid to reach within-class significance (q = 0.011; m = 46 diacylglycerols), although it was not prioritized under criterion 5 because it showed downregulation inconsistent with the sphingolipid-related hypothesis. Cer d18:1/26:0 had a within-class q of 0.052 (m = 13 ceramides), approaching but not reaching the conventional 0.05 threshold. Within-class q-values for the prioritized sphingomyelins were 0.14 (m = 57 sphingomyelins), and for PG 16:1–22:1, 0.19 (m = 168 phosphatidylglycerols). 3.5 Descriptive Pathway Annotation KEGG annotation was available for only one of the four prioritized candidate lipids, Cer d18:1/26:0 (C00195). Because all pathway hits were driven by this single mapped compound, pathway output was considered descriptive only and is presented in Supplementary Fig. S2 . 3.6 Targeted Energy Metabolites Among the 42 metabolites that passed quality filters on the targeted panel, lactate was the only species reaching nominal significance by paired Wilcoxon testing (p = 0.009; median within-pair FC = 1.30, increased at follow-up). This signal did not remain significant after BH correction within the targeted panel. Glycerol-3-phosphate had the largest absolute fold-change estimate on the platform-level analysis, but this apparent change was driven by high missingness (only 7 of 14 pairs had both time points above the detection limit) and a single extreme baseline outlier; the paired Wilcoxon p-value for complete pairs was 0.58. We therefore regard the glycerol-3-phosphate signal as unreliable and the lactate finding as the only noteworthy exploratory observation on this panel. 3.7 Exploratory Clustering and Correlation Analyses Exploratory clustering of nominally changing lipids is shown in Supplementary Fig. S3 . Spearman correlation analysis among the 81 nominally changing species identified only six species with qualifying pairwise correlations (|r| > 0.8 and p < 0.05), an insufficient number to construct an informative network; these results are therefore reported narratively rather than as a figure. These analyses provided supplementary context but were not treated as independent evidence. 4. Discussion In this paired pilot study of 14 ART-naive adults with HIV/TB co-infection, no lipid species remained significant after global multiple-testing correction across 1,230 features (smallest q = 0.30). The dataset should therefore be interpreted as hypothesis-generating rather than confirmatory. Within that constraint, four lipids were mechanistically prioritized because they combined nominal paired differences with large effect sizes, observed non-imputed intensities in all paired samples, bootstrap confidence intervals for fold change that excluded 1.0, and directional consistency with the prespecified sphingolipid-focused framework. Three of these four candidates were sphingolipids—Cer d18:1/26:0, SM 40:5, and SM 37:1—while the fourth was the phosphatidylglycerol species PG 16:1–22:1. The follow-up sample reflects the cumulative effect of multiple interventions begun within the first month of treatment. Zinc supplementation and HRZE started at baseline, while ART was added after 2 weeks. The second blood sample was therefore taken on average 4 weeks after zinc and anti-TB treatment had started, but only 2 weeks after the first ART exposure. Lipid changes should be interpreted as an integrated early-treatment signal, not as the effect of any single intervention. The follow-up sample was also collected during the early post-ART window in which treatment-related immune reconstitution may also be modulating host metabolism. One interpretation is that improved zinc status coincided with altered sphingolipid homeostasis during this period, which is biologically plausible because secretory acid sphingomyelinase is zinc-dependent and links sphingomyelin turnover to ceramide generation [ 8 – 10 ]. However, the present data do not support a zinc-specific interpretation: the concurrent increase in ceramide and two sphingomyelin species argues against a simple explanation based solely on sphingomyelin hydrolysis and instead suggests that multiple processes were acting in parallel during this early period, including sphingolipid turnover, compensatory resynthesis, inflammatory resolution, anti-TB treatment effects, early ART-related remodeling, and general nutritional recovery. The increase in PG 16:1–22:1 is also of interest, but its biological interpretation is less certain. Phosphatidylglycerol is a minor plasma phospholipid with known roles in pulmonary surfactant biology and innate immune signaling [ 21 ]. In the context of pulmonary TB, changes in circulating PG species may be part of treatment-associated remodeling via the Kennedy pathway of de novo phospholipid synthesis and the Lands cycle of acyl-chain remodeling [ 22 , 23 ], but plasma measures can only provide an indirect view of pulmonary events. On the targeted energy panel, lactate was the sole metabolite reaching nominal significance, consistent with increased glycolytic activity that accompanies immune cell activation during early treatment. Whether this observation reflects direct metabolic effects of zinc repletion, a shift in immune cell substrate utilization during CD4 + T-cell recovery, or anti-TB drug effects cannot be determined from these data. The contrast between the two analytical platforms is itself informative: 81 of 1,230 lipid species reached nominal significance, whereas only 1 of 42 energy metabolites did so, despite the targeted panel carrying a lighter multiple-testing burden. One plausible explanation is that membrane phospholipid remodeling operates on a timescale of days to weeks as acyl chains are incorporated and redistributed through the Lands cycle, whereas central carbon intermediates turn over within hours [ 26 ] and may have already reached a new steady state by the 4-week sampling point. Alternatively, the early treatment period may genuinely perturb structural lipid pools more than glycolytic or tricarboxylic acid cycle flux in this clinical context. This largely negative energy metabolomics result should not be dismissed; it narrows the hypothesis space by suggesting that the metabolic signature of early treatment in zinc-supplemented HIV/TB patients is predominantly lipid-centric rather than broadly metabolic. An important methodological point from this pilot dataset is that many nominal lipid signals were dominated by low-abundance features that were highly sensitive to imputation. In these cases, apparent fold changes can be inflated when near-threshold values are contrasted with detectable values. We based biological interpretation on signal completeness and paired robustness rather than large apparent fold changes alone. This approach substantially narrowed the candidate list and avoided overinterpreting low-confidence features. Thirty-six additional complete-data lipids met criteria 1–4 but were not prioritized under criterion 5. Among these, DAG 20:1–20:5 had the largest effect size in the dataset (r = 0.86) and was the only lipid to reach within-class significance after Benjamini–Hochberg correction (q = 0.011, m = 46 diacylglycerols). Its consistent downregulation (13 of 14 participants showed decreased levels) suggests a genuine biological signal that warrants independent investigation, but because this diacylglycerol species falls outside the prespecified sphingolipid-focused framework, it was not prioritized and should instead be considered a parallel candidate for future targeted validation. Additional unprioritized complete-data signals spanned multiple phospholipid subclasses, including PE, PS, LPC, and additional SM and Cer species, consistent with broad phospholipid remodeling during early treatment. These are documented in Supplementary Table S1 and represent candidates for future targeted validation. Routine clinical measures changed over the same interval, including plasma zinc (p = 0.001), CD4 + T-cell counts, HDL-C, albumin, and prealbumin, which increased, and ferritin (p = 0.002) and uric acid (p = 0.013), which decreased. Small dense LDL also rose (p = 0.026), a finding that may reflect early ART-associated lipid remodeling, although the clinical significance of this change in the context of concurrent anti-TB therapy and nutritional recovery remains uncertain. A recent meta-analysis of 14 zinc supplementation trials in type 2 diabetes documented significant increases in HDL-C [ 24 ], though the study populations and clinical contexts are not directly comparable. These changes are consistent with early treatment response and nutritional recovery, but do not help to parse out the specific contribution of zinc supplementation. The same caveat applies to the lipid findings: this study has captured a clinically meaningful early-treatment window, but the metabolic signals observed during that interval are likely to reflect several overlapping processes. The paired design reduced between-subject heterogeneity and allowed early within-subject shifts to be assessed in a clinically understudied population. Few lipidomics studies have focused on ART-naive adults with HIV/TB co-infection at the time of treatment initiation. The combination of untargeted lipidomics with targeted metabolomics provided complementary analytical perspectives, although convergence across platforms was limited. This study has several limitations. As a small single-arm paired study without a comparator group, it is not possible to isolate the specific effects of zinc supplementation from those of the anti-TB therapy initiated at baseline, ART added 2 weeks later, nutritional recovery, or treatment-related immune reconstitution. No lipid remained significant after global BH correction, and the four mechanistically prioritized candidates should be regarded as preliminary signals rather than validated biomarkers. The supervised multivariate model lacked robust predictive performance, and pathway annotation was restricted to a single robust candidate. Secretory acid sphingomyelinase activity was not measured, so the connection between zinc status and sphingolipid metabolism remains hypothetical. The two-time-point design captures only early changes and does not address whether the observed signals persist, intensify, or resolve over longer follow-up. The follow-up sample was obtained during the early post-ART period, within the clinical window in which treatment-related immune reconstitution may be affecting host metabolism, and the single-center design limits generalizability. 