**Use of an Extract of** *Annona muricata* **Linn to Prevent High-Fat Diet Induced Metabolic Disorders in C57BL**/**6 Mice** †

**Sandramara Sasso <sup>1</sup> , Priscilla Cristovam Sampaio e Souza 2, Lidiani Figueiredo Santana 1, Claudia Andréa Lima Cardoso 3, Flávio Macedo Alves 4, Luciane Candeloro Portugal 4, Bernardo Bacelar de Faria 5, Anderson Fernandes da Silva 1, Ana Rita Coimbra Motta-Castro 6,7, Luana Silva Soares 6, Larissa Melo Bandeira 6, Rita de Cássia Avellaneda Guimarães <sup>1</sup> and Karine de Cássia Freitas 1,\***


Received: 4 May 2019; Accepted: 30 June 2019; Published: 2 July 2019

**Abstract:** *Annona muricata* Linn, commonly known as graviola, is one of the most popular plants used in Brazil for weight loss. The aim of this study is to evaluate the therapeutic effects of three different doses (50 mg/kg, 100 mg/kg, and 150 mg/kg) of aqueous graviola leaf extract (AGE) supplemented by oral gavage, on obese C57BL/6 mice. Food intake, body weight, an oral glucose tolerance test (OGTT), an insulin sensitivity test, quantification of adipose tissue cytokines, weight of fat pads, and serum biochemical and histological analyses of the liver, pancreas, and epididymal adipose tissue were measured. AGE had an anti-inflammatory effect by increasing IL-10 at doses of 50 and 100 mg/kg. Regarding the cholesterol profile, there was a significant decrease in LDL-cholesterol levels in the AGE 150 group, and VLDL-cholesterol and triglycerides in the AGE 100 and 150 groups. There was an increase in HDL cholesterol in the AGE 150 group. The extract was able to reduce the adipocyte area of the epididymal adipose tissue in the AGE 100 and 150 groups. According to the histological analysis of the liver and pancreas, no significant difference was found among the groups. There were no significant effects of AGE on OGTT and serum fasting glucose concentration. However, the extract was effective in improving glucose tolerance in the AGE 150 group.

**Keywords:** graviola; weight loss; obesity

#### **1. Introduction**

Obesity is a worldwide public health problem. It increases the risk of metabolic diseases such as hypercholesterolemia, hypertriglyceridemia, insulin resistance, heart disease, type 2 diabetes, atherosclerosis, and cancer [1,2].

The etiology of obesity is complex and multifactorial. Obesity results from the interaction of genetic/epigenetic, environmental, emotional, lifestyle factors and that technically obesity results from a positive energy balance: More energy intake than energy expenditure [3]. Although genetic factors are determinant in the development of obesity, metabolic factors, an unhealthy diet, and a sedentary lifestyle provide conditions for the development of this disorder [4].

Obesity is an increased deposition of white adipose tissue and phenotypic changes in this tissue. It is associated with metabolic changes such as increased production of pro-inflammatory mediators. This leads to organs dysfunction and chronic low-grade inflammation with high levels of proinflammatory cytokines, such as interleukin-6 (IL-6) and tumor necrosis factor-α (TNF-α), and chemokines, such as monocyte chemotactic protein 1 (MCP1), which in turn promotes migration of macrophages into the adipose tissue and increases the release of cytokines. In parallel, low levels of interleukin 10 (IL-10) are observed in obese individuals, and worsens their metabolic profile, since IL-10 inhibits the synthesis of pro-inflammatory cytokines [2,5–8].

To control such abnormalities, several methods have been suggested to regulate obesity and weight gain, including agents that could inhibit fat absorption, control biochemical parameters, such as, serum glucose, serum triglyceride, total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), and very low-density lipoprotein (VLDL) levels, reduce systemic inflammation, and induce weight loss [9]. Medicinal plants, popularly indicated for the treatment of obesity, have been used in many countries to control weight gain and obesity [10].

For these reasons, anti-obesity agents, including infusions and extracts, are widely used for weight control in obese individuals, in addition to reducing biochemical parameters [11].

Tropical countries have a wide variety of flora and a high number of food and medicinal plants. There is information available on the potential functional properties of several of these plants [12]. The investigation of such properties may be of interest for both the pharmaceutical and the food industry [13].

Among the species of pharmaceutical interest, the leaf of *Annona muricata* Linn (Annonaceae), commonly known as soursop or graviola, is used routinely for weight control. It is used in traditional medicine as an antihypertensive, vasodilator, antidiabetic, and hypolipidemic agent due to the presence of several bioactive compounds, such as acetogenins, flavonoids, tannins, alkaloids, coumarins, and terpenoids [14,15]. Therefore, considering the popular use of tea from graviola leaves to prevent obesity and its complications, it is important to verify whether treatment using an aqueous extract of *Annona muricata* Linn could also be beneficial for the treatment of obesity. Thus, the objective of this study is to verify the effects of three different doses of aqueous extract of *Annona muricata* on obese C57BL/6 mice induced by a high-fat diet.

