Dietary triacylglycerol hydroperoxide is not absorbed, yet it induces the formation of other triacylglycerol hydroperoxides in the gastrointestinal tract

The in vivo presence of triacylglycerol hydroperoxide (TGOOH), a primary oxidation product of triacylglycerol (TG), has been speculated to be involved in various diseases. Thus, considerable attention has been paid to whether dietary TGOOH is absorbed from the intestine. In this study, we performed the lymph duct-cannulation study in rats and analyzed the level of TGOOH in lymph following administration of a TG emulsion containing TGOOH. As we successfully detected TGOOH from the lymph, we hypothesized that this might be originated from the intestinal absorption of dietary TGOOH [hypothesis I] and/or the in situ formation of TGOOH [hypothesis II]. To determine the validity of these hypotheses, we then performed another cannulation study using a TG emulsion containing a deuterium-labeled TGOOH (D2-TGOOH) that is traceable in vivo. After administration of this emulsion to rats, we clearly detected unlabeled TGOOH instead of D2-TGOOH from the lymph, indicating that TGOOH is not absorbed from the intestine but is more likely to be produced in situ. By discriminating the isomeric structures of TGOOH present in lymph, we predicted the mechanism by which the intake of dietary TGOOH triggers oxidative stress (e.g., via generation of singlet oxygen) and induces in situ formation of TGOOH. The results of this study hereby provide a foothold to better understand the physiological significance of TGOOH on human health.


Introduction
Lipids are essential nutrients that are widely contained in foods [1]. The major lipid in foods is triacylglycerol (TG), which consists of three fatty acids esterified to a glycerol backbone [2,3]. During excessive food processing and/or inadequate storage conditions, the fatty acid moieties of TG undergo radical (e.g., thermal) and/or singlet oxygen ( 1 O 2 ) (e.g., photo) oxidation to form TG hydroperoxide (TGOOH) (Fig. 1) [4]. Previously, our group and others have shown the presence of a small amount of TGOOH even in fresh edible oils [5] (e.g., canola oil with a peroxide value (POV) of 0.8-2.0 meq/kg containing 10-50 nmol/mL of TGOOH [4]), which suggest that we may ingest a certain amount of TGOOH through daily food intake. Meanwhile, TGOOH is reportedly present in vivo (e.g., in lipoproteins), and such TGOOH is speculated to be involved in the onset and progression of various diseases (e.g., cardiovascular diseases) [6,7]. Due to these, considerable attention has been paid to the absorption of dietary TGOOH in vivo.
To the best of our knowledge, only a few studies have evaluated the absorption of TGOOH. For instance, Kanazawa et al. and Mohr et al. administered oils containing TGOOH (20 μmol [8] and 0.6 μmol [9], respectively) to rats; following this, however, the levels of TGOOH were below the detection limit in biological samples (e.g., intestinal tract and lymph fluid). In line with these studies, Suomela et al. also reported that the levels of TGOOH in the intestinal tract of pigs fed highly oxidized oils (POV: 190 meq/kg) for two weeks were below detection limits [10]. These results collectively indicated that TGOOH is barely absorbed from the gastrointestinal tract. However, given the somewhat low sensitivity of the methods used in these studies [8][9][10] and the above indication of the presence of TGOOH in vivo [6,7], it is worthwhile to provide conclusive data regarding the absorption of dietary TGOOH using more sensitive methods to better understand its physiological significance in vivo.
The conventional methods to detect TGOOH (e.g., UV [8], chemiluminescence [9], and mass spectrometry methods [10]) have detection sensitivities ranging from μmol-pmol levels. Meanwhile, we recently developed a more sensitive method to analyze TGOOH using liquid chromatography-tandem mass spectrometry (HPLC-MS/MS), enabling detection at the fmol level [4]. Using this method, interestingly, we were able to detect TGOOH from the lymph fluid (in the order of fmol/μL) collected from rats given a TG emulsion containing 30 nmol of TGOOH (unpublished preliminary data). From this result, we deduced that a part of dietary TGOOH can be absorbed [hypothesis I]. Alternatively, it may also be possible that the ingested TGOOH triggered oxidative stress (e.g., generation of reactive oxygen species such as free radicals and 1 O 2 [12]), leading to the in situ formation of TGOOH that we detected in the lymph [hypothesis II]. On top of that, our HPLC-MS/MS-based method is not only highly sensitive but also allows us to assess the oxidative stress (i.e., radical or 1 O 2 oxidation) involved in TGOOH formation by discriminating the isomeric structure of TGOOH (Fig. 2) [4,11], such feature is considered to be useful in examining the validity of hypothesis I and/or II.