5. Conclusion supplemented adults with HIV/TB co-infection was accompanied by several within-subject lipid shifts, but none remained significant after global multiple-testing correction. Four lipids—Cer d18:1/26:0, SM 40:5, SM 37:1, and PG 16:1-22:1—showed the most consistent directional evidence among hypothesis-guided candidates, supported by paired effect sizes, signal completeness, and bootstrap confidence intervals. Cer d18:1/26:0 approached within-class significance (q = 0.052 among 13 ceramides), while DAG 20:1-20:5 was the only lipid reaching within-class significance (q = 0.011) and may warrant parallel investigation. These data support focused validation rather than mechanistic inference. A future targeted study pre-specifying these four lipids, together with direct measurement of secretory acid sphingomyelinase activity and an appropriate comparator group, would provide a stronger test of the proposed biology. AI-Assisted Writing Disclosure AI-based language tools were used only for English-language editing during manuscript preparation. No automated software contributed to the study design, data collection, data analysis, or scientific interpretation; all such work was carried out by the authors. The authors retain full responsibility for the scientific content, analyses, and conclusions of this manuscript. Declarations Funding Supported by Shenzhen Science and Technology Program (No. JCYJ20220530163212029) and Shenzhen Clinical Research Center for Emerging Infectious Diseases (No. LCYSSQ20220823091203007). Conflicts of Interest The authors declare no conflicts of interest. Ethics Approval Approved by Shenzhen Third People's Hospital Institutional Ethics Committee (2023-045-02). Consent to Participate Written informed consent was obtained from all participants. Data Availability Available from the corresponding author on reasonable request. Author Contributions Conceptualization: Chao Chen, Fang Zhao. Methodology: Shiyun Chen, Chao Chen. Formal Analysis: Wei Li, Yingying Zhang. Investigation: Wei Li, Fang Zhao, Miaona Liu. Resources: Haitao Zhang, Miaona Liu. Data Curation: Haitao Zhang, Fang Zhao. Writing – Review & Editing: Wei Li, Yingying Zhang. Visualization: Man Rao, Chao Chen, Miaona Liu. Supervision: Fang Zhao, Wei Li. Project Administration: Fang Zhao. All authors have read and approved the final manuscript. References World Health Organization (2024) Global Tuberculosis Report 2024. World Health Organization, Geneva Müller M, Wandel S, Colebunders R et al (2010) Immune reconstitution inflammatory syndrome in patients starting antiretroviral therapy for HIV infection: a systematic review and meta-analysis. Lancet Infect Dis 10:251–261. https://doi.org/10.1016/S1473-3099(10)70026-8 Baum MK, Campa A, Lai S, Lai H, Page JB (2003) Zinc status in human immunodeficiency virus type 1 infection and illicit drug use. Clin Infect Dis 37(Suppl 2):S117–S123. https://doi.org/10.1086/375875 Read SA, Obeid S, Ahlenstiel C, Ahlenstiel G (2019) The role of zinc in antiviral immunity. Adv Nutr 10:696–710. https://doi.org/10.1093/advances/nmz013 Baum MK, Lai S, Sales S, Page JB, Campa A (2010) Randomized, controlled clinical trial of zinc supplementation to prevent immunological failure in HIV-infected adults. Clin Infect Dis 50:1653–1660. https://doi.org/10.1086/652864 Wagnew F, Alene KA, Eshetie S et al (2022) Effects of zinc and vitamin A supplementation on prognostic markers and treatment outcomes of adults with pulmonary tuberculosis: a systematic review and meta-analysis. BMJ Glob Health 7:e008625. https://doi.org/10.1136/bmjgh-2022-008625 Lawson L, Thacher TD, Yassin MA et al (2010) Randomized controlled trial of zinc and vitamin A as co-adjuvants for the treatment of pulmonary tuberculosis. Trop Med Int Health 15:1481–1490. https://doi.org/10.1111/j.1365-3156.2010.02638.x Liu B, Obeid LM, Hannun YA (1997) Sphingomyelinases in cell regulation. Semin Cell Dev Biol 8:311–322. https://doi.org/10.1006/scdb.1997.0153 Schuchman EH (2010) Acid sphingomyelinase, cell signalling, and disease. FEBS Lett 584:1895–1900. https://doi.org/10.1016/j.febslet.2009.11.083 Schissel SL, Keesler GA, Schuchman EH, Williams KJ, Tabas I (1998) The cellular trafficking and zinc dependence of secretory and lysosomal sphingomyelinase, two products of the acid sphingomyelinase gene. J Biol Chem 273:18250–18259. https://doi.org/10.1074/jbc.273.29.18250 Morales A, Lee H, Goñi FM, Kolesnick R, Fernandez-Checa JC (2007) Sphingolipids and cell death. Apoptosis 12:923–939. https://doi.org/10.1007/s10495-007-0721-0 Koethe JR, Lagathu C, Lake JE et al (2020) HIV and antiretroviral therapy-related fat alterations. Nat Rev Dis Primers 6:48. https://doi.org/10.1038/s41572-020-0181-1 Chaudhary NS, Kind T, Willig AL et al (2021) Changes in lipidomic profile by anti-retroviral treatment regimen: an ACTG 5257 ancillary study. Medicine 100:e26588. https://doi.org/10.1097/MD.0000000000026588 Matyash V, Liebisch G, Kurzchalia TV, Shevchenko A, Schwudke D (2008) Lipid extraction by methyl-tert-butyl ether for high-throughput lipidomics. J Lipid Res 49:1137–1146. https://doi.org/10.1194/jlr.D700041-JLR200 Wishart DS, Guo A, Oler E et al (2022) HMDB 5.0: the Human Metabolome Database for 2022. Nucleic Acids Res 50:D622–D631. https://doi.org/10.1093/nar/gkab1062 Fahy E, Sud M, Cotter D, Subramaniam S (2007) LIPID MAPS online tools for lipid research. Nucleic Acids Res 35:W606–W612. https://doi.org/10.1093/nar/gkm324 Thévenot EA, Roux A, Xu Y, Ezan E, Junot C (2015) Analysis of the human adult urinary metabolome variations with age, body mass index, and gender by implementing a comprehensive workflow for univariate and OPLS statistical analyses. J Proteome Res 14:3322–3335. https://doi.org/10.1021/acs.jproteome.5b00354 Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B 57:289–300. https://doi.org/10.1111/j.2517-6161.1995.tb02031.x Cohen J (1988) Statistical power analysis for the behavioral sciences, 2nd edn. Lawrence Erlbaum Associates, Hillsdale Kanehisa M, Goto S (2000) KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Res 28:27–30. https://doi.org/10.1093/nar/28.1.27 Han S, Mallampalli RK (2015) The role of surfactant in lung disease and host defense against pulmonary infections. Ann Am Thorac Soc 12:765–774. https://doi.org/10.1513/AnnalsATS.201411-507FR Kennedy EP, Weiss SB (1956) The function of cytidine coenzymes in the biosynthesis of phospholipides. J Biol Chem 222:193–214. https://doi.org/10.1016/S0021-9258(19)50785-2 Valentine WJ, Shimizu T, Shindou H (2023) Lysophospholipid acyltransferases orchestrate the compositional diversity of phospholipids. Biochimie 215:24–33. https://doi.org/10.1016/j.biochi.2023.08.012 Heidari Seyedmahalleh M, Montazer M, Ebrahimpour-Koujan S, Azadbakht L (2023) The effect of zinc supplementation on lipid profiles in patients with type 2 diabetes mellitus: a systematic review and dose-response meta-analysis of randomized clinical trials. Adv Nutr 14:1374–1388. https://doi.org/10.1016/j.advnut.2023.08.006 Dooley KE, Kaplan R, Mwelase N et al (2020) Dolutegravir-based antiretroviral therapy for patients coinfected with tuberculosis and human immunodeficiency virus: a multicenter, noncomparative, open-label, randomized trial. Clin Infect Dis 70:549–556. https://doi.org/10.1093/cid/ciz256 Jang C, Chen L, Rabinowitz JD (2018) Metabolomics and isotope tracing. Cell 173:822–837. https://doi.org/10.1016/j.cell.2018.03.055 Tables Table 1. Baseline characteristics of 14 participants Characteristic Value n Age, years 41.1 ± 13.0 14 Sex (M/F) 11/3 14 BMI, kg/m² 19.34 ± 4.09 14 Baseline CD4+ T-cells, cells/μL 95.8 ± 75.7 (21–237) 14 Baseline plasma zinc, μmol/L 11.32 ± 1.39 14 Sampling interval, days 29.6 ± 1.9 (28–34) 14 Data are mean ± SD (range where indicated). Table 2. Clinical and biochemical parameters at baseline and follow-up Parameter Baseline Follow-up Change (%) p Sig. Plasma zinc, μmol/L 11.32 ± 1.39 13.06 ± 1.78 +15.3 <0.001 *** CD4+ T-cells, cells/μL 95.8 ± 75.7 191.6 ± 96.0 +100.0 <0.001 *** IL-6, pg/mL 30.66 ± 33.09 8.52 ± 9.66 −72.2 0.058 Total cholesterol, mmol/L 3.49 ± 0.92 4.27 ± 1.19 +22.2 0.064 Triglycerides, mmol/L 1.33 ± 0.92 1.41 ± 0.76 +6.2 0.808 HDL-C, mmol/L 0.89 ± 0.39 1.35 ± 0.51 +51.6 0.011 * LDL-C, mmol/L 2.04 ± 0.67 2.28 ± 0.95 +11.9 0.397 ApoA1, g/L 0.84 ± 0.38 1.05 ± 0.49 +23.8 0.198 ApoB, g/L 0.87 ± 0.15 1.19 ± 0.40 +36.3 0.011 * Lp(a), mg/dL 23.7 ± 31.9 11.7 ± 27.9 −50.6 0.028 * sdLDL, mmol/L 0.49 ± 0.23 0.72 ± 0.27 +46.9 0.026 * Albumin, g/L 36.0 ± 4.8 39.3 ± 4.5 +9.2 0.028 * Prealbumin, mg/L 158.9 ± 71.7 213.7 ± 45.3 +34.4 0.035 * ALT, U/L 24.4 ± 18.0 26.6 ± 18.0 +8.7 0.401 AST, U/L 35.0 ± 27.1 27.7 ± 8.6 −20.9 0.777 GGT, U/L 95.5 ± 147.9 82.6 ± 60.3 −13.5 0.463 Fasting glucose, mmol/L 4.77 ± 0.36 5.25 ± 0.66 +10.1 0.035 * Ferritin, ng/mL 812.5 ± 514.0 401.5 ± 183.2 −50.6 0.002 ** Uric acid, μmol/L 374.1 ± 110.8 543.2 ± 206.1 +45.2 0.013 * eGFR, mL/min/1.73m² 110.4 ± 24.8 106.6 ± 19.5 −3.4 0.241 Data are mean ± SD. P values from two-sided paired Wilcoxon signed-rank tests. Change (%) = ((follow-up − baseline)/baseline) × 100. *P < 0.05; **P < 0.01; ***P < 0.001. HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; ApoA1, apolipoprotein A1; ApoB, apolipoprotein B; Lp(a), lipoprotein(a); sdLDL, small dense LDL; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase; eGFR, estimated glomerular filtration rate. Table 3. Mechanistically prioritized candidate lipid species Lipid species Class p r FC 95% CI Dir. q (global) Cer d18:1/26:0 Cer 0.004 0.73 1.31 1.16–1.67 Up 0.55 SM 40:5 SM 0.007 0.70 1.44 1.05–1.64 Up 0.64 SM 37:1 SM 0.013 0.65 1.17 1.02–1.36 Up 0.69 PG 16:1-22:1 PG 0.002 0.76 1.35 1.02–1.66 Up 0.55 p, exact paired Wilcoxon signed-rank test. r, within-subject effect size (|Z|/√N). FC, median of within-pair fold changes (follow-up/baseline). CI, bootstrap 95% confidence interval (10,000 resamples). Dir., direction of change. q (global), Benjamini–Hochberg adjusted p-value across 1,230 features. All four candidates had observed non-imputed intensities in all 14 paired samples. Table 3 Prioritized candidate lipid species based on paired effect size, signal completeness, and bootstrap confidence intervals. Additional Declarations No competing interests reported. Supplementary Files Supplementarytable.xlsx SupplementaryMaterials.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 29 Apr, 2026 Reviewers agreed at journal 20 Apr, 2026 Reviewers agreed at journal 19 Apr, 2026 Reviewers invited by journal 14 Apr, 2026 Editor assigned by journal 11 Apr, 2026 Submission checks completed at journal 10 Apr, 2026 First submitted to journal 03 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9309714","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":626412803,"identity":"7e02e69a-cc8e-4e0f-9a66-d88a6a4075ef","order_by":0,"name":"Chao Chen","email":"","orcid":"","institution":"Shenzhen Third People’s Hospital","correspondingAuthor":false,"prefix":"","firstName":"Chao","middleName":"","lastName":"Chen","suffix":""},{"id":626412804,"identity":"e24596cf-1729-4b4a-8c75-6495eb931a42","order_by":1,"name":"ShiYun Chen","email":"","orcid":"","institution":"Shenzhen Third People’s Hospital","correspondingAuthor":false,"prefix":"","firstName":"ShiYun","middleName":"","lastName":"Chen","suffix":""},{"id":626412805,"identity":"d37dc91d-3e15-496f-bbcc-fb7fd2ff1be9","order_by":2,"name":"YingYing Zhang","email":"","orcid":"","institution":"Shenzhen University","correspondingAuthor":false,"prefix":"","firstName":"YingYing","middleName":"","lastName":"Zhang","suffix":""},{"id":626412806,"identity":"4145c514-4ab5-489c-8c21-b9faab15decc","order_by":3,"name":"Man Rao","email":"","orcid":"","institution":"Shenzhen Third People’s Hospital","correspondingAuthor":false,"prefix":"","firstName":"Man","middleName":"","lastName":"Rao","suffix":""},{"id":626412807,"identity":"8686f7cf-8b93-48fe-b7b6-ffbb072903d5","order_by":4,"name":"HaiTao Zhang","email":"","orcid":"","institution":"Shenzhen Third People’s Hospital","correspondingAuthor":false,"prefix":"","firstName":"HaiTao","middleName":"","lastName":"Zhang","suffix":""},{"id":626412808,"identity":"6e50b1a7-815a-4c32-827f-ae6e195c396d","order_by":5,"name":"MiaoNa Liu","email":"","orcid":"","institution":"Shenzhen Third People’s Hospital","correspondingAuthor":false,"prefix":"","firstName":"MiaoNa","middleName":"","lastName":"Liu","suffix":""},{"id":626412809,"identity":"333ad794-a925-471c-957d-3299d68a3860","order_by":6,"name":"Wei Li","email":"","orcid":"","institution":"Shenzhen Third People’s Hospital","correspondingAuthor":false,"prefix":"","firstName":"Wei","middleName":"","lastName":"Li","suffix":""},{"id":626412810,"identity":"434e17fe-012b-4fa4-84f9-bef405160138","order_by":7,"name":"Fang Zhao","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5ElEQVRIiWNgGAWjYDACCSCuADGYGRsffDCwsSNOyxkQg735sOGMgrRkErTwHEuT5vlwiLGBkA752c0PHxyouWPXIJFjbGxjcICZgf3w0Q34tDDOOWZscODYs2SgFsPHOQZ3+Bh40tJu4NPCLJFgJv2B7XAyA8iWHINnzAwSPGZ4tbBJpH+TOPAPrMVM2sLgMGMDIS08QJUSB9sO24G9z0CMFgmJnGKDg32HE8CB3GOQlsxGyC/yM9I3Pjjw7bA9OCp//LGx42c/fAyvFhhI3H8A5jtilIOAPbEKR8EoGAWjYAQCAJIgTFr6MO6xAAAAAElFTkSuQmCC","orcid":"","institution":"Shenzhen Third People’s Hospital","correspondingAuthor":true,"prefix":"","firstName":"Fang","middleName":"","lastName":"Zhao","suffix":""}],"badges":[],"createdAt":"2026-04-03 06:55:39","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9309714/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9309714/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107569018,"identity":"684254d4-f673-4276-9fa5-56f529893958","added_by":"auto","created_at":"2026-04-22 17:41:34","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":95956,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePCA score plot of 1,230 lipid species in 14 paired plasma samples.