Collagen hydrolysates (CHs) have been shown to provide multiple health benefits, which have been primarily attributed to their bioactive peptide (BAP) content [1
]. These BAPs can be found in the hydrolysate products, although an increase in the diversity and content of peptides can result from gastrointestinal (GI) digestion [4
]. The BAPs released after the digestion of collagen products, such as Pro-Hyp and Gly-Pro-Hyp, can possess multiple health properties, which include antimicrobial and antihypertensive effects, regulating inflammation, reducing pain associated with osteoarthritis, promoting bone synthesis, stimulating wound healing, as well as antioxidant properties and angiotensin-I-converting enzyme inhibitory effects [3
After digestion, BAPs undergo first pass metabolism, a process defined by hepatic metabolism of compounds following their absorption at the level of the intestinal epithelium that mediates entry into the systemic circulation [8
]. The bioactivity of BAPs depends heavily on their ability to reach the general circulation intact after oral ingestion, otherwise called bioavailability [9
]. Clinical studies have consistently shown that peptides generated from orally ingested collagen precursors, such as gelatin, or collagen hydrolysates, can reach the systemic circulation and be excreted in the urine [4
]. Importantly, the clinical efficacy of CHs has been demonstrated in multiple trials showing reduction of joint discomfort in athletes with functional knee problems and decreased joint pain in osteoarthritis patients [1
]. The BAPs in the bloodstream identified after oral ingestion of CHs and CH precursors, include Ala-Hyp, Pro-Hyp and Gly-Pro-Hyp [4
The assessment of peptide bioavailability using human trials remains costly, lengthy and with limited experimental options for sampling due to ethical restrictions. Instead, animal studies have been used to estimate the bioavailability of BAPs from collagen and collagen precursor products [14
]; however, predictions of bio-absorbability do not always align with human clinical data due to species differences in intestinal permeability and metabolic activity [2
]. Bioavailability studies of food components and pharmaceuticals using animal models have demonstrated poor correlations between rats and humans (r2
= 0.18) as well as dogs and humans (r2
= 0.19) [18
]. Due to such species differences in intestinal permeability and metabolic activity, intestinal cell culture models, rather than animal models, are often used to assess the intestinal transport of food-derived BAPs [2
Caco-2 cells, a human colon carcinoma cell line, has been used regularly to assess for small intestinal (SI) permeability [2
]. Previous work by Feng et al. (2017) [19
] used the Caco-2 model to estimate the transepithelial peptide transport efficiency of bovine CHs. The bioavailability of the CHs, as determined by amino acid (AA) transport, ranged between ~15 and 23%, depending on the hydrolysis method used to generate the CH. Recent work by Song et al. (2020) assessed the bioavailability of BAPs from silver carp skin hydrolysate using in vitro digestion and Caco-2 cells [7
]. They found that, using high-performance liquid chromatography–electrospray ionization tandem mass spectrometry (HPLC-ESI-MS), the transport (%) of Hyp-Gly, Hyp-Gly-Glu and Pro-Gly-Glu-Hyp-Gly was 22.63 ± 5.19, 11.15 ± 0.52 and 18.35 ± 1.20, respectively.
Although in vitro intestinal permeability measures have typically used Caco-2 cells, peptide bioavailability assessments using this cell culture model are not ideal due to the under-expression of peptide transporters such as peptide transporter 1 (PepT1) in these tumorigenic cells. Hence, depending on the compound being assessed, permeability results using Caco-2 cells do not always correlate with human intestinal permeability [18
]. PepT1, otherwise known as SLC15A1, is the main transporter for di- and tri-peptides, which are predominant in CHs and have been indicated to be primarily responsible for the CH-mediated bioactivities [7
]. To overcome the limited PepT1 expression in Caco-2 cells, a non-tumorigenic human small intestinal epithelial cell (HIEC) line can be used. HIEC cells have been shown to be a superior alternative to Caco-2 cells for predicting transporter-mediated absorption of compounds in humans when taken orally [21
]. The HIEC cell model also more accurately represents the physiological in vivo conditions of the SI [22
]. To the best of our knowledge, no study has investigated the transport of CH-derived BAPs using HIEC cells. One study investigating salmon protein hydrolysate peptides and their regulation of oxidative protective genes was investigated using HIEC cells; however, no analysis of peptide bioavailability was completed [25
Methods to accurately quantify di- and tri-peptides to determine their bioavailability have been lacking. Using plasma samples from clinical studies, quantification methods of BAP bioavailability are often calculated using an indirect calculation of Hyp-containing peptides and/or AAs [4
]. Cell culture models also suffer from such limitations in terms of peptide analysis. Feng et al. (2017) assessed the bioavailability of bovine CHs involving Caco-2 cells using an indirect calculation based on the total AAs transported [19
] but peptides were not identified or measured. In the present study, our novel method for targeted BAP quantification using capillary electrophoresis (CE) [26
] was adapted for cell culture media to determine peptide content.
