Lipidomic signatures in lung adenocarcinoma and spontaneous pneumothorax tissues associated with heated tobacco use: a pilot study.

Introduction
Cigarette smoking affects lipid metabolism in both lung and lung cancer tissues. Lipidomics, which allows semi-quantitative analysis of lipid species in biological samples, can be used to infer the impact of smoking on these tissues [1–5]. Although cigarette-induced lipid changes have been characterized, the metabolic effects of heated tobacco products (HTPs) remain largely unexplored [6]. Given that the lipid profile can be regarded as a metabolic “resume” of the tissue, lipidomics enables the generation of new hypotheses regarding the effects of HTPs on lung and lung cancer tissues.
HTPs heat tobacco without combustion, thereby reducing toxic emissions [7–11]. However, recent studies have shown that aerosols from HTPs can still induce oxidative stress, DNA damage [12], and epithelial–mesenchymal transition in airway cells [13]. These findings suggest that HTPs may influence lipid metabolism, which reflects their effect on lung and lung cancer tissues, but direct evidence from human tissues is lacking.
To address this gap, we performed a pilot lipidomic analysis of surgical lung and lung cancer tissues from patients with lung adenocarcinoma (LAD) and spontaneous pneumothorax, comparing profiles among HTP-users, cigarette-smokers, and never-smokers using liquid chromatography-tandem mass spectrometry (LC–MS/MS). This study aimed to identify lipid signatures potentially associated with HTP exposure in lung and lung cancer tissues.

Methods

Patients and tissue samples
Frozen LAD and pneumothorax tissues resected at Hamamatsu University Hospital (2022–2024) were analyzed. In LAD cases, primary tumor and normal lung tissues were obtained. In pneumothorax cases, normal lung tissue was collected from the area farthest from the pulmonary cyst within the same specimen. Among the LAD cases, two patients were HTP users (HTP-user group), and both exhibited the papillary adenocarcinoma (ADC) subtype. Therefore, as controls, we selected two cases of papillary ADC from cigarette smokers (cigarette-smoker group) and two from never-smokers (never-smoker group). Because priority was given to matching samples by the papillary ADC subtype, all never-smokers with available tissue samples were female. In the spontaneous pneumothorax cases, two patients were identified as HTP-users (HTP-user group). Accordingly, two cigarette smokers (cigarette-smoker group) and two never-smokers (never-smoker group) were selected as controls.

Histopathological evaluation
Hematoxylin and eosin staining was used to evaluate inflammatory changes, including lymphocytic infiltration and pleural thickening. Regarding the LAD cases, pathological diagnosis and staging were performed according to the World Health Organization criteria [14] and the 9th edition of the TNM classification for lung cancer [15], respectively.

Lipid extraction from tissue samples
As described in our previous report, we extracted lipids from the frozen tissue samples [16]. In brief, tissue samples (range: 5.2–16.2 mg, mean: 10.8 mg, standard deviation: 2.4 mg) were weighed using an analytical balance CPA224S (Sartorius AG, Göttingen, Germany) with a readability of 0.1 mg, reflecting variation in sample size rather than weighing error (Additional file 1, sheet 1). The tissue samples were subjected to modified Bligh-Dyer method for lipid extraction [16]. During the extraction, 1.6 mmol of 1,2-dilauroyl-sn-glycero-3-PC (Avanti Polar Lipids, Alabaster, AL, USA), phosphatidylcholine (PC) (12:0_12:0) per 1 mg of sample was added to standardize the lipid levels. The extracted lipids were dried using rotary evaporator. We dissolved the dried lipids with methanol proportionally to the tissue weights to keep the PC (12:0_12:0) levels similar among the cases. The dried lipid was dissolved with 20 µL of methanol, and 8 µL of the dissolved lipids were diluted with methanol proportional to the weight of the original tissue samples.