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePrincipal component analysis (PCA) based on 1,230 lipid species detected by untargeted UHPLC-MS/MS lipidomics. Each point represents one plasma sample; circles indicate baseline samples and triangles indicate follow-up samples collected approximately 4 weeks later. Ellipses represent 95 % confidence regions for each group. Thin connecting lines link paired samples from the same participant. Substantial overlap between groups indicates that between-subject variation exceeded the overall within-subject shift over time. Color scheme: dark blue circles (baseline), red-orange triangles (follow-up).\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-9309714/v1/bcd015571186abbc60a5657d.png"},{"id":107706231,"identity":"bc74ffb6-c45f-4277-9c03-0a94022ab5f3","added_by":"auto","created_at":"2026-04-24 09:17:43","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":132628,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExploratory OPLS-DA analysis.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Score plot from orthogonal partial least-squares discriminant analysis (OPLS-DA; ropls 1.34.0, Bioconductor; log10 transformation, mean centering, 1,230 features). The horizontal axis represents the predictive component (t[1]) and the vertical axis represents the orthogonal component (to[1]). Clear separation along the predictive axis reflects the supervised classification, but the model yielded R²Y = 0.997 and Q² = 0.070, indicating high explained variance with limited cross-validated predictive performance. (B) Permutation testing (200 permutations; ropls 1.34.0, Bioconductor; log10 transformation, mean centering; seed = 42). R²Y intercept = 0.99, Q² intercept = 0.03; permutation p(R²Y) = 0.045, p(Q²) = 0.405. The model’s explained variance was nominally significant but predictive performance was not, consistent with overfitting in this high-dimensional, small-sample dataset; accordingly, VIP scores from this model were not used as mandatory feature-selection criteria.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-9309714/v1/7d75c6459ca6f5428df89980.png"},{"id":107706408,"identity":"51a8eeb5-9026-48af-bfa8-54385804a1d0","added_by":"auto","created_at":"2026-04-24 09:18:04","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":97898,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eVolcano plot of 1,230 lipid species.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEach point represents one lipid feature. The horizontal axis shows log₂(fold change) and the vertical axis shows −log₁₀(p value) from paired Wilcoxon signed-rank tests. The horizontal dashed line indicates the nominal threshold of p = 0.05. Cyan points: nominally significant lipids (p \u0026lt; 0.05); light gray: not significant. Stars highlight the four robust candidate lipids that had observed non-imputed intensities in all 14 paired samples, large within-subject effect sizes (r ≥ 0.5), and bootstrap 95 % confidence intervals for fold change excluding 1.0: Cer d18:1/26:0, SM 40:5, SM 37:1, and PG 16:1-22:1 (all increased at follow-up). No lipid species remained significant after global Benjamini–Hochberg correction (all q \u0026gt; 0.30). The volcano plot includes all 1,230 tested features; features with high missingness (\u0026gt;20%) that reached nominal significance are listed in Supplementary Table S1 but were not used for candidate prioritization.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-9309714/v1/5d2ffc1db17d233d404d7110.png"},{"id":107709102,"identity":"d48b5d0d-e616-4699-84ea-5392f3a5ed67","added_by":"auto","created_at":"2026-04-24 09:34:49","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":619130,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9309714/v1/28e34c52-615b-4af8-83e9-f65ece50fac6.pdf"},{"id":107569019,"identity":"615675e8-d752-4f10-acc0-d39c07bee591","added_by":"auto","created_at":"2026-04-22 17:41:34","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":113952,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarytable.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9309714/v1/34d351cf56fd220879260638.xlsx"},{"id":107569022,"identity":"506afb95-8c81-4713-b048-b2d5cfc5a7bc","added_by":"auto","created_at":"2026-04-22 17:41:35","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":303117,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterials.docx","url":"https://assets-eu.researchsquare.com/files/rs-9309714/v1/59f30ace55a6e1266d5db8d4.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Plasma sphingolipid and glycerophospholipid shifts during early treatment in zinc-supplemented adults with HIV–tuberculosis co-infection: a paired exploratory lipidomics study","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eTuberculosis (TB) is the world\u0026rsquo;s deadliest infectious disease with 1.25\u0026nbsp;million deaths and 10.8\u0026nbsp;million new cases in 2023 [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. A substantial proportion of these new TB cases occur among people living with HIV, with the latter being at highest risk. Paradoxical clinical worsening and metabolic instability can occur during the first weeks of treatment, when CD4\u0026thinsp;+\u0026thinsp;T-cell recovery from antiretroviral therapy (ART) begins [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. How early immune reconstitution impacts lipid, glucose, and protein metabolism remains poorly characterized, particularly when nutritional supplementation is also provided.\u003c/p\u003e \u003cp\u003eZinc deficiency is common among people with HIV [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e], particularly in resource-constrained settings where TB is most prevalent. Low zinc status has been linked with adverse clinical outcomes, including more rapid disease progression and more opportunistic infections. Prior work has also suggested that long-term zinc supplementation may decrease the risk of immunological failure [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. In the setting of TB, zinc co-supplementation during anti-TB treatment has been associated with modest improvements in sputum conversion [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. The impact of zinc supplementation on the host lipidome during early immune reconstitution in HIV/TB co-infection has not yet been studied.\u003c/p\u003e \u003cp\u003eOne mechanistic connection between zinc and lipid metabolism is the zinc-dependent secretory acid sphingomyelinase (S-aSMase) pathway, which cleaves sphingomyelin to ceramide and phosphocholine. Activity of this pathway is dependent on the exogenous addition of zinc [\u003cspan additionalcitationids=\"CR9\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Ceramide is a bioactive second messenger involved in many aspects of cellular function, including programmed cell death and inflammation [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Consequently, zinc availability could plausibly influence sphingomyelin/ceramide balance and thereby perturb phospholipid composition. To our knowledge, evidence on whether oral zinc supplementation is associated with changes in plasma sphingolipid\u0026ndash;glycerophospholipid profiles during early treatment in HIV/TB co-infection is lacking.\u003c/p\u003e \u003cp\u003eLipid metabolism is already perturbed in HIV infection, and the initiation of ART further reconfigures the plasma lipidome. ART initiation, including integrase inhibitor-based regimens, has been associated with changes in body composition and lipid metabolism in treatment-naive patients [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. At the molecular level, Chaudhary et al. reported that different ART regimens induced divergent responses in ceramide, sphingomyelin, and phosphatidylethanolamine species [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. These studies suggest that the plasma lipidome is responsive to ART in HIV. In contrast, the lipidomic effects of nutritional supplementation, including zinc, have been largely overlooked, particularly during the first weeks of treatment when immune reconstitution, anti-TB drug effects, and nutritional repletion all converge.\u003c/p\u003e \u003cp\u003eAccordingly, paired plasma samples from 14 ART-naive adults with HIV/TB co-infection who received zinc supplementation during the first month of treatment were analyzed in this study. In this cohort, zinc supplementation and standard anti-TB therapy were initiated at baseline, whereas ART was started 2 weeks later. The primary objective was to identify within-subject lipid shifts that were sufficiently consistent and analytically robust to warrant targeted future investigation. Given the small exploratory sample size and pilot design of the present study, candidate lipids were prioritized using a combined assessment of effect sizes, confidence intervals, imputation resistance, and transparent reporting of global false-discovery rates.\u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Study Design and Ethical Approval\u003c/h2\u003e \u003cp\u003eThis prospective paired single-arm study was conducted at Shenzhen Third People\u0026rsquo;s Hospital between 2023 and 2024 as a substudy nested within a larger randomized controlled trial (NCT05847715) of zinc supplementation in HIV/TB co-infected adults. The protocol was approved by the Institutional Ethics Committee of Shenzhen Third People\u0026rsquo;s Hospital (approval number: 2023-045-02). Written informed consent was obtained from all participants. The study was conducted in accordance with the Declaration of Helsinki.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Study Population\u003c/h2\u003e \u003cp\u003eAdults between 18 and 70 years of age with confirmed HIV infection and newly diagnosed active pulmonary TB were eligible. All participants were ART-naive at recruitment. Standard anti-TB therapy was initiated at enrollment, whereas ART was started at the scheduled 2-week follow-up visit. Inclusion criteria were: (1) confirmed HIV-1 infection by Western blot or nucleic acid testing; (2) active pulmonary TB diagnosed by sputum smear, culture, or GeneXpert MTB/RIF; (3) no prior ART exposure; and (4) willingness to adhere to the supplementation protocol. Patients with hepatic insufficiency (ALT\u0026thinsp;\u0026gt;\u0026thinsp;5\u0026times; upper limit of normal), renal failure (eGFR\u0026thinsp;\u0026lt;\u0026thinsp;30 mL/min/1.73 m\u0026sup2;), concurrent malignancy, pregnancy, or recent use of zinc-containing preparations were excluded.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Intervention and Sample Collection\u003c/h2\u003e \u003cp\u003e Participants began oral zinc gluconate (40 mg elemental zinc per day) and standard first-line anti-TB therapy (isoniazid, rifampin, ethambutol, and pyrazinamide; HRZE) on the day of enrollment, after baseline blood sampling. An intermediate follow-up visit was conducted approximately 2 weeks later, at which antiretroviral therapy (ART) was initiated. The ART regimen consisted of lamivudine, tenofovir disoproxil fumarate, and dolutegravir, with dolutegravir dosed twice daily because of concomitant rifampin use [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. No blood sample from the 2-week visit was included in the present paired lipidomics analysis. The second fasting blood sample for the present study was collected approximately 4 weeks after enrollment. Thus, the follow-up sample reflected approximately 4 weeks of zinc supplementation and anti-TB therapy, as well as approximately 2 weeks of ART exposure (Supplementary Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). Plasma was separated within 2 hours, aliquoted, and stored at \u0026minus;\u0026thinsp;80\u0026deg;C. No untreated or non-zinc comparator arm was included in the present lipidomics substudy. Zinc supplementation adherence was assessed by pill counts at each visit; all 14 participants maintained full adherence.