Another limitation to previous in vitro studies investigating BAP bioavailability has been the sole use of intestinal cell cultures without consideration of the subsequent hepatic first pass effects on the intestinally transported BAPs. Some reports have used liver cell culture models, often using human hepatocellular carcinoma (HepG2) cell line, to assess the hepatic metabolism of xenobiotics and drug transporters [8
]. Previous work has also shown that Pro-Gly can increase PepT1 expression in HepG2 cells, although no assessment of the hepatic effects on Pro-Gly was investigated [29
]. Previous studies from our laboratory have assessed the bioavailability of dietary components using a Caco-2/HepG2 co-culture model of first pass metabolism by applying digests from a human simulated gut digestion model [8
]. Similar in vitro models have assessed the oral bioavailability of compounds, such as xenobiotics, and have shown very good correlations with in vivo data from humans and animal models [30
]. In general, there is a major gap in the literature with respect to the study of the hepatic first pass effects on BAPs following their intestinal cell absorption.
In this study, a combination of in vitro gut digestion together with HIEC-6/HepG2-mediated transport and metabolism was used to investigate the bioavailability of BAPs generated after CH digestion. Direct quantification of BAP bioavailability was performed using CE. The aim of this study was to use this novel combination of techniques and cell lines to improve our understanding of the bioavailability and metabolism of CH-derived BAPs that have postulated health promoting properties.
This work was the first to utilize a HIEC-6/HepG2 co-culture to predict the bioavailability of BAPs after the digestion of two CHs using an optimized CE method. This novel combination of cell lines provided further insight into the high degree of BAP transport by utilizing HIEC-6 cells, which more accurately represents the physiological in vivo conditions than previously utilized Caco-2 cells. In terms of the key observations related to di-peptide transport, the Papp
for all the di-peptides measured for both CHs were between 1 and 10 × 10−6
cm/s. Previous work, establishing the relationship between in vitro (Papp
) and in vivo absorption, have ranked compounds as poorly, moderate, or well absorbed to corresponding Papp
]. Poorly absorbed compounds are below 1 × 10−6
cm/s, moderately between 1 and 10 × 10−6
cm/s, and well absorbed compound are above 10 × 10−6
cm/s. Thus, the di-peptides measured in the present study can be considered moderately bioavailable, except for Ala-Hyp after CH-GL treatment, which was 0.7254 ± 0.1947 × 10−6
cm/s. It is possible that the moderate and high degree of bioavailability of collagen-derived BAPs are related to the clinically significant health benefits associated with CH intake.
A relatively high (59%) monolayer transport of Gly-Pro-Hyp with a Papp
value of approximately 9 × 10−6
cm/s was noted after CH-GL treatment. The Papp
of Gly-Pro-Hyp observed with the CH-GL treatment could thus be in the range of a moderately to well absorbed compound. The above Papp
value was much greater than previously reported for Gly-Pro-Hyp by Sontakke et al. (2016), who using Caco-2 cells followed by LC-MS/MS analysis, showed relatively low cumulative amounts of the tri-peptide transported across the monolayer with a Papp
value of 1.09 ± 0.03 × 10−6
]. The Gly-Pro-Hyp peptide exhibits multiple health promoting properties, most notably inhibition of dipeptidylpeptidase-IV (DPP-IV) [39
]. In patients with type 2 diabetes, DPP-IV inhibitors are used to control postprandial glycemia [39
]. Future work is needed assessing the in vivo bioavailability and health modulating properties of this peptide in association with the CH-GL treatment.
In the present work, a markedly lower degree of transport for Pro-Hyp (Papp
= 1.912 ± 0.4794 × 10−6
) as compared to Gly-Pro-Hyp was observed with the CH-GL treatment. Similarly, the apparent permeability reported by Sontakke et al. (2016) for Pro-Hyp (0.13 ± 0.03 × 10−6
cm/s) was significantly lower than their value for Gly-Pro-Hyp [15
]. The Papp
of Pro-Hyp observed in the present study, however, was greater than the values reported by Sontakke et al. (2016) [15
] and Feng et al. (2017) (1.45 ± 0.17 × 10−6
]. As noted by the above, the permeation of Gly-Pro-Hyp was greater than Pro-Hyp, even though Gly-Pro-Hyp is a larger molecular weight peptide. Peptide transport across the intestinal layer via paracellular pathways is primarily dependent on the charge and molecular size of the compound. Since both peptides are uncharged, it is conceivable that active transporters were involved in the relatively greater transport of Gly-Pro-Hyp. Overall, there is a paucity of research pertaining to BAP intestinal transporters, which requires more research using representative physiological models. Pro-Hyp has been shown to decrease the loss of chondrocytes, which synthesize articular cartilage [41
]. In animal models designed to promote cartilage damage, Pro-Hyp inhibited cartilage thinning [41
]. Accordingly, Pro-Hyp is considered to be one of the major bioactive components linked with the clinical efficacy of CHs towards treatment of osteoarthritis.
