**2. Results**

#### *2.1. Physicochemical Properties of Fucoidans from Various Sources*

The basic physicochemical properties of AnF, LjF, and KcF were determined and are summarized in Table 1. The sulfate content ranged from 22% to 28%, and the sulfate content of AnF was lower compared with that of LjF or KcF. The MWs of AnF and LjF were similar, which were both much lower than that of KcF (*p* < 0.05). According to our monosaccharide composition analysis, these three fucoidans were mainly composed of fucose, and the contents of fucose were similar in AnF and LjF, while KcF had a much higher fucose content (*p* < 0.05). The above results were consistent with that reported in previous literatures [32–34]. In general, the main differences between these three fucoidans were the type of glycosidic linkages, MWs, and fucose content. The bioactivities of fucoidans with various physicochemical properties can be different. Therefore, we further investigated the effects of these three fucoidans on alleviating postprandial hyperglycemia in vitro and in vivo.

**Table 1.** Compositions of fucoidans from various sources.


AnF: fucoidan from *Ascophyllum nodosum*; LjF: fucoidan from *Laminaria japonica*; KcF: fucoidan from *Kjellmaniella crassifolia*; FvF: fucoidan from *Fucus vesiculosus*; MW, molecular weight; Man, mannose; GlcN, glucosamine; Rha, rhamnose; GlcA, glucuronic acid; Glc, glucose; Gal, galactose; Xyl, xylose; Fuc, fucose.

#### *2.2. E*ff*ect of Fucoidans on Inhibiting Glucose Transport in a Caco-2 Monolayer Cell Model*

Here, we used a Caco-2 monolayer cell model to evaluate the influences of various fucoidans on intestinal glucose transport. The Caco-2 monolayer cell model is a well-established in vitro model for studying the transport of substrates (e.g., glucose, nutrients, and drugs) through the intestine [35]. The prerequisite for the simulation of in vivo intestinal processes is the differentiation of the Caco-2 monolayer cell, which expresses a tissue-typical cell membrane and transport proteins. In addition, it has been verified that there is a high expression of endogenous SGLT1 in polarized Caco-2 cells [36]. As shown in Figure 1A, the transepithelial electrical resistance (TEER) value rose rapidly from 49 Ω\*cm<sup>2</sup> on the fourth day to about 476 Ω\*cm<sup>2</sup> after a 16-day incubation. Then, this cell monolayer model tended to be completed, and the resistance value changed little on the 21st day compared to that on the 16th day, which indicated that the cells started to differentiate. Moreover, the activities of alkaline phosphatase (ALP) on both sides of the transwell chamber were measured to judge the degree of cell differentiation and the success of the cell model [37]. As shown in Figure 1B, the ratio of ALP activities significantly increased to about 8 during the 21 days incubation, which indicated that the degree of differentiation in Caco-2 cells was high, and the cells displayed obvious polarization. The above results

indicated that a Caco-2 monolayer cell model was successfully constructed, which was then used to study the e ffects of fucoidans on the transmembrane transport of glucose. The inhibitory activity of specific fucoidan on the transport of 2-Deoxy-2-[(7-nitro-2,1,3-benzoxadiazol-4-yl) amino]-D-glucose (2-NBDG) using the Caco-2 monolayer model is shown in Figure 1C. These results showed that only 400 μg/mL of an AnF solution could significantly inhibit the transport of 2-NBDG in this Caco-2 monolayer cell model compared with the Control group (*p* < 0.05), while the same concentration of LjF and KcF could not. Compared to the e ffects of AnF and LjF on 2-NBDG transport, our results indicate that the type of glycosidic linkages may play a crucial role in inhibiting glucose transport. In addition, both LjF and KcF had no marked e ffects on glucose transport in the Caco-2 monolayer cell model, which indicated that MW and fucose content may not play a pivotal role in that. We have previously reported that FvF could significantly reduce glucose transport in a Caco-2 monolayer cell mode [30]. In general, only AnF and FvF, with type II structures, could inhibit the glucose transport in a Caco-2 monolayer cell model, which indicated the importance of the type of glycosidic linkages in inhibiting glucose transport.