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Clinical and Biochemical Assessments\u003c/h2\u003e \u003cp\u003eFasting blood was collected for routine clinical parameters at each time point: complete blood count, hepatic function (ALT, AST, GGT, albumin, prealbumin), renal function (uric acid, eGFR), fasting glucose, ferritin, and a lipid panel (total cholesterol, triglycerides, HDL-C, LDL-C, apolipoprotein A1, apolipoprotein B, lipoprotein(a), small dense LDL). Plasma zinc was measured by inductively coupled plasma mass spectrometry. CD4\u0026thinsp;+\u0026thinsp;T-cell counts were determined by flow cytometry, and interleukin-6 (IL-6) by chemiluminescent immunoassay.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Untargeted Lipidomics Analysis\u003c/h2\u003e \u003cp\u003eLipids were extracted from 100 \u0026micro;L plasma using methyl tert-butyl ether/methanol (Matyash protocol) [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Chromatographic separation used a Waters ACQUITY UPLC CSH C18 column (2.1 \u0026times; 100 mm, 1.7 \u0026micro;m) at 55\u0026deg;C with acetonitrile/water (60:40) and isopropanol/acetonitrile (90:10) mobile phases, both containing 10 mM ammonium formate and 0.1% formic acid. Data were acquired in positive and negative ESI modes on a Thermo Scientific Q Exactive Plus mass spectrometer. QC samples were prepared by pooling equal aliquots from all study samples and injected regularly throughout the run. Pearson correlation coefficients among QC replicates were \u0026gt;\u0026thinsp;0.99. Lipid species were identified by matching accurate mass and MS/MS spectra against HMDB [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e], LIPID MAPS [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e], and an in-house spectral library. The extraction of metabolite ion peaks and the identification of metabolites are carried out using the quantitative software Skyline [v.21.1.0.146]. This will result in a data matrix containing information such as the identification results and quantitative results of the metabolites. Subsequently, further information analysis and processing will be conducted on this table. for peak detection, alignment, and feature extraction. A total of 1,230 species spanning 10 subclasses passed basic quality control.\u003c/p\u003e \u003cp\u003eData preprocessing: Peak intensities were normalized to internal standards. Among the 1,230 detected species, 696 features had missingness\u0026thinsp;\u0026le;\u0026thinsp;20% across the 28 observations; the remaining 534 features had missingness\u0026thinsp;\u0026gt;\u0026thinsp;20%. Missing values were imputed using k-nearest neighbors for features judged to be missing at random and by the minimum observed value after normalization for values judged to be below the detection limit. For exploratory hypothesis screening, paired Wilcoxon tests were applied to all 1,230 features to avoid excluding potentially informative low-abundance species at the testing stage. However, candidate prioritization and biological interpretation were restricted to features with observed non-imputed intensities in all 14 paired samples (i.e., complete-data features), as described in Section \u003cspan refid=\"Sec9\" class=\"InternalRef\"\u003e2.7\u003c/span\u003e. Features with missingness\u0026thinsp;\u0026gt;\u0026thinsp;20% were retained in unsupervised (PCA) and exploratory supervised (OPLS-DA) multivariate overviews, where they contribute to the overall variance structure. No batch correction was applied because all samples were analyzed in a single analytical batch. Because low-abundance features can generate unstable fold-change estimates after imputation, we prespecified a sensitivity analysis to distinguish robust observed signals from imputation-sensitive signals in the downstream interpretation.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Targeted Energy Metabolomics\u003c/h2\u003e \u003cp\u003eEighty MRM transitions covering central carbon and energy metabolites were monitored using a Waters ACQUITY H-ClassD UPLC coupled to an AB Sciex QTRAP 6500\u0026thinsp;+\u0026thinsp;mass spectrometer in scheduled multiple reaction monitoring mode. Separation used a BEH Amide column (2.1 \u0026times; 100 mm, 1.7 \u0026micro;m) at 60\u0026deg;C with gradient elution in both positive and negative ESI modes. Peak integration and quantification were performed using MultiQuant software (version 4.0, SCIEX). Fifty-one metabolites were quantifiable above the lower limit of quantification, and 42 of these passed quality filters for paired statistical testing (\u0026ge;\u0026thinsp;5 complete pairs with both time points above the detection limit).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.7 Statistical Analysis\u003c/h2\u003e \u003cp\u003eAll statistical analyses were performed in R (version 4.3.3; R Foundation for Statistical Computing, Vienna, Austria). Multivariate analyses were conducted using the ropls package (version 1.34.0, Bioconductor 3.18) [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Imputation of missing values (described in Section \u003cspan refid=\"Sec7\" class=\"InternalRef\"\u003e2.5\u003c/span\u003e) was performed using the impute package (version 1.76.0, Bioconductor; k\u0026thinsp;=\u0026thinsp;10). Figures were generated using the matplotlib library (version 3.8, Python 3.11) from precomputed R outputs. Given the exploratory pilot design and the small paired sample size (n\u0026thinsp;=\u0026thinsp;14), the study was not powered to confirm findings across 1,230 lipid features. Instead, we adopted a robustness-based candidate prioritization framework rather than a fold-change-threshold screening framework.\u003c/p\u003e \u003cp\u003eFor each of the 1,230 lipid species, paired differences between baseline and follow-up were assessed using the exact Wilcoxon signed-rank test (wilcox.test with exact\u0026thinsp;=\u0026thinsp;TRUE, stats package, base R). All reported p-values in the final manuscript were derived from the same statistical implementation, and all figures and tables were generated from a single precomputed master results table to ensure internal consistency. This broad exploratory screen included high-missingness features whose fold-change estimates may be unstable; downstream candidate prioritization was restricted to complete-data features as described below. Global Benjamini\u0026ndash;Hochberg (BH) adjusted q-values were calculated and are reported transparently, but were not used as the sole basis for candidate nomination in this pilot dataset [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. To describe signal strength beyond p-values, we calculated the Wilcoxon effect size (r = |Z|/\u0026radic;N), with r\u0026thinsp;\u0026ge;\u0026thinsp;0.5 considered large [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Fold change was summarized as the median of within-pair fold changes (follow-up/baseline for each participant) and accompanied by bootstrap 95% confidence intervals (10,000 resamples with replacement, seed\u0026thinsp;=\u0026thinsp;42; boot package, version 1.3\u0026ndash;28.1).\u003c/p\u003e \u003cp\u003ePrimary candidate lipids were prioritized using five criteria considered jointly: (1) nominal paired p\u0026thinsp;\u0026lt;\u0026thinsp;0.05; (2) large within-subject effect size (r\u0026thinsp;\u0026ge;\u0026thinsp;0.5); (3) observed non-imputed intensities across all 14 paired samples; (4) bootstrap 95% confidence intervals for fold change that excluded 1.0; and (5) directional consistency with the a priori sphingolipid-related biological hypothesis (upregulated species). Complete data lipids that met criteria 1\u0026ndash;4 but were not prioritized under criterion 5 are reported transparently in Supplementary Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e. Lipids with repeated values at or below the imputation threshold (\u0026lt;\u0026thinsp;1 \u0026times; 10⁻⁴) were classified as imputation-sensitive exploratory signals and were not used for mechanistic interpretation.\u003c/p\u003e \u003cp\u003ePCA was used for an unsupervised overview. For exploratory supervised analysis, OPLS-DA was performed on log10-transformed intensities with mean centering (ropls 1.34.0; 1 predictive and 1 orthogonal component). Model validity was assessed using a 200-fold permutation test (seed\u0026thinsp;=\u0026thinsp;42). Because cross-validated predictive performance was limited, VIP scores were not used as mandatory criteria for feature selection. Class-level BH correction within lipid subclasses was performed as a supplementary analysis. KEGG mapping and pathway annotation were used descriptively when available [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Exploratory Spearman correlations among nominally changing lipids were calculated; however, only six species met the threshold (|r| \u0026gt; 0.8 and p\u0026thinsp;\u0026lt;\u0026thinsp;0.05), yielding insufficient nodes for a meaningful network visualization, and this analysis was therefore not presented as a standalone figure. Routine clinical parameters were compared using paired Wilcoxon signed-rank tests (two-sided p\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\u003c/p\u003e \u003cp\u003eFor the targeted metabolomics panel, paired Wilcoxon tests and BH-adjusted q-values were calculated separately from the lipidomics dataset. Only metabolites with at least five complete paired observations above the detection limit were included.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Participant Characteristics\u003c/h2\u003e \u003cp\u003eFourteen HIV-1-infected adults (11 male and 3 female; mean age 41.1\u0026thinsp;\u0026plusmn;\u0026thinsp;13.0 years; mean BMI 19.34\u0026thinsp;\u0026plusmn;\u0026thinsp;4.09 kg/m2) completed the paired sampling protocol (all ART-naive at enrollment). The mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD time between baseline and follow-up sampling was 29.6\u0026thinsp;\u0026plusmn;\u0026thinsp;1.9 days (range 28\u0026ndash;34; Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Zinc supplementation and standard four-drug anti-TB therapy were begun after the baseline sampling, and ART was initiated at the scheduled 2-week follow-up visit. Plasma zinc increased from 11.32\u0026thinsp;\u0026plusmn;\u0026thinsp;1.39 \u0026micro;mol/L to 13.06\u0026thinsp;\u0026plusmn;\u0026thinsp;1.78 \u0026micro;mol/L (p\u0026thinsp;=\u0026thinsp;0.001). CD4\u0026thinsp;+\u0026thinsp;T-cell counts rose from 95.8\u0026thinsp;\u0026plusmn;\u0026thinsp;75.7 to 191.6\u0026thinsp;\u0026plusmn;\u0026thinsp;96.0 cells/\u0026micro;L (p\u0026thinsp;=\u0026thinsp;0.0004), consistent with early treatment-related immune recovery. IL-6 declined by 72.2% (30.66\u0026thinsp;\u0026plusmn;\u0026thinsp;33.09 to 8.52\u0026thinsp;\u0026plusmn;\u0026thinsp;9.66 pg/mL; p\u0026thinsp;=\u0026thinsp;0.058).