In this study, we first collected and analyzed lymph from rats following administration of a TG emulsion containing 32 nmol of TGOOH to confirm the reproducibility of our preliminary experiment. As we successfully detected TGOOH from the lymph, we then administered a TG emulsion containing a deuterium-labeled TGOOH (D2-TGOOH) that is traceable in vivo. By doing so, we aimed to determine the validity of the above hypotheses (i.e., the presence of D2-TGOOH in the lymph would indicate the absorption of ingested TGOOH from the intestine [hypothesis I], while the presence of unlabeled TGOOH would suggest the in situ formation of TGOOH [hypothesis II]). The results obtained from this study will thereby elucidate the absorption of dietary TGOOH in vivo, which provides a foothold to understand the effects of TGOOH on human health.  Chemical structures of TG 18:1/18:1/18:1 and its oxidation products. The hydroperoxyl group (OOH) position and the geometrical (E/Z) structure of TG 18:1/18:1/18:1; OOH depend on the oxidative stress involved (i.e., radicals and/or singlet oxygen). The shorthand notation of lipids was in accordance with the LIPID MAPS nomenclature [55]. For instance, TG 18:1/18:1(9E); 8OOH(sn-2)/18:1 refers to 1, 3-dioleoyl-2-(8-hydroperoxy-9E-octadecamonoenoyl)-TG.

Preparation of TG
HPLC-MS/MS analysis was conducted with an ExionLC HPLC/ UHPLC system (SCIEX, Tokyo, Japan) equipped with a 4000 QTRAP mass spectrometer according to a previously described method with slight modifications [4]. Samples were subjected to chromatographic separation using an analytical column (Inertsil SIL-100A; 5 μm, 2.1 × 250 mm; GL Sciences) attached to a cartridge guard column E (Inertsil SIL-100A; 5 μm, 1.5 × 10 mm; GL Sciences). Hexane-2-propanol-acetic acid (100:0.6:0.5, v/v/v) was used as the mobile phase (flow rate of 0.2 mL/min). To promote ionization, methanol-2-propanol (1:1, v/v) containing 0.2 mM sodium acetate (flow rate of 0.2 mL/min) was mixed with the eluate at the post column using a gradient mixer. The column temperature was set at 40 • C. Ionization conditions (electrospray-ionization: ESI) were set as follows: curtain gas, 20 psi; collision gas, 3 psi; ion spray voltage, 5500 V; temperature, 400 • C; ion source gas 1, 50 psi; ion source gas 2, 70 psi. The concentration of each TGOOH isomer in samples was determined based on external calibration curves obtained from the synthesized standards. Limit of quantification (LOQ) and the minimum concentration of which the coefficient value within ±20% (n=3) were determined according to a validation guideline [18].

Animals
All procedures were performed with protocols approved by the Tohoku University Ethics Review Board (approval number: 2019AgA-024). Male Sprague-Dawley rats (n=7, for experiment I; n=7, for experiment II; aged 8 weeks) weighing 290-320 g were obtained from CLEA Japan, Inc. (Tokyo, Japan). Prior to the experiments, all animals were acclimatized for 1 week. During this period, animals were caged and placed in the animal experimental room with controlled temperature and 12 h light/dark cycle, while given free access to water and commercial rodent chow (CE-2; CLEA Japan, Inc.).