#### **2. Materials and Methods**

#### *2.1. Extraction of Plant Material*

Leaves of *Annona muricata* Linn were collected in June 2015 from an adult specimen that produces flowers and fruits, in the municipality of Campo Grande, Mato Grosso do Sul state, Brazil. The tree was properly identified. The geographical coordinates defined by manual GPS were 22◦29 42.6" S and 054◦37 1.6" W. A voucher specimen (number 53,928) was deposited at the Herbarium CGMS of the Federal University of Mato Grosso do Sul, Brazil. The extract of leaves of *Annona muricata* Linn was prepared by immersing 1 kg of leaf powder into 3 L of distilled water for 48 h, then lyophilizing this until a dry powder was obtained. Then, the extract was stored at room temperature and protected from light until use [14].

#### *2.2. Quantification of Total Phenols and Flavonoids*

The total phenols of aqueous graviola leaf extract (AGE) were determined by the Folin-Ciocalteu reagent method [16]. Samples and a standard curve of gallic acid were read at 760 nm. The result was expressed as mg of gallic acid per g of extract. For the quantification of the flavonoids, the colorimetric method of aluminum chloride was used [17]. The absorbances were read at 415 nm with a UV-Vis spectrophotometer. To calculate the concentration of flavonoids, an analytical curve was prepared using quercetin as standard. The results are expressed as mg quercetin per g of extract.

#### *2.3. Quantification of Condensed Tannins*

The extract was dissolved in water at a concentration of 50 <sup>μ</sup>g·mL−<sup>1</sup> using the valinine reaction [18]. The absorbance reading was performed using a spectrophotometer at 510 nm. The quantification was performed using an external calibration curve with catechin as standard. The results are expressed as mg equivalent of catechin per g of extract.

#### *2.4. Assay of Antioxidant Activity Using the 2,2-Diphenyl-1-Picrylhydrazyl Free Radical (DPPH)*

The sequestering capacity was measured using DPPH solution. The absorbances were read at 517 nm with a spectrophotometer. The percentage of DPPH radical sequestration inhibition was calculated according to the equation:

$$\text{Percent inhibition activity (\%)} = \left[ (\text{A0} - \text{A1}) / \text{A0} \right] 100 \tag{1}$$

where *A*<sup>0</sup> is the absorbance of the control, and *A*<sup>1</sup> is the absorbance in the presence of the compound. Subsequently, the mean inhibitory concentration (IC 50) was calculated. It represents the concentration of the sample required to capture 50% of the DPPH [19].

#### *2.5. Isolation and Identification of Compounds*

The extract was fractionated by XAD-2 (Supelco, Bellefonte, PA, USA) on column chromatography (30 cm × 3 cm). The extract (3.16 g) was eluted with 0.5 L of water, followed by 0.5 L of methanol, and again eluted with 0.2 L of ethyl acetate. An aliquot of 0.89 g of the methanolic fraction was dissolved into 50 mL of methanol and fractionated by chromatography using a Sephadex LH-20 (Amersham Pharmacia Biotech, Uppsala, Sweden) on column chromatography (70 cm × 3 cm) at a rate of 0.2 mL·min−1. Twenty-five fractions of 2 mL were collected. The fractions were combined according to their chemical behavior on thin layer chromatography (silica gel plates) using ethyl acetate:n-propanol:water (123:7:70 *v*/*v*/*v*) as the eluent. The fractions 2–4, 6–8 and 11–14 were purified using polyvinylpolypyrrolidone (Sigma, St. Louis, MO, USA) on column chromatography (10 cm × 2 cm) by eluting them with methanol. The result is the identification of compounds. An aliquot of 0.54 g of the ethyl acetate fraction was dissolved into 10 mL of methanol and fractionated by chromatography using a Sephadex LH-20 (Amersham Pharmacia Biotech, Uppsala, Sweden) on column chromatography (80 cm <sup>×</sup> 2 cm) by eluting it with methanol at a rate of 0.3 mL·min<sup>−</sup>1. Twenty-eight fractions of 5 mL were collected. The fractions were combined according to their chemical behavior on thin layer chromatography (silica gel plates) using ethyl acetate:methanol (60:40 *v*/*v*) as the eluent. The fractions 10–13, 18–19 and 22–25 resulted in the isolation of the other compounds. The identification of the compounds was carried out using 1H and 13C nuclear resonance (Advance 300 MHz, Brucher, Ettlingen, Germany) and mass spectrometry (Shimadzu Corp. Shimadzu, Kyoto). Their chemical structures were confirmed by comparison with literature data [20–22].

#### *2.6. Ethics Statement*

All animal experiments were submitted and approved by the Ethics Committee on Animal Use, Federal University of Mato Grosso do Sul (Protocol no. 682/2015).