**Figure 1.** Inhibitory e ffects of fucoidans on 2-Deoxy-2-[(7-nitro-2,1,3-benzoxadiazol-4-yl) amino]-D-glucose (2-NBDG) transport in a Caco-2 monolayer cell model. The transepithelial electrical resistance (TEER) of Caco-2 monolayer cell model ( **A**); alkaline phosphatase (ALP) activity ratio of Caco-2 monolayer cell model (**B**), calculated by the ratio of ALP activities on the apical side to the basolateral side; inhibitory e ffects of fucoidans on 2-NBDG transport ( **C**). Control, Caco-2 monolayer cell model treated with HBSS bu ffer; *Ascophyllum nodosum* (AnF), Caco-2 monolayer cell model treated with 400 μg/mL of AnF; or *Laminaria japonica* (LjF): Caco-2 monolayer cell model treated with 400 μg/mL of LjF; *Kjellmaniella crassifolia* (KcF): Caco-2 monolayer cell model treated with 400 μg/mL of KcF. Data are expressed as the mean ± SEM. \* *p* < 0.05, compared with the Control group.

#### *2.3. E*ff*ect of Fucoidans on Intestinal Glucose Uptake Using an Everted Gut Sac Model*

Glucose can be transported from the intestinal lumen into small intestinal enterocytes by SGLT1, which is located at the apical brush border [38]. SGLT1 was originally expressed on the serous side of the intestine and was moved to the outside in an everted gu<sup>t</sup> sac model. Thus, the everted gu<sup>t</sup> sac model can be used as an e fficient semi-in vivo tool for studying substrates and drug absorption mechanisms, as well as the role of compounds on regulating SGLT1 activity in the intestine [39,40]. Therefore, the inhibition of glucose absorption was evaluated using the everted gu<sup>t</sup> sac model to explore the inhibitory effect on SGLT1 activity. As shown in Figure 2, 100 μg/mL of AnF could effectively decrease the glucose absorption compared with the Control group (*p* < 0.05), and this inhibitory effect was elevated with an increased AnF concentration, while LjF and KcF showed no significant inhibitory effect in the range from 100 μg/mL to 600 μg/mL. All in all, the above results demonstrated that only AnF with a type II structure had a marked inhibition on glucose intake in a dose-dependent manner via inhibiting SGLT1 activity, which was consistent with the results in Section 2.2. Once again, the type of glycosidic linkages was shown to play an important role in suppressing glucose absorption from the intestinal lumen.

**Figure 2.** Inhibition rate of different concentrations of fucoidans on glucose uptake in an everted gu<sup>t</sup> sac model. Control, the everted gu<sup>t</sup> sac model treated with Krebs–Ringer bicarbonate buffer; AnF, the everted gu<sup>t</sup> sac model treated with various concentrations of AnF; LjF, the everted gu<sup>t</sup> sac model treated with various concentrations of LjF; KcF, the everted gu<sup>t</sup> sac model treated with various concentrations of KcF. Red represents the results of treatment with 100 μg/mL of specific fucoidan; green represents the results of treatment with 400 μg/mL of specific fucoidan; black represents the result of treatment with 600 μg/mL of fucoidan. Data are expressed as the mean ± SEM. \* *p* < 0.05, specific concentration of AnF treated group compared with the Control group.

#### *2.4. E*ff*ects of Fucoidans on OGTT in Kunming Mice*

Intestinal glucose absorption is mediated by SGLT1 [41]. First, the carbohydrates in food are degraded to monosaccharides (such as glucose, galactose, etc.) through the action of various glycosidases. Next, glucose is then transported into the cells on the mucosal side by SGLT1. In addition, it has been confirmed that fucoidans have a certain inhibitory effect on α-glucosidase [42] and cannot be digested by gastric and pancreatic enzymes [43]. Therefore, the OGTT was used to explore the inhibitory effects of fucoidans on glucose absorption via inhibiting SGLT1 activity, which can avoid the influence of glucosidase. As shown in Figure 3, the administration of 200 mg/kg of AnF effectively suppressed the elevation in the postprandial blood glucose level and the areas under curve after glucose loading in Kunming mice compared with the Control group (*p* < 0.05), while LjF and KcF treatments could not. The above results verified the efficient decrease in postprandial blood glucose conferred by AnF treatment, as compared to LjF and KcF in Kunming mice, which was consistent with our results from in vitro and semi-in vivo assays (shown in Sections 2.2 and 2.3). Thus, we further investigated and evaluated the hypoglycemic effect of AnF in db/db mice.