This lipid extract was purified by a Strata® SI-1 Silica cartridge (100 mg, 1 mL, Phenomenex Inc., CA, USA) as follows. The cartridge was washed with methanol (4 mL) and then equilibrated with hexane-2- The extraction recovery rates of TG 18:1/18:1/18:1; OOH from the lymph and the test lipid emulsion were checked as follows. The samples (blank lymph or emulsion) were spiked with 100 pmol of the synthesized TG 18:1/18:1/18:1; OOH standard. The extraction of TG 18:1/ 18:1/18:1; OOH was performed as described above. Extracts were analyzed by HPLC-MS/MS. Extraction recovery rates were calculated by comparison with the control (i.e., blank lymph or emulsion that were not added with the standard) [25]. The results of the recovery test are shown in SI 2. Extraction and analysis were performed as described above. The extraction recovery rates of TG 18:1/18:1/18:1[D2]; OOH from the lymph and the test lipid emulsion were checked as described above. The results are shown in SI 3.

Results and discussion
As previously mentioned, we recently developed a novel and highly sensitive HPLC-MS/MS-based method that enables the detection of TGOOH at the fmol level [4]. On top of that, this method also allows for the selective detection of different molecular species of TGOOH, including isomers with minor structural differences near the OOH group (e.g., OOH positional isomers) that leads to assess the oxidative stress (i. e., radical or 1 O 2 oxidation) [4,11]. Hence, we considered that this method may be a useful tool for obtaining more conclusive data regarding the absorption of dietary TGOOH.
As mentioned in the Introduction, we were able to detect TGOOH in the lymph collected from a small number of rats administered a TG emulsion containing 30 nmol of TGOOH (unpublished preliminary data). To confirm the reproducibility of the experiment, we performed the present study (Experiment I) under similar conditions but with a larger sample size (n=7). To prepare the TG emulsion, we mixed TG 18:1/18:1/18:1 with distilled water containing sodium taurocholate and fatty acid-free bovine serum albumin and sonicated the mixture [19,20]. The resultant emulsion, which was then used as the test sample in Experiment I (Fig. 4A), fortuitously contained 32 nmol of TG 18:1/18:1/18:1; OOH. This may be due to the fact that the commercial TG 18:1/18:1/18:1 standard used to prepare the TG emulsion originally contained a small amount of TG 18:1/18:1/18:1; OOH (data not shown). Also, the heat generated during the sonication might have caused oxidation of TG 18:1/18:1/18:1 [31], to yield an emulsion with the above characteristic (i.e., containing 32 nmol of TG 18:1/18:1/18:1; OOH).
After the lymph-cannulation surgery, lymph was collected from the thoracic duct for 2 h (blank lymph) before the administration of the test emulsion [22]. As shown in Fig. 4B (Fig. 4C). To the best of our knowledge, this is the first study that detected such TGOOH in the lymph after gastric administration of TGOOH mostly owing to our highly sensitive and selective HPLC-MS/MS-based analytical method. In addition, our extraction method allows for the efficient isolation of TGOOH from lymph (extraction recovery >80% (SI 2)), enabling better detection of TGOOH. Fig. 5 shows the time-dependent changes in the concentration of TG 18:1/18:1/18:1; OOH (i.e., TG 18:1/18:1/18:1; 9OOH and TG 18:1/18:1/18:1; 10OOH species) in the lymph. According to this data, the lymphatic concentrations of TGOOH isomers reached their highest at 1-2 h following the administration of TG emulsion. Also, since the emulsion contained a substantial amount of unoxidized TG 18:1/18:1/18:1 whose level in lymph reached its peak at 1-2 h following administration of test emulsion (Fig. 6) [19,20], we deduced that dietary TGOOH can be absorbed from the intestine in the same manner as TG [hypothesis I]. However, considering that only certain TGOOH isomers were present in the lymph and results of previous studies that administered other lipid hydroperoxides to animals [32,33], it is also possible that the ingested TG 18:1/18:1/18:1; OOH triggered  oxidative stress (e.g., generation of free radicals and/or 1 O 2 ) that led to the in situ formation of TG 18:1/18:1/18:1; 9OOH and TG 18:1/18:1/18:1; 10OOH [hypothesis II]. Therefore, we aimed to determine the validity of the above hypotheses using a deuterium-labeled TGOOH (D2-TGOOH) that is traceable in vivo. In this case, the presence of D2-TGOOH in the lymph would indicate the absorption of ingested TGOOH from the intestine [hypothesis I], while the presence of unlabeled TGOOH would suggest the in situ formation of TGOOH [hypothesis II]).