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Clinical Metabolic Parameters\u003c/h2\u003e \u003cp\u003eSeveral other routinely obtained clinical parameters changed in the expected directions during the study period (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). HDL-C increased by 51.6% (0.89\u0026thinsp;\u0026plusmn;\u0026thinsp;0.39 to 1.35\u0026thinsp;\u0026plusmn;\u0026thinsp;0.51 mmol/L; p\u0026thinsp;=\u0026thinsp;0.011), apolipoprotein B rose by 36.3% (p\u0026thinsp;=\u0026thinsp;0.011), and lipoprotein(a) decreased by 50.6% (p\u0026thinsp;=\u0026thinsp;0.028). Small dense LDL increased by 46.9% (0.49\u0026thinsp;\u0026plusmn;\u0026thinsp;0.23 to 0.72\u0026thinsp;\u0026plusmn;\u0026thinsp;0.27 mmol/L; p\u0026thinsp;=\u0026thinsp;0.026). Albumin increased by 9.2% (p\u0026thinsp;=\u0026thinsp;0.028) and prealbumin by 34.4% (p\u0026thinsp;=\u0026thinsp;0.035). Fasting glucose rose modestly (p\u0026thinsp;=\u0026thinsp;0.035). Ferritin declined from 812.5\u0026thinsp;\u0026plusmn;\u0026thinsp;514.0 to 401.5\u0026thinsp;\u0026plusmn;\u0026thinsp;183.2 ng/mL (p\u0026thinsp;=\u0026thinsp;0.002), and uric acid increased from 374.1\u0026thinsp;\u0026plusmn;\u0026thinsp;110.8 to 543.2\u0026thinsp;\u0026plusmn;\u0026thinsp;206.1 \u0026micro;mol/L (p\u0026thinsp;=\u0026thinsp;0.013). Total cholesterol, triglycerides, LDL-C, apolipoprotein A1, and hepatic transaminases did not change significantly. All changes occurred during the first month of treatment, during which zinc supplementation and anti-TB therapy were started at baseline and ART was introduced 2 weeks later; none can be attributed confidently to a single intervention.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Multivariate Lipidomic Profiling\u003c/h2\u003e \u003cp\u003ePCA based on the 1,230 detected lipid species showed substantial overlap between baseline and follow-up samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), indicating that between-subject variation exceeded the overall within-subject shift over time. Exploratory OPLS-DA yielded R\u0026sup2;Y\u0026thinsp;=\u0026thinsp;0.997 and Q\u0026sup2; = 0.070 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Permutation testing confirmed that the model\u0026rsquo;s explained variance was nominally significant (p(R\u0026sup2;Y)\u0026thinsp;=\u0026thinsp;0.045), but predictive performance was not (p(Q\u0026sup2;)\u0026thinsp;=\u0026thinsp;0.405; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB), consistent with overfitting in this high-dimensional, small-sample dataset. Accordingly, the supervised model was interpreted cautiously and used only as an ancillary visualization rather than as a basis for feature selection.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Prioritization of Differential Lipid Species\u003c/h2\u003e \u003cp\u003eAfter global Benjamini\u0026ndash;Hochberg correction across 1,230 lipid species tested by paired Wilcoxon signed-rank tests, no feature remained significant (smallest q\u0026thinsp;=\u0026thinsp;0.30 for DAG 20:1\u0026ndash;20:5; all other q\u0026thinsp;\u0026gt;\u0026thinsp;0.55). We therefore restricted inference to candidate prioritization based on within-subject robustness rather than confirmatory significance testing.\u003c/p\u003e \u003cp\u003eIn total, 81 lipid species showed nominal paired differences at the uncorrected level (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Of these 81, 51 had complete observed data in all 14 paired samples, whereas 30 had one or more imputed values. Among imputed features, apparent fold changes were often driven by transitions between near-threshold and detectable values and were therefore not considered reliable for biological interpretation. These imputation-sensitive signals are listed in Supplementary Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e for transparency.\u003c/p\u003e \u003cp\u003eForty complete-data lipid species met criteria 1\u0026ndash;4 (nominal p\u0026thinsp;\u0026lt;\u0026thinsp;0.05, r\u0026thinsp;\u0026ge;\u0026thinsp;0.5, non-imputed intensities in all 14 pairs, and bootstrap CI excluding 1.0). Of these, four were prioritized under criterion 5 for mechanistic interpretation because they showed upregulation consistent with the a priori sphingolipid-related hypothesis: Cer d18:1/26:0 (r\u0026thinsp;=\u0026thinsp;0.73, median FC\u0026thinsp;=\u0026thinsp;1.31, 95% bootstrap CI 1.16\u0026ndash;1.67), SM 40:5 (r\u0026thinsp;=\u0026thinsp;0.70, FC\u0026thinsp;=\u0026thinsp;1.44, CI 1.05\u0026ndash;1.64), SM 37:1 (r\u0026thinsp;=\u0026thinsp;0.65, FC\u0026thinsp;=\u0026thinsp;1.17, CI 1.02\u0026ndash;1.36), and PG 16:1\u0026ndash;22:1 (r\u0026thinsp;=\u0026thinsp;0.76, FC\u0026thinsp;=\u0026thinsp;1.35, CI 1.02\u0026ndash;1.66) (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Three of the four prioritized candidates were sphingolipids. The remaining 36 complete-data lipids that met criteria 1\u0026ndash;4 but were not prioritized under criterion 5 are reported in Supplementary Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e; these include DAG 20:1\u0026ndash;20:5 (r\u0026thinsp;=\u0026thinsp;0.86, FC\u0026thinsp;=\u0026thinsp;0.95, direction: decreased), SM 39:2 (r\u0026thinsp;=\u0026thinsp;0.60, FC\u0026thinsp;=\u0026thinsp;1.13, direction: increased), and additional PE, PS, and SM species.\u003c/p\u003e \u003cp\u003eAs a supplementary analysis, within-class Benjamini\u0026ndash;Hochberg correction was performed for all lipids within each subclass using paired Wilcoxon p-values. DAG 20:1\u0026ndash;20:5 was the only lipid to reach within-class significance (q\u0026thinsp;=\u0026thinsp;0.011; m\u0026thinsp;=\u0026thinsp;46 diacylglycerols), although it was not prioritized under criterion 5 because it showed downregulation inconsistent with the sphingolipid-related hypothesis. Cer d18:1/26:0 had a within-class q of 0.052 (m\u0026thinsp;=\u0026thinsp;13 ceramides), approaching but not reaching the conventional 0.05 threshold. Within-class q-values for the prioritized sphingomyelins were 0.14 (m\u0026thinsp;=\u0026thinsp;57 sphingomyelins), and for PG 16:1\u0026ndash;22:1, 0.19 (m\u0026thinsp;=\u0026thinsp;168 phosphatidylglycerols).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Descriptive Pathway Annotation\u003c/h2\u003e \u003cp\u003eKEGG annotation was available for only one of the four prioritized candidate lipids, Cer d18:1/26:0 (C00195). Because all pathway hits were driven by this single mapped compound, pathway output was considered descriptive only and is presented in Supplementary Fig. \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e3.6 Targeted Energy Metabolites\u003c/h2\u003e \u003cp\u003eAmong the 42 metabolites that passed quality filters on the targeted panel, lactate was the only species reaching nominal significance by paired Wilcoxon testing (p\u0026thinsp;=\u0026thinsp;0.009; median within-pair FC\u0026thinsp;=\u0026thinsp;1.30, increased at follow-up). This signal did not remain significant after BH correction within the targeted panel. Glycerol-3-phosphate had the largest absolute fold-change estimate on the platform-level analysis, but this apparent change was driven by high missingness (only 7 of 14 pairs had both time points above the detection limit) and a single extreme baseline outlier; the paired Wilcoxon p-value for complete pairs was 0.58. We therefore regard the glycerol-3-phosphate signal as unreliable and the lactate finding as the only noteworthy exploratory observation on this panel.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e3.7 Exploratory Clustering and Correlation Analyses\u003c/h2\u003e \u003cp\u003eExploratory clustering of nominally changing lipids is shown in Supplementary Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e. Spearman correlation analysis among the 81 nominally changing species identified only six species with qualifying pairwise correlations (|r| \u0026gt; 0.8 and p\u0026thinsp;\u0026lt;\u0026thinsp;0.05), an insufficient number to construct an informative network; these results are therefore reported narratively rather than as a figure. These analyses provided supplementary context but were not treated as independent evidence.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eIn this paired pilot study of 14 ART-naive adults with HIV/TB co-infection, no lipid species remained significant after global multiple-testing correction across 1,230 features (smallest q\u0026thinsp;=\u0026thinsp;0.30). The dataset should therefore be interpreted as hypothesis-generating rather than confirmatory. Within that constraint, four lipids were mechanistically prioritized because they combined nominal paired differences with large effect sizes, observed non-imputed intensities in all paired samples, bootstrap confidence intervals for fold change that excluded 1.0, and directional consistency with the prespecified sphingolipid-focused framework. Three of these four candidates were sphingolipids\u0026mdash;Cer d18:1/26:0, SM 40:5, and SM 37:1\u0026mdash;while the fourth was the phosphatidylglycerol species PG 16:1\u0026ndash;22:1.\u003c/p\u003e \u003cp\u003eThe follow-up sample reflects the cumulative effect of multiple interventions begun within the first month of treatment. Zinc supplementation and HRZE started at baseline, while ART was added after 2 weeks. The second blood sample was therefore taken on average 4 weeks after zinc and anti-TB treatment had started, but only 2 weeks after the first ART exposure. Lipid changes should be interpreted as an integrated early-treatment signal, not as the effect of any single intervention. The follow-up sample was also collected during the early post-ART window in which treatment-related immune reconstitution may also be modulating host metabolism. One interpretation is that improved zinc status coincided with altered sphingolipid homeostasis during this period, which is biologically plausible because secretory acid sphingomyelinase is zinc-dependent and links sphingomyelin turnover to ceramide generation [\u003cspan additionalcitationids=\"CR9\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. However, the present data do not support a zinc-specific interpretation: the concurrent increase in ceramide and two sphingomyelin species argues against a simple explanation based solely on sphingomyelin hydrolysis and instead suggests that multiple processes were acting in parallel during this early period, including sphingolipid turnover, compensatory resynthesis, inflammatory resolution, anti-TB treatment effects, early ART-related remodeling, and general nutritional recovery.\u003c/p\u003e \u003cp\u003eThe increase in PG 16:1\u0026ndash;22:1 is also of interest, but its biological interpretation is less certain. Phosphatidylglycerol is a minor plasma phospholipid with known roles in pulmonary surfactant biology and innate immune signaling [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. In the context of pulmonary TB, changes in circulating PG species may be part of treatment-associated remodeling via the Kennedy pathway of de novo phospholipid synthesis and the Lands cycle of acyl-chain remodeling [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e], but plasma measures can only provide an indirect view of pulmonary events. On the targeted energy panel, lactate was the sole metabolite reaching nominal significance, consistent with increased glycolytic activity that accompanies immune cell activation during early treatment. Whether this observation reflects direct metabolic effects of zinc repletion, a shift in immune cell substrate utilization during CD4\u0026thinsp;+\u0026thinsp;T-cell recovery, or anti-TB drug effects cannot be determined from these data. The contrast between the two analytical platforms is itself informative: 81 of 1,230 lipid species reached nominal significance, whereas only 1 of 42 energy metabolites did so, despite the targeted panel carrying a lighter multiple-testing burden. One plausible explanation is that membrane phospholipid remodeling operates on a timescale of days to weeks as acyl chains are incorporated and redistributed through the Lands cycle, whereas central carbon intermediates turn over within hours [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e] and may have already reached a new steady state by the 4-week sampling point. Alternatively, the early treatment period may genuinely perturb structural lipid pools more than glycolytic or tricarboxylic acid cycle flux in this clinical context. This largely negative energy metabolomics result should not be dismissed; it narrows the hypothesis space by suggesting that the metabolic signature of early treatment in zinc-supplemented HIV/TB patients is predominantly lipid-centric rather than broadly metabolic.