**Figure 3.** Effects of fucoidans on the oral glucose tolerance test (OGTT) in Kunming mice. Curve of OGTT (**A**) and area under curve (**B**). Control, Kunming mice gavaged with PBS; AnF, Kunming mice gavaged with 200 mg/kg of AnF; LjF: Kunming mice gavaged with 200 mg/kg of LjF; KcF: Kunming mice gavaged with 200 mg/kg of KcF. Data are expressed as the mean ± SEM. \* *p* < 0.05, compared with the Control group. Six mice of each group were analyzed.

#### *2.5. E*ff*ects of AnF on the OGTT in db*/*db Mice*

OGTT is the gold standard for DM diagnosis [44] and is used to evaluate the function of β cells and the individual ability to regulate postprandial blood glucose levels. Thus, a glucose solution was gavaged after 15 h of fasting at the end of a 4-week feeding trial in db/db mice. The AnF group showed remarkable suppression of OGTT and the area under curve in db/db mice (*p* < 0.05), which was comparable to the effects of metformin (Metf) (Figure 4). This result indicated that AnF can improve blood glucose homeostasis in mice with diabetes.

**Figure 4.** Effects of AnF on OGTT in leptin receptor-deficient (db/db) mice. Curve of OGTT (**A**) and area under curve (**B**). Control, C57BL/6J mice; Model, db/db mice; metformin (Metf), db/db mice with 200 mg/kg/d of metformin; AnF: db/db mice with 200 mg/kg/d of AnF. Data are expressed as the mean ± SEM. \* *p* < 0.05, compared with the Control group; # *p* < 0.05, compared with the Model group. Six mice of each group were analyzed.

#### *2.6. E*ff*ects of AnF on Body Weight in db*/*db Mice*

During the 4-weeks feeding trial, db/db mice gained much more weight compared to normal C57BL/6J mice (*p* < 0.05) (Figure 5), while both AnF and metformin induced a trend of weight loss in db/db mice. In addition, another study in our lab showed that AnF significantly reduced the body weight gain of mice, which were fed with a high-fat diet [28]. The above results indicated that AnF could effectively lower the body weight in mice with DM.

**Figure 5.** Effect of AnF on body weight change in db/db mice. Body weight change (**A**) and body weight gain (**B**). Control, C57BL/6J mice; Model, db/db mice; Metf, db/db mice with 200 mg/kg/d of metformin; AnF, db/db with 200 mg/kg/d AnF. Data are expressed as the mean ± SEM. \* *p* < 0.05, compared with the Control group. Six mice of each group were analyzed.

.

#### *2.7. E*ff*ects of AnF on Glucose-Insulin Homeostasis in db*/*db Mice*

DM is characterized by hyperglycemia and systemic insulin resistance [45]. Thus, we investigated the effect of AnF on glucose-insulin homeostasis in db/db mice. As shown in Figure 6A, AnF significantly alleviated fasting hyperglycemia compared with the Model group (*p* < 0.05). The administration of AnF as well as metformin, resulted in an effective decrease in hemoglobin A1c (HbA1c) level (*p* < 0.05) compared with the Model group (Figure 6B), which indicated that AnF had a long-term effect on alleviating hyperglycemia. As shown in Figure 6C,D, the results showed the significant effect of AnF on suppressing hyperinsulinemia and lowering the homeostasis model assessment-insulin resistance (HOMA-IR) index in db/db mice (*p* < 0.05). All of these analyses confirmed the glucose–insulin homeostasis effects of AnF in db/db mice.

**Figure 6.** Effects of AnF on glucose-insulin homeostasis in db/db mice. Fasting blood glucose (**A**); hemoglobin A1c (HbA1c) (**B**); Fasting insulin (**C**); homeostasis model assessment-insulin resistance (HOMAI-IR) (**D**). Control, C57BL/6J mice; Model, db/db mice; Metf, db/db mice with 200 mg/kg/d of metformin; AnF, db/db with 200 mg/kg/d AnF. Data are expressed as the mean ± SEM. \* *p* < 0.05, compared with the Control group; # *p* < 0.05, compared with the Model group. Six mice of each group were analyzed.