#### *2.7. Acute Oral Toxicity*

The acute toxicity test of the AGE was performed in female Wistar rats (*Rattus norvegicus*) based on the OECD Guidelines 425 (Organization for Economic Co-operation and Development) [23]. For the test, the animals were divided into two groups (*n* = 5): A control group that received saline solution, and the treatment group that received the aqueous extract of *Annona muricata* Linn orally (gavage) at a dose of 2000 mg/kg. After treatment, the animals were observed at 30 min, 1 h, 2 h, 3 h, 4 h, 6 h, 12 h, 24 h, and then daily for 14 days.

At the same time, the hippocratic screening test was carried out to quantify the effects of abnormal morphological and behavioral signs of toxicity. Furthermore, changes in body weight, water and food intake, as well as excreta production, were also evaluated [24].

At the end of 14 days, the animals were euthanized (ketamine and xylazine). The organs (heart, lung, liver, spleen, pancreas, and kidneys) were removed, weighed, and analyzed macroscopically to investigate possible changes [25].

#### *2.8. Animals and Experimental Design*

C57BL/6 adult male mice (*n* = 55, 12 weeks of age) were divided into two groups based on body weight, as follows: SHAM group (*n* = 11), treated with standard diet AIN-93M [26], and HFD group (*n* = 44), treated with a hyperlipidic diet. After 12 weeks, the animals of the HFD group were divided into four homogenous groups according to weight and value of fasting blood glucose and concomitantly supplemented (oral gavage) with aqueous graviola leaf extract in different doses: HFD SALINE group (HFD + saline), AGE 50 mg/kg group (HFD + aqueous graviola leaf extract of 50 mg/kg) (*n* = 11), AGE 100 mg/kg group (HFD + aqueous graviola leaf extract of 100 mg/kg) (*n* = 11), and AGE 150 mg/kg group (HFD + aqueous graviola leaf extract of 150 mg/kg) (*n* = 11). The SHAM group also received saline solution at this stage of the study. Each group had ad libitum access to water and food during the experimental period. The composition of the experimental diets is show in the Table 1 below.


**Table 1.** Composition of experimental diets (g/kg diet).

The mice were anesthetized (Ketamine and xylazine, 75 and 10 mg/kg, respectively), and euthanized by cardiac puncture when they reached 35 weeks of age. The blood and the organs were collected for subsequent analyses.

#### *2.9. Body Weight and Diet Intake*

The mice were weighed weekly to observe weight changes until the end of the study. Food intake was measured three times per week.

The energy intake was calculated by multiplying the amount of diet ingested (g/day/animal) by the energy density of each diet, expressed in kcal/day per animal. In addition, the calculation of the feed efficiency index (FEI) was performed using the following equation:

$$\text{Free efficiency index} = \frac{(FW - IW)}{TF} \tag{2}$$

where *FW* is the final body weight in grams, *IW* is the initial body weight in grams, and *TF* is the total amount of food ingested in grams [27].

#### *2.10. Biochemical Analysis*

Serum glucose, serum triglyceride, total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), and very low-density lipoprotein (VLDL) levels were analyzed by the enzymatic colorimetric test, according to the manufacturer's instructions (Labtest®, Lagoa Santa, Minas Gerais, Brazil). The atherogenic index was determined by the ratio between total cholesterol and HDL cholesterol [14].

#### *2.11. Oral Glucose Tolerance Test*

The oral glucose tolerance test (OGTT) was performed one day prior to initiating treatment with the AGE or saline solution, and three days prior to the euthanasia of animals after six hours of fasting. Fasting glucose was verified via flow rate (time 0) using a G-Tech® glucometer (G-TECH Free, Infopia Co., Ltd. South Korea). Then, the animals received a D-glucose solution (Sigma Aldrich, Duque de Caxias, Rio de Janeiro, Brazil), at 2 g/kg of body weight, by gavage. A blood glucose reading was performed 15, 30, 60 and 120 min after glucose application. The area under the curve (AUC) was calculated for each mouse, and the mean was calculated for each experimental group [28].

#### *2.12. Insulin Sensitivity Test*

The insulin sensitivity test was performed five days before euthanasia. Glycemia was verified with the animals in a fed state (time 0). Then, 0.75 units of insulin (NovoRapid®, 100 U/mL, Novo Nordisk, Bagsvaerd, Denmark) per kg of animal weight was injected intraperitoneally. The blood glucose reading was performed at 15, 30 and 60 min using a G-Tech® glucometer (G-TECH Free, Korea). The area under the curve (AUC) was calculated for each animal, and the mean was calculated for each experimental group [28].

#### *2.13. Quantification of Cytokines in the Adipose Tissue*

Epididymal adipose tissue was collected, weighed (100 mg) and stored at –80 ◦C. For protein extraction, the epididymal adipose tissue was thawed on ice and homogenized in 1 mL of RIPA (RIPA Lysis Buffer, 10×, Cat. no. 20–188, MERCK, Darmstadt, Germany). A cocktail of protease inhibitors was added (Protease Inhibitor Cocktail Set Calbiochem, Cat. no. 539131, MERCK, Darmstadt, Germany).