\u003c/p\u003e \u003cp\u003eAn important methodological point from this pilot dataset is that many nominal lipid signals were dominated by low-abundance features that were highly sensitive to imputation. In these cases, apparent fold changes can be inflated when near-threshold values are contrasted with detectable values. We based biological interpretation on signal completeness and paired robustness rather than large apparent fold changes alone. This approach substantially narrowed the candidate list and avoided overinterpreting low-confidence features. Thirty-six additional complete-data lipids met criteria 1\u0026ndash;4 but were not prioritized under criterion 5. Among these, DAG 20:1\u0026ndash;20:5 had the largest effect size in the dataset (r\u0026thinsp;=\u0026thinsp;0.86) and was the only lipid to reach within-class significance after Benjamini\u0026ndash;Hochberg correction (q\u0026thinsp;=\u0026thinsp;0.011, m\u0026thinsp;=\u0026thinsp;46 diacylglycerols). Its consistent downregulation (13 of 14 participants showed decreased levels) suggests a genuine biological signal that warrants independent investigation, but because this diacylglycerol species falls outside the prespecified sphingolipid-focused framework, it was not prioritized and should instead be considered a parallel candidate for future targeted validation. Additional unprioritized complete-data signals spanned multiple phospholipid subclasses, including PE, PS, LPC, and additional SM and Cer species, consistent with broad phospholipid remodeling during early treatment. These are documented in Supplementary Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e and represent candidates for future targeted validation.\u003c/p\u003e \u003cp\u003eRoutine clinical measures changed over the same interval, including plasma zinc (p\u0026thinsp;=\u0026thinsp;0.001), CD4\u0026thinsp;+\u0026thinsp;T-cell counts, HDL-C, albumin, and prealbumin, which increased, and ferritin (p\u0026thinsp;=\u0026thinsp;0.002) and uric acid (p\u0026thinsp;=\u0026thinsp;0.013), which decreased. Small dense LDL also rose (p\u0026thinsp;=\u0026thinsp;0.026), a finding that may reflect early ART-associated lipid remodeling, although the clinical significance of this change in the context of concurrent anti-TB therapy and nutritional recovery remains uncertain. A recent meta-analysis of 14 zinc supplementation trials in type 2 diabetes documented significant increases in HDL-C [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e], though the study populations and clinical contexts are not directly comparable. These changes are consistent with early treatment response and nutritional recovery, but do not help to parse out the specific contribution of zinc supplementation. The same caveat applies to the lipid findings: this study has captured a clinically meaningful early-treatment window, but the metabolic signals observed during that interval are likely to reflect several overlapping processes.\u003c/p\u003e \u003cp\u003eThe paired design reduced between-subject heterogeneity and allowed early within-subject shifts to be assessed in a clinically understudied population. Few lipidomics studies have focused on ART-naive adults with HIV/TB co-infection at the time of treatment initiation. The combination of untargeted lipidomics with targeted metabolomics provided complementary analytical perspectives, although convergence across platforms was limited.\u003c/p\u003e \u003cp\u003eThis study has several limitations. As a small single-arm paired study without a comparator group, it is not possible to isolate the specific effects of zinc supplementation from those of the anti-TB therapy initiated at baseline, ART added 2 weeks later, nutritional recovery, or treatment-related immune reconstitution. No lipid remained significant after global BH correction, and the four mechanistically prioritized candidates should be regarded as preliminary signals rather than validated biomarkers. The supervised multivariate model lacked robust predictive performance, and pathway annotation was restricted to a single robust candidate. Secretory acid sphingomyelinase activity was not measured, so the connection between zinc status and sphingolipid metabolism remains hypothetical. The two-time-point design captures only early changes and does not address whether the observed signals persist, intensify, or resolve over longer follow-up. The follow-up sample was obtained during the early post-ART period, within the clinical window in which treatment-related immune reconstitution may be affecting host metabolism, and the single-center design limits generalizability.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003esupplemented adults with HIV/TB co-infection was accompanied by several within-subject lipid shifts, but none remained significant after global multiple-testing correction. Four lipids\u0026mdash;Cer d18:1/26:0, SM 40:5, SM 37:1, and PG 16:1-22:1\u0026mdash;showed the most consistent directional evidence among hypothesis-guided candidates, supported by paired effect sizes, signal completeness, and bootstrap confidence intervals. Cer d18:1/26:0 approached within-class significance (q = 0.052 among 13 ceramides), while DAG 20:1-20:5 was the only lipid reaching within-class significance (q = 0.011) and may warrant parallel investigation. These data support focused validation rather than mechanistic inference. A future targeted study pre-specifying these four lipids, together with direct measurement of secretory acid sphingomyelinase activity and an appropriate comparator group, would provide a stronger test of the proposed biology.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAI-Assisted Writing Disclosure\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAI-based language tools were used only for English-language editing during manuscript preparation. No automated software contributed to the study design, data collection, data analysis, or scientific interpretation; all such work was carried out by the authors. The authors retain full responsibility for the scientific content, analyses, and conclusions of this manuscript.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSupported by Shenzhen Science and Technology Program (No. JCYJ20220530163212029) and Shenzhen Clinical Research Center for Emerging Infectious Diseases (No. LCYSSQ20220823091203007).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no conflicts of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics Approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eApproved by Shenzhen Third People\u0026apos;s Hospital Institutional Ethics Committee (2023-045-02).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to Participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWritten informed consent was obtained from all participants.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAvailable from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConceptualization: Chao Chen, Fang Zhao. Methodology: Shiyun Chen, Chao Chen. Formal Analysis: Wei Li, Yingying Zhang. Investigation: Wei Li, Fang Zhao, Miaona Liu. Resources: Haitao Zhang, Miaona Liu. Data Curation: Haitao Zhang, Fang Zhao. Writing \u0026ndash; Review \u0026amp; Editing: Wei Li, Yingying Zhang. Visualization: Man Rao, Chao Chen, Miaona Liu. Supervision: Fang Zhao, Wei Li. Project Administration: Fang Zhao. All authors have read and approved the final manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eWorld Health Organization (2024) Global Tuberculosis Report 2024. World Health Organization, Geneva\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM\u0026uuml;ller M, Wandel S, Colebunders R et al (2010) Immune reconstitution inflammatory syndrome in patients starting antiretroviral therapy for HIV infection: a systematic review and meta-analysis. Lancet Infect Dis 10:251\u0026ndash;261. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S1473-3099(10)70026-8\u003c/span\u003e\u003cspan address=\"10.1016/S1473-3099(10)70026-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaum MK, Campa A, Lai S, Lai H, Page JB (2003) Zinc status in human immunodeficiency virus type 1 infection and illicit drug use. Clin Infect Dis 37(Suppl 2):S117\u0026ndash;S123. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1086/375875\u003c/span\u003e\u003cspan address=\"10.1086/375875\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRead SA, Obeid S, Ahlenstiel C, Ahlenstiel G (2019) The role of zinc in antiviral immunity. Adv Nutr 10:696\u0026ndash;710. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/advances/nmz013\u003c/span\u003e\u003cspan address=\"10.1093/advances/nmz013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaum MK, Lai S, Sales S, Page JB, Campa A (2010) Randomized, controlled clinical trial of zinc supplementation to prevent immunological failure in HIV-infected adults. Clin Infect Dis 50:1653\u0026ndash;1660. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1086/652864\u003c/span\u003e\u003cspan address=\"10.1086/652864\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWagnew F, Alene KA, Eshetie S et al (2022) Effects of zinc and vitamin A supplementation on prognostic markers and treatment outcomes of adults with pulmonary tuberculosis: a systematic review and meta-analysis. BMJ Glob Health 7:e008625. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/bmjgh-2022-008625\u003c/span\u003e\u003cspan address=\"10.1136/bmjgh-2022-008625\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLawson L, Thacher TD, Yassin MA et al (2010) Randomized controlled trial of zinc and vitamin A as co-adjuvants for the treatment of pulmonary tuberculosis. Trop Med Int Health 15:1481\u0026ndash;1490. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1365-3156.2010.02638.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1365-3156.2010.02638.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu B, Obeid LM, Hannun YA (1997) Sphingomyelinases in cell regulation. Semin Cell Dev Biol 8:311\u0026ndash;322. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1006/scdb.1997.0153\u003c/span\u003e\u003cspan address=\"10.1006/scdb.1997.0153\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchuchman EH (2010) Acid sphingomyelinase, cell signalling, and disease. FEBS Lett 584:1895\u0026ndash;1900. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.febslet.2009.11.083\u003c/span\u003e\u003cspan address=\"10.1016/j.febslet.2009.11.083\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchissel SL, Keesler GA, Schuchman EH, Williams KJ, Tabas I (1998) The cellular trafficking and zinc dependence of secretory and lysosomal sphingomyelinase, two products of the acid sphingomyelinase gene. J Biol Chem 273:18250\u0026ndash;18259. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1074/jbc.273.29.18250\u003c/span\u003e\u003cspan address=\"10.1074/jbc.273.29.18250\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorales A, Lee H, Go\u0026ntilde;i FM, Kolesnick R, Fernandez-Checa JC (2007) Sphingolipids and cell death. Apoptosis 12:923\u0026ndash;939. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10495-007-0721-0\u003c/span\u003e\u003cspan address=\"10.1007/s10495-007-0721-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKoethe JR, Lagathu C, Lake JE et al (2020) HIV and antiretroviral therapy-related fat alterations. Nat Rev Dis Primers 6:48. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41572-020-0181-1\u003c/span\u003e\u003cspan address=\"10.1038/s41572-020-0181-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChaudhary NS, Kind T, Willig AL et al (2021) Changes in lipidomic profile by anti-retroviral treatment regimen: an ACTG 5257 ancillary study. Medicine 100:e26588. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1097/MD.0000000000026588\u003c/span\u003e\u003cspan address=\"10.1097/MD.0000000000026588\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMatyash V, Liebisch G, Kurzchalia TV, Shevchenko A, Schwudke D (2008) Lipid extraction by methyl-tert-butyl ether for high-throughput lipidomics. J Lipid Res 49:1137\u0026ndash;1146. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1194/jlr.D700041-JLR200\u003c/span\u003e\u003cspan address=\"10.1194/jlr.D700041-JLR200\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWishart DS, Guo A, Oler E et al (2022) HMDB 5.0: the Human Metabolome Database for 2022. Nucleic Acids Res 50:D622\u0026ndash;D631. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/nar/gkab1062\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkab1062\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFahy E, Sud M, Cotter D, Subramaniam S (2007) LIPID MAPS online tools for lipid research. Nucleic Acids Res 35:W606\u0026ndash;W612. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/nar/gkm324\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkm324\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTh\u0026eacute;venot EA, Roux A, Xu Y, Ezan E, Junot C (2015) Analysis of the human adult urinary metabolome variations with age, body mass index, and gender by implementing a comprehensive workflow for univariate and OPLS statistical analyses. J Proteome Res 14:3322\u0026ndash;3335. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1021/acs.jproteome.5b00354\u003c/span\u003e\u003cspan address=\"10.1021/acs.jproteome.5b00354\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBenjamini Y, Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B 57:289\u0026ndash;300. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.2517-6161.1995.tb02031.x\u003c/span\u003e\u003cspan address=\"10.1111/j.2517-6161.1995.tb02031.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCohen J (1988) Statistical power analysis for the behavioral sciences, 2nd edn. Lawrence Erlbaum Associates, Hillsdale\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKanehisa M, Goto S (2000) KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Res 28:27\u0026ndash;30. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/nar/28.1.27\u003c/span\u003e\u003cspan address=\"10.1093/nar/28.1.27\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHan S, Mallampalli RK (2015) The role of surfactant in lung disease and host defense against pulmonary infections. Ann Am Thorac Soc 12:765\u0026ndash;774. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1513/AnnalsATS.201411-507FR\u003c/span\u003e\u003cspan address=\"10.1513/AnnalsATS.201411-507FR\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKennedy EP, Weiss SB (1956) The function of cytidine coenzymes in the biosynthesis of phospholipides. J Biol Chem 222:193\u0026ndash;214. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0021-9258(19)50785-2\u003c/span\u003e\u003cspan address=\"10.1016/S0021-9258(19)50785-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eValentine WJ, Shimizu T, Shindou H (2023) Lysophospholipid acyltransferases orchestrate the compositional diversity of phospholipids. Biochimie 215:24\u0026ndash;33. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.biochi.2023.08.012\u003c/span\u003e\u003cspan address=\"10.1016/j.biochi.2023.08.012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHeidari Seyedmahalleh M, Montazer M, Ebrahimpour-Koujan S, Azadbakht L (2023) The effect of zinc supplementation on lipid profiles in patients with type 2 diabetes mellitus: a systematic review and dose-response meta-analysis of randomized clinical trials. Adv Nutr 14:1374\u0026ndash;1388. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.advnut.2023.08.006\u003c/span\u003e\u003cspan address=\"10.1016/j.advnut.2023.08.006\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDooley KE, Kaplan R, Mwelase N et al (2020) Dolutegravir-based antiretroviral therapy for patients coinfected with tuberculosis and human immunodeficiency virus: a multicenter, noncomparative, open-label, randomized trial. Clin Infect Dis 70:549\u0026ndash;556. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/cid/ciz256\u003c/span\u003e\u003cspan address=\"10.1093/cid/ciz256\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJang C, Chen L, Rabinowitz JD (2018) Metabolomics and isotope tracing. Cell 173:822\u0026ndash;837. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.cell.2018.03.055\u003c/span\u003e\u003cspan address=\"10.1016/j.cell.2018.03.055\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1. Baseline characteristics of 14 participants\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"650\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 46.1538%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCharacteristic\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 38.4615%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eValue\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.3846%;\"\u003e\n \u003cp\u003e\u003cstrong\u003en\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 46.1538%;\"\u003e\n \u003cp\u003eAge, years\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 38.4615%;\"\u003e\n \u003cp\u003e41.1 \u0026plusmn; 13.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.3846%;\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 46.1538%;\"\u003e\n \u003cp\u003eSex (M/F)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 38.4615%;\"\u003e\n \u003cp\u003e11/3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.3846%;\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 46.1538%;\"\u003e\n \u003cp\u003eBMI, kg/m\u0026sup2;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 38.4615%;\"\u003e\n \u003cp\u003e19.34 \u0026plusmn; 4.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.3846%;\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 46.1538%;\"\u003e\n \u003cp\u003eBaseline CD4+ T-cells, cells/\u0026mu;L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 38.4615%;\"\u003e\n \u003cp\u003e95.8 \u0026plusmn; 75.7 (21\u0026ndash;237)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.3846%;\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 46.1538%;\"\u003e\n \u003cp\u003eBaseline plasma zinc, \u0026mu;mol/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 38.4615%;\"\u003e\n \u003cp\u003e11.32 \u0026plusmn; 1.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.3846%;\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 46.1538%;\"\u003e\n \u003cp\u003eSampling interval, days\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 38.4615%;\"\u003e\n \u003cp\u003e29.6 \u0026plusmn; 1.9 (28\u0026ndash;34)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.3846%;\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eData are mean \u0026plusmn; SD (range where indicated).\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2. Clinical and biochemical parameters at baseline and follow-up\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"613\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eParameter\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eBaseline\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFollow-up\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eChange (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ep\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSig.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003ePlasma zinc, \u0026mu;mol/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e11.32 \u0026plusmn; 1.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e13.06 \u0026plusmn; 1.78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e+15.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eCD4+ T-cells, cells/\u0026mu;L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e95.8 \u0026plusmn; 75.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e191.6 \u0026plusmn; 96.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e+100.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eIL-6, pg/mL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e30.66 \u0026plusmn; 33.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e8.52 \u0026plusmn; 9.66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e\u0026minus;72.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.058\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eTotal cholesterol, mmol/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e3.49 \u0026plusmn; 0.92\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e4.27 \u0026plusmn; 1.19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e+22.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.064\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eTriglycerides, mmol/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e1.33 \u0026plusmn; 0.92\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e1.41 \u0026plusmn; 0.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e+6.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.808\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eHDL-C, mmol/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e0.89 \u0026plusmn; 0.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e1.35 \u0026plusmn; 0.51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e+51.6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eLDL-C, mmol/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e2.04 \u0026plusmn; 0.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e2.28 \u0026plusmn; 0.95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e+11.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.397\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eApoA1, g/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e0.84 \u0026plusmn; 0.38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e1.05 \u0026plusmn; 0.49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e+23.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.198\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eApoB, g/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e0.87 \u0026plusmn; 0.15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e1.19 \u0026plusmn; 0.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e+36.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eLp(a), mg/dL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e23.7 \u0026plusmn; 31.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e11.7 \u0026plusmn; 27.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e\u0026minus;50.6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.028\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003esdLDL, mmol/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e0.49 \u0026plusmn; 0.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e0.72 \u0026plusmn; 0.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e+46.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.026\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eAlbumin, g/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e36.0 \u0026plusmn; 4.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e39.3 \u0026plusmn; 4.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e+9.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.028\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003ePrealbumin, mg/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e158.9 \u0026plusmn; 71.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e213.7 \u0026plusmn; 45.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e+34.