Our work assessing Hyp-Gly demonstrated transport (%) values of 62.41 ± 11.11 and 82.53 ± 36.53 for CH-GL and CH-OPT, respectively. Song et al. (2020) showed lower transport of Hyp-Gly (22.63 ± 5.19%) from silver carp skin hydrolysate after in vitro digestion and Caco-2 assessment using HPLC-ESI-MS analysis [7
]. The greater degree of transport observed in our study may be attributed to the more physiologically relevant cell culture model used; the under expression of PepT1 in Caco-2 cells could significantly decrease the amount of peptide traveling across the intestinal layer. In contrast, the Papp
values for Hyp-Gly (6.740 ± 1.200 × 10−6
after CH-GL and 5.593 ± 2.476 × 10−6
after CH-OPT) were lower compared to Song et al. (2020), which was 10.00 × 10−6
]. Apart from the different intestinal cell types used, variances in the quality of the established monolayer due to differences in passage number, cell conditions, and culture duration could impact the intestinal transport coefficients [42
]. The high bioavailability of Hyp-Gly in the present work coincides with in vivo studies showing that this antiplatelet peptide is present in blood after CH ingestion and thereby could provide anti-thrombotic protection [7
Although there were no differences in di-peptide bioavailability between the two tested CHs, CH-GL showed significant Gly-Pro-Hyp content after first pass liver metabolism, whereas none was observed after CH-OPT. This difference in bioavailability could be attributed to the presence of other peptides found within the CHs, as the digestion and bioavailability of BAPs can be affected by the presence of other peptides, proteins, or food components [2
]. Increased peptide absorption could also occur due to synergisms with other peptides present in the digests as dietary AAs and protein hydrolysates can increase PepT1 expression [2
]. Previous work by our group has established that CH-GL and CH-OPT have different peptide profiles, both pre- and post-digestion, with some peptide sequences being found in one CH and not the other [5
]. The synergistic effects of BAPs are still under investigation; however, hormonal responses can be influenced by the presence of other proteins or peptides consumed. For example, the glucose-dependent insulinotropic polypeptide response and gastric emptying were greater when milk protein hydrolysates were ingested compared to whole milk protein sources [2
]. Furthermore, colonic motility contractions were increased after whey hydrolysates compared to whey protein concentrates [2
]. Further work on identifying and understanding synergistic effects affecting peptide transport, bioavailability and bioactivity, is required, particularly for CH-derived BAPs.
To our knowledge, the present study has been the first to determine the impact of hepatic first pass effects on BAPs after their intestinal transport. A direct and targeted method of BAPs quantification using CE allowed for an in-depth analysis of BAP content following their first pass effects. The presence of HepG2 cells in the basolateral compartment could potentially have affected permeability assessments, as previous work reporting Papp
has used only intestinal cell monolayers. The effect of HepG2 cells in a co-culture on Papp
has not been fully established. Some preliminary reports have demonstrated that the presence of Pro-Gly increases PepT1 expression in HepG2 cells [29
], although further work is needed assessing peptide transport as affected by modulation of PepT1 expression by di-peptides. The use of a co-culture of intestinal and hepatic cell lines has been well established to understand bioavailability (%), although assessments of Papp
were not reported [8
]. Future work to incorporate hepatic effects on peptide transport should be investigated, especially considering that the expression of PepT1 may be regulated by the presence of BAPs [29
The hepatic first pass effects on BAPs have not been well studied. Most published work discussed above investigating “bioavailability” only used Caco-2 cells thereby determining intestinal transport only, but this does not represent systemic availability. The degree that hepatic first pass effects affected peptide content in this study was unexpected; however, such studies investigating BAPs have not been previously performed. In that regard, it has been well established that there is high hepatic metabolism for small peptides [44
], but hepatic upregulation of BAPs has not been studied previously. The importance of assessing the contribution of hepatic action is clearly demonstrated in our work. For example, Ala-Hyp was increased after incubating with HepG2 cells up to 304.9 ± 57.2% after treatment with CH-GL digests. Although both CHs were derived from bovine collagen, there was a significant difference in the hepatic first pass effects on Pro-Hyp. Hepatic action on Pro-Hyp was greater after CH-GL treatment (151.4 ± 24.3%) compared to CH-OPT (63.63 ± 8.63%); this was surprising as the content of Pro-Hyp that traversed across the intestinal layer was not significantly different between the treatments. The difference in hepatic first pass effects on Pro-Hyp might be due to the presence of Gly-Pro-Hyp that was solely noted to be intestinally transported after CH-GL treatment; this tri-peptide could conceivably be metabolized further by hepatic cells to contribute to the Pro-Hyp content. Such hepatic production of Pro-Hyp would not be expected with CH-OPT as Gly-Pro-Hyp was not appreciably transported across the intestinal layer with this treatment. The increase in BAP production for all the di-peptides during hepatic action could also have occurred due to the metabolism of unidentified longer chain peptides that travelled across the epithelium. In that respect, further work into identifying and assessing other collagen-derived BAPs is needed.