The supernatant was collected after centrifugation at 4 ◦C and stored again at –80 ◦C until cytokine analysis, according to the recommendations of the manufacturer (MILLIPLEX MAP/Mouse Cytokine/Chemokine and Adipocyte Magnetic Bead panel) (Millipore, Billerica, MA, USA). The concentrations of the following cytokines were analyzed: IL-10, IL-6, MCP-1, and TNF-α using the MCYTOMAG-70K kit, and adiponectin using the MADCYMAG-72K kit. The concentration of the cytokines IL-10, IL-6, MCP-1, and TNF-α in the adipose tissue was expressed as cytokine picograms in relation to protein content (mg of protein). For adiponectin, the values were expressed as nanograms of cytokines in relation to protein content (mg of protein). Protein quantification was based on the bicinchoninic acid assay (BCA) following the manufacturer's recommendations (BCA Protein Assay kit) (MERCK, Darmstadt, Germany) [29,30].

#### *2.14. Assessment of Body Fat and Liver Weight*

After euthanasia, the liver and fat pads of white adipose tissue (omental, epididymal, perirenal, retroperitoneal, and mesenteric) were dissected and weighed. The adiposity index was calculated as the total sum of visceral white adipose tissue (g) divided by the final body weight of the animal × 100 and expressed as percentage of adiposity [31].

#### *2.15. Histopathological Analysis*

Samples of the liver, pancreas, and epididymal adipose tissue were fixed with 10% formalin solution. After fixation, the specimens were dehydrated, embedded in paraffin, cut in a microtome to a thickness of 5 mm each, and stained with hematoxylin-eosin. An expert pathologist performed the histological analysis of the liver and pancreas. For the analysis of treatment effects on the hepatocytes, a scoring system was used [32]. In the evaluation of the architecture of the pancreas, there were changes in the Islets of Langerhans and pancreatic acini, and inflammation was observed [33,34]. For the analysis of the adipocyte area of the epididymal adipose tissue, the images were initially taken using a LEICA DFC 495 digital camera system (Leica Microsystems, Wetzlar, Germany) integrated into a LEICA DM 5500B microscope (Leica Microsystems, Wetzlar, Germany), with a magnification of 20X. The images were analyzed using the LEICA Application Suite software, version 4.0 (Leica Microsystems, Wetzlar, Germany), and the mean area of 100 adipocytes per sample was determined [35].

#### *2.16. Statistical Analyses*

The results were expressed as mean ± MSE (mean standard error). For multiple comparisons of parametric results, an ANOVA followed by a Tukey post-test were performed. The Student t-test was performed for comparison between two groups. The chi-square test was used to evaluate associations in histological analyses. A significance level of *p* < 0.05 was adopted. Statistical analysis was performed using the software Jandel Sigma Stat, version 3.5 (Systat software, Incs., San Jose, CA, USA), and Sigma Plot, version 12.5 (Systat Software Inc., San Jose, CA, USA).

#### **3. Results**

#### *3.1. Chemical Composition*

The content of phenols, flavonoids, and tannins in AGE was 156.37 ± 1.2 mg/g, 92.07 ± 1.8 mg/g and 42.99 <sup>±</sup> 0.6 mg/g, respectively. The antioxidant activity of IC50 was 12 <sup>±</sup> 0.1 <sup>μ</sup>g·mL<sup>−</sup>1. In addition, six compounds were isolated and identified in the extract: kaempferol-3-*O*-a-l-rhamnopyranoside, quercetin 3-*O*-rutinoside, kaempferol 3-*O*-rutinoside, luteolin, quercetin, and sitosterol-3-*O*-β-d-glucopyranoside.

#### *3.2. Acute Oral Toxicity*

The results showed no signs of systemic toxicity. There are no changes in body weight, water consumption, food intake, and excretion of urine and feces. In addition, no changes in the Hippocratic screening test were observed, such as motor and/or sensory and neurological changes, as no animals died. The weight of the liver, spleen, pancreas, lungs, heart, and kidneys did not show significant differences among groups. Macroscopic changes in the organs of the animals were not visualized (Supplementary Material Figure S1).

#### *3.3. E*ff*ects of AGE on Body Weight and Food Intake*

At the beginning of the experiment, the animals in the HFD group did not present significant differences in body weight when compared to animals in the SHAM group (*p* = 0.971) (Table 2).


**Table 2.** Initial and final weight, weight gain and food intake assessment during obesity induction between the first and the 12th week.

SHAM: Standard diet. HFD: Hyperlipidic diet. Values represent the mean ± mean standard error;, \*\* *p* ≤ 0.01, \*\*\* *p* ≤ 0.001 vs. SHAM SALINE. Student *t* test.

However, with the HFD, the weight evolution evidenced a greater gain of body weight in the HFD group, with maintenance of a significant difference from the fourth week up to the 12th week (*p* = 0.005) compared to the control group (*p* ≤ 0.001) (Figure 1A).