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.035\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eALT, U/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e24.4 \u0026plusmn; 18.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e26.6 \u0026plusmn; 18.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e+8.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.401\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eAST, U/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e35.0 \u0026plusmn; 27.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e27.7 \u0026plusmn; 8.6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e\u0026minus;20.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.777\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eGGT, U/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e95.5 \u0026plusmn; 147.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e82.6 \u0026plusmn; 60.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e\u0026minus;13.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.463\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eFasting glucose, mmol/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e4.77 \u0026plusmn; 0.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e5.25 \u0026plusmn; 0.66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e+10.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.035\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eFerritin, ng/mL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e812.5 \u0026plusmn; 514.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e401.5 \u0026plusmn; 183.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e\u0026minus;50.6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e**\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eUric acid, \u0026mu;mol/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e374.1 \u0026plusmn; 110.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e543.2 \u0026plusmn; 206.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e+45.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.013\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 30.5057%;\"\u003e\n \u003cp\u003eeGFR, mL/min/1.73m\u0026sup2;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e110.4 \u0026plusmn; 24.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.5759%;\"\u003e\n \u003cp\u003e106.6 \u0026plusmn; 19.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e\u0026minus;3.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.9086%;\"\u003e\n \u003cp\u003e0.241\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.52529%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eData are mean \u0026plusmn; SD. P values from two-sided paired Wilcoxon signed-rank tests. Change (%) = ((follow-up \u0026minus; baseline)/baseline) \u0026times; 100. *P \u0026lt; 0.05; **P \u0026lt; 0.01; ***P \u0026lt; 0.001. HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; ApoA1, apolipoprotein A1; ApoB, apolipoprotein B; Lp(a), lipoprotein(a); sdLDL, small dense LDL; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase; eGFR, estimated glomerular filtration rate.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3. Mechanistically prioritized candidate lipid species\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"587\" class=\"fr-table-selection-hover\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.0853%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLipid species\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eClass\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2389%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ep\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003e\u003cstrong\u003er\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFC\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.8703%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e95% CI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 8.02048%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDir.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.6519%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eq (global)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.0853%;\"\u003e\n \u003cp\u003eCer d18:1/26:0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003eCer\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2389%;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003e0.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003e1.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.8703%;\"\u003e\n \u003cp\u003e1.16\u0026ndash;1.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 8.02048%;\"\u003e\n \u003cp\u003eUp\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.6519%;\"\u003e\n \u003cp\u003e0.55\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.0853%;\"\u003e\n \u003cp\u003eSM 40:5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003eSM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2389%;\"\u003e\n \u003cp\u003e0.007\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003e0.70\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003e1.44\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.8703%;\"\u003e\n \u003cp\u003e1.05\u0026ndash;1.64\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 8.02048%;\"\u003e\n \u003cp\u003eUp\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.6519%;\"\u003e\n \u003cp\u003e0.64\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.0853%;\"\u003e\n \u003cp\u003eSM 37:1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003eSM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2389%;\"\u003e\n \u003cp\u003e0.013\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003e0.65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003e1.17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.8703%;\"\u003e\n \u003cp\u003e1.02\u0026ndash;1.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 8.02048%;\"\u003e\n \u003cp\u003eUp\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.6519%;\"\u003e\n \u003cp\u003e0.69\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.0853%;\"\u003e\n \u003cp\u003ePG 16:1-22:1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003ePG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2389%;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003e0.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.04437%;\"\u003e\n \u003cp\u003e1.35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15.8703%;\"\u003e\n \u003cp\u003e1.02\u0026ndash;1.66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 8.02048%;\"\u003e\n \u003cp\u003eUp\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 13.6519%;\"\u003e\n \u003cp\u003e0.55\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003ep, exact paired Wilcoxon signed-rank test. r, within-subject effect size (|Z|/\u0026radic;N). FC, median of within-pair fold changes (follow-up/baseline). CI, bootstrap 95% confidence interval (10,000 resamples). Dir., direction of change. q (global), Benjamini\u0026ndash;Hochberg adjusted p-value across 1,230 features. All four candidates had observed non-imputed intensities in all 14 paired samples.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3\u0026nbsp;\u003c/strong\u003ePrioritized candidate lipid species based on paired effect size, signal completeness, and bootstrap confidence intervals.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"biological-trace-element-research","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bter","sideBox":"Learn more about [Biological Trace Element Research](https://www.springer.com/journal/12011)","snPcode":"12011","submissionUrl":"https://submission.nature.com/new-submission/12011/3","title":"Biological Trace Element Research","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"zinc supplementation, HIV-tuberculosis co-infection, lipidomics, sphingolipid metabolism, phosphatidylglycerol, exploratory analysis","lastPublishedDoi":"10.21203/rs.3.rs-9309714/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9309714/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eBackground: Zinc deficiency is frequent in HIV and may impact sphingolipid metabolism via the zinc-dependent secretory acid sphingomyelinase. To date, associations of early treatment with zinc-supplemented adults with HIV/tuberculosis (TB) co-infection with plasma lipidomic alterations have not been explored.\u003c/p\u003e\n\u003cp\u003eMethods: Paired fasting plasma samples were collected at baseline and ~4 weeks from 14 antiretroviral therapy (ART)-naive adults with HIV/TB co-infection initiating oral zinc gluconate (40 mg/day elemental zinc), standard anti-TB therapy (from baseline), and ART (from week 2). Untargeted UHPLC-MS/MS lipidomics detected 1,230 lipid species, and targeted metabolomics quantified 42 energy-related metabolites. Candidate lipids were prioritized by paired Wilcoxon p-values, within-subject effect sizes, signal completeness, and bootstrap confidence intervals.\u003c/p\u003e\n\u003cp\u003eResults: At follow-up, plasma zinc concentrations were higher (11.32 ± 1.39 to 13.06 ± 1.78 μmol/L; p = 0.001), and CD4+ T-cell counts had doubled (95.8 ± 75.7 to 191.6 ± 96.0 cells/μL; p = 0.0004). No lipid met the global Benjamini–Hochberg correction (all q \u0026gt; 0.30). From the 81 nominally significant species (p \u0026lt; 0.05), four lipids with complete observed data, largest effect sizes (r = 0.65–0.76), and bootstrap confidence intervals (CIs) that excluded 1.0 were mechanistically prioritized: Cer d18:1/26:0 (fold change [FC] 1.31), SM 40:5 (FC 1.44), SM 37:1 (FC 1.17), and PG 16:1-22:1 (FC 1.35), all increased at follow-up. Lactate was the only energy metabolite with nominal significance (p = 0.009).\u003c/p\u003e\n\u003cp\u003eConclusion: These hypothesis-generating results highlight four candidate lipids with consistent, within-subject shifts in response to early HIV/TB treatment. The effects of zinc supplementation, anti-TB therapy, ART, and immune reconstitution cannot be distinguished.\u003c/p\u003e","manuscriptTitle":"Plasma sphingolipid and glycerophospholipid shifts during early treatment in zinc-supplemented adults with HIV–tuberculosis co-infection: a paired exploratory lipidomics study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-22 17:41:26","doi":"10.21203/rs.3.rs-9309714/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-04-29T14:48:48+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"156237261254050295476335716543410031972","date":"2026-04-20T14:37:10+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"151576745364714790282999243669172758490","date":"2026-04-19T20:24:37+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-14T19:29:19+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-11T14:16:12+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-11T01:43:51+00:00","index":"","fulltext":""},{"type":"submitted","content":"Biological Trace Element Research","date":"2026-04-03T06:46:37+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"biological-trace-element-research","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bter","sideBox":"Learn more about [Biological Trace Element Research](https://www.springer.com/journal/12011)","snPcode":"12011","submissionUrl":"https://submission.nature.com/new-submission/12011/3","title":"Biological Trace Element Research","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"29fa38b3-31b7-40ce-93b0-da617ef87d16","owner":[],"postedDate":"April 22nd, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-04-29T14:48:48+00:00","index":17,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-22T17:41:26+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-22 17:41:26","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9309714","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9309714","identity":"rs-9309714","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}