To begin with, we synthesized  (Fig. 8)) [27][28][29][30]. Based on these results, we confirmed that these fragmentations occur near the OOH group by α-cleavage [30] just like the unlabeled TG 18:1/18:1/18:1; OOH and that the deuterium located at the α-carbon of the carbonyl group does not affect the fragmentation patterns. Using a normal-phase HPLC, we were able to separate and selectively detect the twelve  ; OOH isomers. The shorthand notation of lipids was in accordance with the LIPID MAPS nomenclature [55]. T. Takahashi et al. distilled water containing sodium taurocholate and fatty acid-free bovine serum albumin before sonicating the mixture [19,20]. The resultant emulsion, which was then used as the test sample in Experiment II (Fig. 9), contained 87 nmol of TG 18:1/18:1/18:1; OOH and 24 nmol of TG 18:1/18:1/18:1[D2]; OOH. The level of unlabeled TG 18:1/18:1/18:1; OOH in this emulsion was higher than the one used in Experiment I, most likely due to the higher degree of TG 18:1/18:1/18:1 oxidation that occurred during the preparation steps (e.g., during sonication). However, in line with this change, we found that the lymphatic concentration of TG 18:1/18:1/18:1; OOH was also increased compared to the result from Experiment I (c.f., Fig. 10), thus, we believe that the HPLC-MS/MS measurements of TGOOH were accurate in both  experiments. Following the administration of the test emulsion, we clearly detected TG 18:1/18:1/18:1; 9OOH and TG 18:1/18:1/18:1; 10OOH from the lymph (Fig. 9A). Also, the lymphatic concentrations of these isomers reached their highest at 1-2 h following the administration (Fig. 10), which were consistent with our findings in Experiment I. While we barely detected  (Fig. 9B). However, since there is only a mass difference of 2 Da between D2 and unlabeled isotopic pairs, it is likely that most of these signals were derived from the unlabeled TG 18:1/18:1/18:1; OOH isotopes (SI 4) [34]. In fact, when the signals of unlabeled isotope isomers were subtracted [35], the concentrations of all TG 18:1/18:1/18:1[D2]; OOH isomers were under the LOQ (1-10 fmol/injection). Hence, the above results (i.e., the presence of unlabeled TGOOH instead of D2-TGOOH in the lymph) indicated that TGOOH is not absorbed from the intestine but is likely to be produced in situ.
Since we confirmed that TGOOH is not absorbed in its intact form, this suggests the possibility that TGOOH undergoes degradation in the gastrointestinal tract to yield breakdown products that can be absorbed from the intestine. In fact, some previous studies have reported that lipid hydroperoxides are efficiently reduced to hydroxyl species in the gastrointestinal tract [36,37]. Therefore, to confirm whether such phenomenon occurs in the present study, we incubated TGOOH in rat gastric and small intestinal mucosa homogenates and analyzed the formation of the hydroxyl species (TGOH) using HPLC-MS/MS. After 30 min of incubation, we found that the TGOOH level decreased to around 20% (SI 5A) while the level of TGOH significantly increased (SI 5B).
These results suggested that TGOOH decomposes rapidly (e.g., through reductive degradation) in the gastrointestinal environment. More importantly, we clearly detected TGOH in the lymph sample collected in Experiment I (SI 6C). This TGOH may derive from the intestinal absorption of TGOH and/or the in situ formation of TGOH, and thus, we tried to trace the D2-TGOH in the lymph samples collected in Experiment II. Interestingly, we did not detect D2-TGOH in the lymph (SI 6D). Altogether, these data suggest that TGOH is not absorbed from the intestine but is most likely present as the reduction product of the in situ formed TGOOH. The possible mechanism of TGOOH formation is explained in the following paragraph.