Then, the animals of groups receiving a hyperlipidic diet began treatments with different doses of the extract or saline solution. At the 12th week, there was a significant difference in body weight in relation to the SHAM SALINE group (Table 3). However, this difference was not stable throughout the treatment. At the end of the 24th week, this group had a statistically similar body weight compared to the other groups (Figure 1B).

**Table 3.** Initial and final weight, weight gain, and food intake of control animals and animals treated with AGE between the 13th and the 24th week.


SHAM SALINE: standard diet + saline solution. HFD SALINE: hyperlipidic diet + saline solution. AGE 50: hyperlipidic diet + 50 mg/kg of aqueous graviola leaf extract. AGE 100: hyperlipidic diet + 100 mg/kg of aqueous graviola leaf extract. AGE 150: hyperlipidic diet + 150 mg/kg of aqueous graviola leaf extract. Values represent the mean ± mean standard error. In the same line, \* *p* ≤ 0.05, \*\* *p* ≤ 0.01, \*\*\* *p* ≤ 0.001 vs. SHAM SALINE; § *p* ≤ 0.05 vs. HFD SALINE; ANOVA followed by post Tukey test.

At the end of the treatment with the extract, the groups AGE 50 mg/kg and AGE 100 mg/kg presented a lower weight gain in comparison to the other groups (Table 3). The group AGE 50 mg/kg presented a statistical difference in relation to the SHAM SALINE group (*p* = 0.034). The group AGE 100 mg/kg had a significant body weight loss compared to the SHAM SALINE (*p* = 0.003) and HFD SALINE (*p* = 0.034) groups.

Food intake during the induction period was significantly higher in the SHAM group than in the HFD group (*p* ≤ 0.001). However, the caloric intake and the FEI were significantly higher in the HFD group than in the SHAM group (*p* ≤ 0.001) (Table 2). During the treatment period with the extract, similar results were observed for food intake, but daily caloric intake was significantly higher in the groups AGE 50 mg/kg (*p* = 0.007), AGE 100 mg/kg (*p* = 0.048) and AGE 150 mg/kg (*p* ≤ 0.001), compared to the SHAM SALINE group. However, the FEI was significantly lower in the group AGE 100 mg/kg when compared to the SHAM SALINE group (*p* = 0.021) and the HFD SALINE (*p* = 0.035) group (Table 3). That is, there was a lower feed conversion capacity into body mass in the group AGE 100 mg/kg.

**Figure 1.** Effects of a hyperlipidic diet and AGE on weight gain. (**A**) Body weight of animals fed on the standard diet (SHAM) and on the hyperlipidic diet (HFD) during obesity induction for 12 consecutive weeks (0: initial weight). (**B**) Weight of control animals (SHAM SALINE: standard diet + saline solution. HDF SALINE: hyperlipidic diet + saline solution) and of animals treated with aqueous graviola leaf extract (AGE) (AGE 50: hyperlipidic diet + 50 mg/kg AGE. AGE 100: hyperlipidic diet + 100 mg/kg AGE. AGE 150: hyperlipidic diet + 150 mg/kg AGE) (depicted on the graph from the first day of treatment up to the 24th week). Values represent mean ± mean standard error. \* *p* ≤ 0.05, \*\* *p* ≤ 0.01 and \*\*\* *p* ≤ 0.001 vs. SHAM SALINE. Student t-test (Figure 1A) and ANOVA followed by post Tukey test (Figure 1B).

#### *3.4. E*ff*ects of AGE on Serum Biochemical Parameters*

In this experimental model, the AGE was not able to decrease serum fasting glucose concentrations at the end of the study (*p* = 0.242) (Figure 2A).

**Figure 2.** Evaluation of serum parameters. (**A**) Blood glucose (mg/dL), (**B**) total cholesterol (mg/dL), (**C**) LDL-cholesterol (mg/dL), (**D**) HDL-cholesterol (mg/dL), (**E**) VLDL-cholesterol (mg/dL), (**F**) triglycerides (mg/dL), and (**G**) atherogenic index of control animals (SHAM SALINE: standard diet + saline solution, HFD SALINE: hyperlipidic diet + saline solution), and of animals treated with aqueous graviola leaf extract (AGE) at 50, 100, and 150 mg/kg + hyperlipidic diet between the 13th and the 24th week of study. Values represent mean ± mean standard error. \* *p* < 0.05, \*\* *p* < 0.01 vs. SHAM SALINE; § *p* < 0.05 vs. HFD SALINE. ANOVA followed by post Tukey test.

The AGE was also not able to significantly change the total serum cholesterol and HDL concentration (Figure 2B,D). However, serum HDL-cholesterol levels showed a 30.35% increase in concentration in the group treated with AGE 150 mg/kg (61.57 ± 6.47 mg/dL), compared to the HFD SALINE group (42.88 ± 5.40 mg/dL) (Figure 2D). For total serum cholesterol, the percentage decrease in AGE-treated groups was 4.92% for AGE 50 mg/kg (200.84 ± 8.30 mg/dL), 20.54% for AGE 100 mg/kg (167.85 ± 8.22 mg/dL), and 17.49% for AGE 150 mg/kg (174.30 ± 13.10 mg/dL) compared to the HFD SALINE group (211.24 ± 16.33 mg/dL) (Figure 2B).