Lipid analysis by LC–MS/MS
As reported previously, the diluted lipids were subjected to LC–MS/MS [16]. In brief, we applied one µL of the diluted lipids to an Acclaim 120 C18 column (150 mm × 2.1 mm, 3 μm) (Thermo Fisher Scientific, Waltham, MA, USA) and measured using a Q Exactive™ Hybrid Quadrupole-Orbitrap™ Mass Spectrometer (Thermo Fisher Scientific) with an UltiMate 3000 (Thermo Fisher Scientific). Spectral data were recorded by Xcalibur v3.0 Software (Thermo Fisher Scientific). Ionization was performed in both positive and negative electrospray ionization modes. Lipid peak identification and semi-quantification were performed using LipidSearch™ software version 5.2 (Mitsui Knowledge Industry, Tokyo, Japan) in LC–MS product search mode. Default mammalian lipid identification settings (plasma, tissue, cells) were applied. The precursor mass tolerance was set to ± 5 ppm, and product ion tolerance to ± 8 ppm. A product ion threshold of 1% and a retention-time tolerance of 0.5 min were used for MS/MS matching. Lipids were accepted when they fulfilled LipidSearch identification criteria (grades A–C) based on precursor–product ion matching and software-built-in scoring. Signal thresholds followed the LipidSearch default peak detection settings, and peaks not meeting the minimum product ion threshold (1%) or lacking characteristic MS/MS fragments required for annotation were excluded (used parameters are shown in Additional file 1, sheet 2). An area value of a lipid species was divided by that of the standard PC (12:0_12:0) in the corresponding case for normalization to tissue weight. The complete list of normalized lipid peaks is presented in Additional file 1, sheet 3.

Data processing
First, for each lipid species identified using LipidSearch™ software, the normalized area values of all lipid species belonging to the same lipid class [17] were summed (Additional file 1, sheet 4). Next, for each lipid class, the area value for each group was divided by the mean area value of the never-smoker group to calculate the relative intensity. The heatmap visualization of the relative intensities was generated using MetaboAnalyst 6.0 (https://www.metaboanalyst.ca/MetaboAnalyst/ModuleView.xhtml). Finally, for each lipid class, the mean relative intensity in the HTP-user and cigarette-smoker, and never-smoker groups was divided by the mean value in the never-smoker group to calculate the fold change (FC). Lipid classes with a FC ≥ 2 or FC ≤ 0.5 in the HTP-user group were screened as lipid profiles potentially affected by HTP use. To assess extraction and instrumental variability, the area value of the internal standard PC (12:0/12:0) was normalized to tissue weight (mg), and the coefficient of variation (CV) of this weight-normalized internal standard signal across all samples was calculated.

Results

Patient characteristics
In LAD cases (Table 1), the HTP-user and cigarette-smoker groups comprised male patients, while the never-smokers were female. In the HTP-user group, case 1 had smoked 20 sticks/day of IQOS (Philip Morris International, Lausanne, Switzerland) for three years until 25 days before surgery. Case 2 had smoked 20 sticks/day of Ploom TECH (Japan Tobacco Inc., Tokyo, Japan) for one year until three days before surgery, after previously smoking 20 sticks/day of IQOS for six months. Both cases had prior cigarette histories before switching exclusively to HTP use.

In pneumothorax cases (Table 2), all patients were male. In the HTP-user group, case 1 had smoked 15 sticks/day of Ploom X (Japan Tobacco Inc., Tokyo, Japan) for one year until the day before surgery, and case 2 had smoked 10 sticks/day of IQOS for two years until the day before surgery. Both cases had previous cigarette exposures before switching to HTP use.

Histopathological findings
In the LAD and spontaneous pneumothorax cases, histopathology revealed emphysematous changes and inflammatory infiltrates in both HTP-user and cigarette-smoker groups (Additional file 2, Supplemental Fig. 1, 2).

Screening of lipid signature influenced by HTP use
In the LC–MS/MS analysis, a total of 1,982 lipid species belonging to 24 lipid classes were identified. The weight-normalized internal standard signal (PC (12:0/12:0) area value per tissue weight) showed a CV of 44.5% across samples. The relative intensity of each lipid class is visualized in a heatmap (Fig. 1). Some lipid classes exhibited HTP-user group-specific alterations in LAD (Fig. 1, left panel) and spontaneous pneumothorax cases (Fig. 1, right panel).

Lipid classes with a FC ≥ 2 or ≤ 0.5 in the HTP-user group were screened in both LAD and spontaneous pneumothorax cases (Table 3). In LAD cases, acylcarnitine (AcCa), coenzyme Q10 (CoQ10), triglyceride (TG), and phosphatidylethanolamine (PE) were high in cancer tissue, whereas TG, monohexosylceramide (Hex1Cer), and CoQ10 were high in normal lung. In contrast, cholesteryl ester (ChE) and cholesterol (Ch) were consistently reduced in both cancer and normal tissues, accompanied by low lysophospholipids (lysosphingomyelin [LSM], lysophosphatidylethanolamine [LPE], lysophosphatidylcholine [LPC]) and phosphatidylserine (PS) in normal lung tissue. In spontaneous pneumothorax cases, CoQ10 was high, whereas AcCa was low.