As already mentioned, since radical and 1 O 2 oxidation yield characteristically different lipid hydroperoxide isomers [13], we can assess the oxidative stress involved in TGOOH formation based on its isomeric structure [4,11]. In the present study, TG 18:1/18:1/18:1; 9OOH and TG 18:1/18:1/18:1; 10OOH were the predominant isomers detected in the lymph, indicating that 1 O 2 oxidation is likely to be involved in TGOOH formation. This reaction may be triggered by the intake of TGOOH. Previously, dietary lipid hydroperoxides have been reported to cause oxidative stress that induces pro-inflammatory changes in the gastrointestinal tract [32,33,38,39]. Separately, other studies reported that an inflamed intestinal mucosa is highly infiltrated with neutrophils (e.g., in ulcerative colitis patients [40,41]), which generate 1 O 2 via the myeloperoxidase system of activated phagocytosis as the initial defense mechanism against tissue damage [42][43][44]. From these facts, we deduce that the administered TGOOH might trigger oxidative stress by generating 1 O 2 via neutrophil infiltration into the gastrointestinal tract, which subsequently induced 1 O 2 oxidation of TG in situ. To further validate this assertion, we analyzed other TGOOH molecular species such as TG 18:1/18:1/18:2; OOH [4] because the lymph contained a relatively T. Takahashi et al. large amount of unoxidized TG 18:1/18:1/18:2 (15%) (the remaining 85% was TG 18:1/18:1/18:1) (data not shown). As a result, in support of our findings, we found 1 O 2 oxidation-specific isomers (i.e., TG 18:1/18:1/18:2; 10OOH and TG 18:1/18:1/18:2; 12OOH [4]) in the lymph, despite the emulsion containing undetectable levels of these isomers (Fig. 11). These findings may serve as an important basis to formulate effective strategies to prevent and treat various diseases. For instance, considering that continuous accumulation of lipid hydroperoxides in vivo has been linked to the onset and progression of diseases (e. g., atherosclerosis [27], diabetes [45] and Alzheimer's disease [46]), the use of antioxidants (e.g., as supplements and food additives) that specifically inhibit 1 O 2 oxidation, such as carotenoids [47] may be useful for attenuating in situ formation of TGOOH, thus reducing the risk of diseases. To substantiate this, further works are needed to elucidate the association of TGOOH in the development of diseases.
This study has two main limitations. First, we could only trace a small portion of the degradation products that may be formed from the administered TGOOH. Based on the reports by Kanazawa et al. [8,48] and other investigators that proposed the degradation pathway of lipid hydroperoxides [37,49,50], it appears that TGOOH is degraded in the gastrointestinal tract to yield breakdown products such as aldehydes. Since such compounds appear to be transported into the systemic circulation [48,51,52], future studies should broaden the scope of analysis to include other biological samples (e.g., blood and liver). Second, we were unable to perform experiments using a wider range of TGOOH doses that is achievable through daily food intake. The actual dietary intake of TGOOH in humans may be higher than the doses we used in the present study, and thus, future studies should perform under such conditions to strengthen the cause-effect relationship. Such data will help to clarify the mechanisms of the dose-dependent relationship between the intake amount of lipid hydroperoxides and the degree of oxidative stress damage in the intestine which leads to various diseases [53,54]. Nonetheless, as far as the present study goes, we successfully achieved our main goal of elucidating the absorption of TGOOH owing to the use of the isotope-labeling technique and our highly sensitive and selective HPLC-MS/MS-based method.

Conclusions
In this study, we investigated the absorption of TGOOH by performing lymph-cannulation and utilizing our novel HPLC-MS/MS method for the quantification of lymphatic TGOOH levels. As a result, we found that dietary TGOOH is not likely to be absorbed from the intestine. Rather, the presence of TGOOH in lymph is more likely due to the in situ formation of TGOOH, presumably via 1 O 2 oxidation of TG triggered by the administered TGOOH. Altogether, these findings provide a foothold to better understand the effects of TGOOH on human health.

Declaration of competing interest
The authors declare no conflicts of interest.