In this study, the decrease in LDL-cholesterol concentration in the AGE-treated groups seems to be directly associated with the dose given. That is, the higher the AGE dose, the greater the decrease. There was a significant difference (*p* = 0.038) between AGE 150 mg/kg in relation to the HFD SALINE group (Figure 2C). In addition, AGE was able to significantly decrease triglyceride concentrations in the treated groups at doses of 100 mg/kg (*p* = 0.026) and 150 mg/kg (*p* = 0.025) compared to the SHAM SALINE group (Figure 2F). There was also a decrease in VLDL cholesterol using AGE 100 mg/kg and 150 mg/kg (*p* = 0.030) compared to the SHAM SALINE group (Figure 2E). Regarding the atherogenic index, which evaluates the risk of developing cardiovascular diseases, the group AGE 150 mg/kg presented a significantly lower mean value (*p* = 0.025) than the HFD SALINE group. In addition, the HFD SALINE group presented a significantly higher value in relation to the SHAM SALINE group (Figure 2G).

#### *3.5. E*ff*ects of AGE on Insulin Sensitivity and Glucose Tolerance*

The OGTT was performed prior to the beginning of the AGE treatment. No significant increases in fasting glycemia were observed between the hyperlipidic and the normolipidic diet groups. However, there was a significant increase (*p* ≤ 0.05) in the glycemia of animals at 15 min in the groups HFD SALINE and AGE 150 mg/kg, in relation to the group fed on a normolipidic diet (SHAM SALINE). At 30 min, all groups fed on a hyperlipidic diet had a significant increase in glycemia in relation to the SHAM SALINE group (Figure 3A).

**Figure 3.** Evaluation of the glycemic profile before and at the end of the treatment with AGE. (**A**) Oral glucose tolerance test prior to the beginning of treatment (12th week). (**B**) Area under the curve (AUC) of blood glucose of animals evaluated prior to the beginning of treatment (12th week). (**C**) Oral glucose tolerance test at the end of treatment (24th week). (**D**) Area under the curve (AUC) of glycemia of animals evaluated at the end of treatment (24th week). SHAM SALINE: standard diet + saline solution. HFD SALINE: hyperlipidic diet + saline solution. AGE 50: hyperlipidic diet + 50 mg/kg of aqueous graviola leaf extract. AGE 100: hyperlipidic diet + 100 mg/kg of aqueous graviola leaf extract. AGE 150: hyperlipidic diet + 150 mg/kg of aqueous graviola leaf extract. Values represent mean ± mean standard error. \* *p* < 0.05 vs. SHAM SALINE, § *p* < 0.05 vs. HFD SALINE, # *p* < 0.05 vs. AGE 150. ANOVA followed by post Tukey test.

The OGTT performed at the end of the experiment indicated that the AGE 150 mg/kg dose was able to significantly reduce blood glucose (*p* ≤ 0.05) at 15 min, in relation to the groups SHAM SALINE, HFD SALINE and AGE 50 mg/kg. However, there was no significant difference between AGE 100 mg/kg and AGE 150 mg/kg. The area under the curve did not indicate a significant difference for the comparison among groups (Figure 3B).

The insulin sensitivity test performed at the end of the treatment with AGE or saline solution did not present a statistical difference in glycemia at the times analyzed after administration of insulin. This result is confirmed by observing the total area under the curve among the groups that received AGE or SALINE during the experimental period (Figure 4A,B).

**Figure 4.** Evaluation of the glycemic profile at the end of the treatment with AGE. (**A**) Insulin sensitivity test performed at the end of treatment. (**B**) Area under the curve (AUC) of the insulin sensitivity test at the end of treatment. SHAM SALINE: standard diet + saline solution. HFD SALINE: hyperlipidic diet + saline solution. AGE 50: hyperlipidic diet + 50 mg/kg of aqueous graviola leaf extract. AGE 100: hyperlipidic diet + 100 mg/kg of aqueous graviola leaf extract. AGE 150: hyperlipidic diet + 150 mg/kg of aqueous graviola leaf extract. Each column represents the mean, and the bar represents the mean standard error. ANOVA.

#### *3.6. E*ff*ects of AGE on Anti- and Pro-inflammatory Cytokines, Chemokines and Adiponectin*

The animals treated with AGE showed an increase in IL-10 concentration, with a significant difference for the groups AGE 100 mg/kg (*p* = 0.021) and AGE 50 mg/kg (*p* = 0.042) when compared to the HFD SALINE group (Figure 5A). In analyzing MCP-1, no significant changes were observed between the groups studied (*p* = 0.840) (Figure 5B). As shown in Figure 5C,D, the levels of proinflammatory cytokines TNF-α (*p* = 0.640) and IL-6 (*p* = 0.768) also did not differ between study groups. Furthermore,

there was no significant difference (*p* = 0.244) in the levels of adiponectin in the adipose tissue of mice (Figure 5D).