Lipid profile alterations according to HTP brands
Brand-specific patterns were observed among HTP-users. In LAD cases (Fig. 2), the IQOS-exclusive user (case 1) exhibited higher levels of CoQ10, TG, AcCa, and PE in cancer tissue, as well as TG and Hex1Cer in normal lung tissue. In contrast, the Ploom TECH user (case 2, who had switched from IQOS) exhibited a higher level of CoQ10 in normal lung tissue. Both brands demonstrated consistently lower levels of ChE, Ch in cancer and normal lung tissues, as well as lower levels of lysophospholipids (LSM, LPE and LPC) in normal lung tissue. In spontaneous pneumothorax cases (Fig. 3), the Ploom X user (case 1) had a higher level of CoQ10. In contrast, AcCa levels remained low across all cases in both the HTP-user and cigarette-smoker groups.

Discussion
In this study, we identified lipid profiles in lung cancer and lung tissue that may have been altered in response to HTP use.
AcCa, which showed the highest elevation in cancer tissue, is known to transport long-chain fatty acids into mitochondria and participate in energy production through β–oxidation [18]. CoQ10, which was high in LAD and spontaneous pneumothorax cases, functions as a coenzyme in the mitochondrial electron transport chain for adenosine triphosphate production and possesses antioxidant properties [19]. In spontaneous pneumothorax cases, two HTP-user cases showed low AcCa levels. By contrast, CoQ10 was high only in case 1 (Ploom X user), but not in case 2 (IQOS user). A plausible hypothesis is that HTP use may induce oxidative stress, leading to (i) mitochondrial dysfunction caused by oxidative stress may lead to impaired AcCa metabolism and its subsequent accumulation [20] and (ii) enhanced biosynthesis of the antioxidant CoQ10 may occur as a cellular defense response. The pattern observed in spontaneous pneumothorax cases indicates that AcCa and CoQ10 may show stress-responsive changes similar to those observed in the normal lung tissue of LAD cases.
TG was high in both tumor and normal lung tissues of LAD cases. TG is the form in which three acyl groups are esterified to glycerol. Glycerol, which is added to HTPs as a volatile agent, has been suggested to deposit in the lung and trigger inflammation [21]. Notably, high TG was most prominent in the case of long-term IQOS use (case 1). Compositional analyses of HTPs have shown that IQOS contains a higher amount of glycerol compared to Ploom TECH [22]. One possible explanation is that glycerol may deposit in lung tissue and could potentially be acylated to form TG, leading to intracellular accumulation [23]. Also, the observed increase in TG levels in the long-term IQOS smoker of this study is consistent with previous reports showing that IQOS contains more glycerol than Ploom TECH.
PE was markedly high in the cancer tissue of the long-term IQOS user (case 1). A plausible hypothesis is that enhanced PE synthesis and uptake may contribute to cancer cell proliferation [24, 25]. In support of this hypothesis, recent lipidomic studies in early-onset lung cancer have identified specific PE species as early metabolic signatures associated with lung tumorigenesis and smoking-related lipid remodeling [26]. Taken together, these findings suggest that the elevated PE observed in our HTP-users may reflect similar stress-induced membrane remodeling and early tumor-associated metabolic reprogramming. Our observation may provide a basis for generating the hypothesis that HTP use indirectly promotes lung cancer growth.
Hex1Cer was prominently high in the normal lung tissue of the long-term IQOS smoker (case 1). One possible explanation is that smoking-related oxidative stress increases ceramide [4], which cells may glycosylate into Hex1Cer as a mechanism to avoid apoptosis and acquire stress resistance [27]. Thus, elevated Hex1Cer may reflect stress-adaptive responses of normal lung tissue to HTP-related stress.
Both ChE and Ch were low in LAD cancer and normal lung tissues. Low levels of ChE and Ch in LAD cases may be explained by oxidative stress–induced metabolic reprogramming. Under oxidative stress, cells may shift their energy metabolism toward mitochondrial fatty acid β–oxidation. This process can be accompanied by enhanced mobilization and breakdown of neutral lipids and cholesteryl esters stored in lipid droplets, thereby supplying fatty acids for mitochondrial oxidation [28]. Accordingly, the reduced levels of cholesterol and cholesteryl esters might reflect stress-driven lipid catabolism fueling β–oxidation.
In the HTP-user group, normal lung tissue exhibited low levels of lysophospholipids (LSM, LPE, and LPC). Since these metabolites are mainly generated through phospholipase A2 (PLA2)–mediated hydrolysis of membrane phospholipids [29], their reduction may be explained by suppressed PLA2 activity [30] or enhanced reacylation via the Lands’ cycle [31] under HTP-induced oxidative stress. Such alterations could indicate adaptive membrane remodeling and modulation of inflammatory signaling in response to chronic HTP exposure.
In the HTP-user group, normal lung tissue exhibited low-level PS. Given that PS is a critical membrane phospholipid involved in apoptotic signaling and immune regulation [32], possible explanations for low-level PS include impaired biosynthesis under oxidative stress, enhanced membrane remodeling through the Lands’ cycle [31], or increased externalization and clearance during stress-induced apoptosis. These alterations may suggest that HTP exposure could influence stress responses and immune regulation in normal lung tissue through PS depletion.
The low level of AcCa observed in the HTP-user and cigarette-smoker groups of spontaneous pneumothorax cases may be explained primarily by substrate limitation. In emphysematous and fibrotic lung tissue with marked inflammatory cell infiltration, fatty acid uptake and activation are likely impaired, leading to insufficient acyl-CoA supply for mitochondrial transport via the carnitine shuttle [33]. As a result, AcCa levels may decline due to reduced substrate availability.
Several findings in this study indicated oxidative-stress–related perturbations. Given that HTP aerosols are chemically characterized by high levels of glycerol, significant nicotine delivery, and measurable carbonyl species arising from low-temperature pyrolysis, as reported in recent analytical reviews [34, 35], these constituents likely contribute to the oxidative burden associated with HTP use. Thus, the chemical composition of HTP aerosols provides a mechanistic context that supports our hypothesis.
Recent mechanistic studies further support the notion that HTP exposure can trigger oxidative stress in lung tissues. For example, IQOS cigarette smoke extract activates the NF-E2 p45-related factor 2 (NRF2) pathway and upregulates canonical oxidative stress–responsive genes, including Hmox1, Gsta1, Gsta3, and Nqo1 [36]. These NRF2-driven responses are consistent with the oxidative signatures observed with conventional cigarette smoke. Considering these findings, the lipidomic alterations identified in our study may represent downstream metabolic consequences of HTP-induced oxidative stress.
In conclusion, this exploratory pilot study suggests that HTP use may be associated with distinct alterations in lipid metabolism in lung cancer and lung tissues. Observed patterns included elevations of AcCa, CoQ10, PE, and TG, together with reductions in ChE, Ch, and several lysophospholipids, which may reflect oxidative stress–driven metabolic reprogramming, membrane remodeling, and stress-adaptive responses. These findings are exploratory and should be interpreted with caution, given the small sample size and the fact that all HTP-users had a history of conventional cigarette smoking. Nonetheless, this pilot analysis highlights the possibility that HTPs exert unique effects on lipid homeostasis in the lung, and underscores the need for validation in larger, prospective studies including exclusive HTP users.