No previous studies have combined simulated digestion together with HIEC-6/HepG2-mediated transport and metabolism to investigate the bioavailability of CH-derived BAPs. A notable finding was that Gly-Pro-Hyp had a 12.24 ± 1.12% bioavailability with the CH-GL treatment after intestinal transport and hepatic first pass effects. A possible comparison might be made with the in vivo studies by Skov et al. (2019), which determined the postprandial plasma concentration of Gly-Pro-Hyp in a human clinical trial using 1
H NMR analysis [4
]. The initial Gly-Pro-Hyp content in the plasma was ~ 400 µM, and the Gly-Pro-Hyp content increased after 2 h to ~ 1050 µM, which would represent a 162.5% increase. It should be noted, however, that the method by which plasma Gly-Pro-Hyp was calculated by Skov et al. (2019), involved summing the individual AA measurements of Gly, Pro and Hyp, as no peptide sequencing or targeted quantification of Gly-Pro-Hyp was done. As digestion breaks down peptides into their AA components, it is possible that the summed plasma content of Gly, Pro, and Hyp indicated a greater apparent bioavailability of Gly-Pro-Hyp than provided via direct measurement of the tri-peptide.
To further understand the bioactivity of specific BAPs, rapid, accurate and efficient methods of identification and quantification are necessary. Previous work assessing CH-derived peptide bioavailability using Caco-2 cells have had significant limitations in terms of endpoint analysis. Feng et al. (2017) [19
] assessed bovine CH bioavailability according to an indirect calculation of total AA transported. Furthermore, no peptide sequencing using proteomics methods or quantification was done. Three major AAs found in collagen are Gly, Pro and Hyp, but no Pro content was detected for all the hydrolysates assessed [19
]; therefore, established BAPs sequences such as Pro-Hyp, Gly-Pro-Hyp, Gly-Pro, were likely not found. Future studies can utilize emerging technologies such as the CE methodology described herein towards the identification and quantitation of BAPs.
Despite their limitations, cell culture models continue to provide a platform to predict the bioavailability of BAPs, as animal studies often to do not correlate with human data, and human trials are long, associated with increased costs and have ethical restrictions [2
]. Comparisons of cell culture models to human in vivo data generally support the use of the former to assess intestinal transport [22
]. Discrepancies involving in vitro assessments of kinetics and peptide activity may occur, however, if the digestive and metabolic processes are not sufficiently considered [2
]. Cell culture models must therefore accurately replicate the digestion, transport, and metabolism of the bioactive components of interest. For this reason, in this study, the bioavailability of CH-derived BAPs after in vitro digestion was determined using a novel co-culture of HIEC-6/HepG2 cells rather than a Caco-2 monolayer, as the expression of a key peptide transporter PepT1 is under-expressed in Caco-2 cells and predictions of peptide bioavailability could be misleading. Previous work has confirmed that HIEC cells more accurately represent the physiological in vivo conditions of the SI compared to Caco-2 cells [22
]. Further studies can adopt and standardize this HIEC-6/HepG2 co-culture method, which could be adapted to investigate the first pass effects of bioactive food components, nutraceuticals and supplements.
As demonstrated in this study, similarly sourced and marketed CH products can contain different peptide profiles [5
] and have varying degrees of peptide bioavailability. These findings are pertinent since BAPs must undergo first pass metabolism [9
] for CHs and collagen-derived peptides to exert their bioactivity, such as on joint tissues including bone, cartilage and muscle. The bioavailability of collagen BAPs has been related to the clinically significant health benefits associated with CH intake, such as decreasing pain associated with OA, improving joint discomfort, and increasing bone mineral density [1
]. Therefore, the different degree of BAP bioavailability seen after hepatic first pass effects between the CH products could modify their clinical efficacy. As consumers continue to use an increasing variety of over-the-counter CHs, assessing the bioavailability and bioactivity of BAPs from various CHs using higher throughput models is advantageous. This model provides a higher throughput method to assess peptide bioavailability before clinical studies are undertaken, which are often costly, long and have various ethical constraints.