**Figure 5.** Effects of AGE on anti- and pro-inflammatory cytokines, chemokines and adiponectin. (**A**) Interleukin-10 (pg/mg protein). (**B**) Interleukin-6 (pg/mg protein). (**C**) Monocyte-1 chemotactic protein (pg/mg protein). (**D**) Tumor necrosis factor alpha (pg/mg protein). (**E**) Adiponectin (ng/mg protein) of control animals (SHAM SALINE: standard diet + saline solution. HFD SALINE: hyperlipidic diet + saline solution) and of animals treated with aqueous graviola leaf extract (AGE) at 50, 100 and 150 mg/kg + hyperlipidic diet between the 13th and the 24th week of study. The cytokines are measured in adipose tissue. Values represent mean <sup>±</sup> mean standard error. § *p* < 0.05 vs. HFD SALINE. ANOVA/Tukey. Kruskal-Wallis test.

#### *3.7. E*ff*ects of AGE on Fat Pads, Adiposity Index and Liver Weight*

The AGE at the doses studied was not able to reduce the rate of adiposity of the animals. However, there was a decrease, although not significant, in the weight of fat pads of groups treated with aqueous graviola leaf extract: AGE 50 mg/kg: omental (22.22%), mesenteric (6.79%), retroperitoneal (8.79%), perirenal (9.41%); AGE 100 mg/kg: omental (11.11%), epididymal (14.34%), mesenteric (21.36%), retroperitoneal (31.04%); and AGE 150 mg/kg: omental (33.33%), epididymal (6.17%), mesenteric (6.08%), retroperitoneal (3.16%), and perirenal (2.35%), all compared to the HFD SALINE group (Table 4).


**Table 4.** Effects of AGE on fat pads, adiposity index, and liver weight.

Values represent mean ± mean standard error. \* *p* ≤ 0.05, \*\*\* *p* ≤ 0.001 vs. SHAM SALINE; ANOVA followed by post Tukey test.

#### *3.8. E*ff*ects of AGE on Liver, Pancreas, and Epididymal Adipose Tissue*

The histological analysis of the pancreas showed no statistical differences among groups regarding pancreatic acini (*p* = 0.400), Islet of Langerhans (*p* = 0.291), and inflammation (*p* = 0.458) (Table 5, Figure 6). However, the atrophy/necrosis was less frequent in the pancreas of animals treated with AGE, especially in the AGE 100 mg/kg group (Table 5).

**Table 5.** Results for changes observed in the pancreas of the animals in each experimental group.


Data presented as relative frequency (absolute frequency). Value of *p* in the chi-square test.

Similarly, the liver histological analysis also showed that the treatment with AGE did not change the quantification of steatosis (*p* = 0.881), microvesicular steatosis (*p* = 0.501), lobular inflammation (*p* = 0.501), balloonization (*p* = 0.192), Mallory's Hyaline (*p* = 0.408), apoptosis (*p* = 1.00), and glycogenate nucleus (*p* = 0.408) (Table 5, Figure 6). However, ballooning was more frequent in the HFD SALINE group compared to the groups that received AGE at different concentrations when fed on a hyperlipidic diet. Furthermore, hepatic steatosis was also frequent in the experimental groups that received a hyperlipidic diet or a normolipidic diet. However, the carbohydrate content in normolipidic diet was high, which may have contributed to this result (Table 6).

**Figure 6.** Histological analysis of the liver (Black arrows indicate hepatic steatosis, arrow head lobular inflammation and red arrows indicate ballooning) and pancreas of each experimental group. 20× magnification. Bar scale: 100 μm.


**Table 6.** Results for changes observed in the liver of animals in each experimental group.


**Table 6.** *Cont.*

Data presented as relative frequency (absolute frequency). Value of *p* in the chi-square test.

Regarding adipocytes, AGE at the dose of 100 mg/kg (4682.52 <sup>±</sup> 476.91 <sup>μ</sup>m2) and at the dose of 150 mg/kg (4410.54 <sup>±</sup> 426.73 <sup>μ</sup>m2) was able to significantly reduce the adipocyte area of the epididymal adipose tissue compared to the HFD SALINE group (6675.10 <sup>±</sup> 736.87 <sup>μ</sup>m2) (Figure 7).

**Figure 7.** Histological analysis of the epididymal adipose tissue of each experimental group. (**A**) SHAM SALINE group. (**B**) HFD SALINE group. (**C**) AGE 50 mg/kg group. (**D**) AGE 100 mg/kg group. (**E**) AGE 150 mg/kg group. 20<sup>×</sup> magnification. Bar scale: 100 <sup>μ</sup>m. **(F)** Adipocyte area (μm2) of the groups studied. Values represent mean <sup>±</sup> mean standard error. § *<sup>p</sup>* <sup>&</sup>lt; 0.05 vs. HFD SALINE. ANOVA followed by *post* Tukey test.