#### *2.8. E*ff*ects of AnF on GLP-1 Level in db*/*db Mice*

Increasingly, studies have demonstrated that SGLT1 in the intestine serves as a sensor for the acute glucose-induced GLP-1 secretion [46,47]. The short-term inhibition of SGLT1 in the intestine can delay the absorption of glucose, and this unabsorbed glucose can stimulate L cells to secrete GLP-1 [48,49]. GLP-1 can act on pancreatic β cells to increase insulin release in a glucose-dependent manner and decrease pancreatic glucagon secretion, which both contribute to the antihyperglycemic effect [50]. Studies have indicated that the inhibition of SGLT1 in the intestine can not only increase the content of serum active GLP-1 (aGLP-1) level but also significantly increase the serum total GLP-1 (tGLP-1) level [51,52]. The effects of AnF on inhibiting postprandial blood glucose level by in vitro (Caco-2 monolayer model), semi-in vivo (everted gu<sup>t</sup> sac model), and in vivo assays indicate that AnF can inhibit SGLT1 activity. Thus, we investigated the effect of AnF on the level of GLP-1 in db/db mice further. As shown in Figure 7, AnF effectively increased the serum content of tGLP-1 and aGLP-1 compared with the Model group, but this was still lower compared to the normal levels (*p* < 0.05), which indicated the moderate inhibition of SGLT1 by AnF can cause GLP-1 release without side effects on the digestive system.

**Figure 7.** Effects of AnF on serum total glucagon-like peptide-1 (tGLP-1) and active GLP-1 (aGLP-1) levels in db/db mice. tGLP-1 (**A**) and aGLP-1 (**B**) levels. Control, C57BL/6J mice; Model, db/db mice; Metf, db/db mice with 200 mg/kg/d of metformin; AnF, db/db with 200 mg/kg/d AnF. Data are expressed as the mean ± SEM. \* *p* < 0.05, compared with the Control group; # *p* < 0.05, compared with the Model group; and & *p* < 0.05, compared with the Metf group. Six mice of each group were analyzed.

#### *2.9. Interaction Study between Fucoidans and SGLT1*

Our in vitro and in vivo assays demonstrated that AnF could effectively decrease the glucose transport and postprandial blood glucose levels, while LjF and KcF could not, which may have been due to the various inhibitions of SGLT1 activity by fucoidans. Thus, we validated the binding affinities between these three fucoidans (AnF, LjF, and KcF) and SGLT1 protein further using SPR. The response units were recorded in real-time as sensor grams using the BIAcore system (Figure 8), and the kinetic parameters (such as the binding constant (*K*a, <sup>M</sup>−1s−1), dissociation constant (*K*d, 1/s), and average dissociation constant (*K*D, M)) were also summarized. These results showed that AnF and FvF with type II structures bound directly to SGLT1, and the *K*D was 3.4 × 10−<sup>6</sup> M and 9.696 × 10−<sup>6</sup> M, respectively. In contrast, no significant affinity was detected between KcF/LjF and SGLT1. Therefore, we concluded that the inhibitory effect of AnF on SGLT1 was due to the strong binding between them, which could effectively block the glucose transport capacity of SGLT1. As expected, AnF and FvF, with type II structures, bound directly to SGLT1, while KcF and LjF, with type I structures, could not, which was consistent with the effects of these respective molecules on glucose transport and postprandial blood glucose levels in vitro and in vivo assays and our previous study [30]. As SGLT1 is a crucial factor for regulating glucose transport and postprandial blood glucose levels, these data indicated that SGLT1 was probably a potential target for fucoidans with type II structure to exert the hypoglycemic effects.

**Figure 8.** Surface plasmon resonance (SPR) kinetic binding analysis of the interactions of Na+/glucose cotransporter 1 (SGLT1) with AnF (**A**), FvF (**B**), KcF (**C**), and LjF (**D**). ND, not detected.