Limitations
This pilot study was limited by the small sample size and prior cigarette exposure among all HTP-users, making it difficult to distinguish independent HTP effects. Histopathological findings showed emphysematous changes and inflammatory infiltrates in both the HTP-user and cigarette-smoker groups in the LAD and spontaneous pneumothorax cases. Given that no pathological findings specific to the HTP-user group were observed, it is impossible from a pathological standpoint to distinguish the effects of a prior history of cigarette smoking from those of HTP use. Therefore, from a lipidomics perspective as well, we surmise that excluding the influence of cigarette smoking in the current cohort is not feasible. Because all HTP-user cases with available tissue samples had a history of cigarette smoking, and because younger individuals were included in the HTP-user group (case 2 in the LAD cohort and case 1 in the spontaneous pneumothorax cohort), and all available papillary ADC samples from never-smokers were obtained from female patients, potential confounders such as smoking history, age, sex, and comorbidities could not be controlled. Larger studies with exclusive HTP-users are warranted.
The weight-normalized internal standard signal showed a CV of 44.5% across samples. This relatively high CV likely reflects variability in manual extraction steps, including incomplete solvent transfer and re-dissolution of dried lipids, and therefore represents the overall technical variability of our workflow.
Abbreviations: AcCa, acylcarnitine; Ch, cholesterol; ChE, cholesteryl ester; CoQ10, coenzyme Q10; FC, fold change; Hex1Cer, monohexosylceramide; HTP, heated tobacco product; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; LSM, lysosphingomyelin; PE, phosphatidylethanolamine; PS, phosphatidylserine; TG, triglyceride.

Supplementary Information
Below is the link to the electronic supplementary material.