#### **4. Discussion**

The choice of *Annona muricata* Linn for this study is because this plant is used for the treatment of obesity and its comorbidities. However, more scientific evidence is needed to support the notion that this plant extract can be used for treating obese patients [36].

Plants interact with the environment to survive and are influenced by many factors, such as pathogen attacks, temperature, circadian rhythm, water availability, nutrients, pollutants, and pesticides, all of which can cause stress. In response, plants produce secondary metabolites such as flavonoids, coumarins, saponins, alkaloids, tannins, and glucosinolates, among others. Thus, plants of the same species grown in different environments may present different concentrations of a certain secondary metabolic compound [37]. In a previous study, a high concentration of tannins and a medium concentration of flavonoids and saponins were identified in the methanolic and aqueous leaf extracts of *A. muricata* [38]. These substances were absent from the aqueous graviola leaf extract produced in this study. However, another study identified a low concentration of flavonoids and a high concentration of tannins, alkaloids, phenols, saponins, and phytosterols in AGE [39]. Thus, the different concentrations of the chemical composition of *Annona muricata* Linn leaves found in the literature and in our study may be related to the mentioned factors.

Studies have indicated that two hundred and twelve bioactive compounds have already been identified in *Annona muricata* Linn. Phenolic compounds are the major phytochemicals responsible for the antioxidant activity of *Annona muricata* Linn [36,40–42].

Acute toxicity tests using AGE found in the literature corroborate the results presented here. In the literature, the administration of a single dose of 2000 mg/kg and 5000 mg/kg of AGE to mice was not able to induce changes in animal behavior or mortality, or visible macroscopic changes in organs after euthanasia on the 14th day of the experiment [14].

Experimental models with modified diets can simulate pathophysiological changes in rodents that are similar to what occurs in humans. Such experiments allow understanding of the specific mechanisms of obesity and its metabolic changes. However, the feed composition and duration of experimental period have not been consistently established in the literature. In general, high-fat diets and physical inactivity are used in these models and are also the main risk factors for humans [40,43].

In experimental models with a high-fat diet, the increase in body weight is significant after two weeks of treatment, and after four weeks of induction this model shows different obesity phenotypes. However, long-term induction leads to obesity-related comorbidities such as moderate hyperglycemia and glucose intolerance [43]. Furthermore, in another model using C57BL/6J mice fed on a high-fat diet and 10% fructose after 16 weeks of treatment, the animals developed central obesity, dyslipidemia, arterial hypertension, insulin resistance, systemic oxidative stress, inflammation, and steatohepatitis. These are the main characteristics of metabolic syndrome [44].

In our study, we exposed mice to a 58% lipid diet for 12 weeks to induce obesity. After this period, our results indicated a significant increase in the weight of the HFD group compared to the SHAM group. The weight gain in the HFD group is consistent with the higher caloric intake evidenced by the FEI. Furthermore, at the end of the experimental period, we verified a significant increase in total and LDL cholesterol, and in the atherogenic index, of the HFD group in relation to the control group, which indicates that the model allowed the desired changes.

Among the various medicinal plants used for weight reduction, *Annona muricata* Linn is the second most used by the Brazilian population [16]. In our study, AGE at the doses 50 mg/kg and 100 mg/kg represented the popularly understood relationship between graviola and weight loss. Despite an observed decrease in body weight, no reduction in caloric intake was observed in the groups treated with AGE during the experimental period. Thus, weight reduction in this study is probably not related to a lower caloric intake.

Effective medicinal plants for weight loss have phenolic compounds among their chemical constituents, such as flavonoids, which modulate lipid metabolism and increase the rate of basal metabolism [45]. Quercetin and kaempferol stand out among flavonoids with an antiobesity effect. These were identified in AGE in the literature. We also found them in our study [42,46].

In the current study, regarding the cholesterol profile, no significant effects of the aqueous extract were observed on total cholesterol and HDL in two of the groups that received AGE. However, a significant increase in HDL cholesterol was observed in the group treated with AGE 150 mg/kg compared to the HFD SALINE group. We also observed a significant (*p* = 0.038) decrease in LDL-cholesterol, VLDL-cholesterol (*p* = 0.030), and triglyceride (*p* = 0.026) concentrations. In a study with streptozotocin-induced diabetic rats, AGE was able to significantly reduce plasma lipid concentrations. However, no difference was observed in relation to the diabetic group treated with 10 IU/kg of insulin [14].

The mechanisms of action of the aqueous extract of graviola on metabolism are not fully understood. However, several studies have reported isolated chemical compounds such as tannins, flavonoids, saponins, and coumarins, among other constituents, as being responsible for hypoglycemic, hypolipidemic, hypotensive, anti-inflammatory, and hepatic tissue changes, among other properties [47].

In our study, we observed a decrease in the atherogenic index after treatment with AGE, mainly in the group treated with 150 mg/kg. This decrease is directly related to the decrease in the development of cardiovascular diseases. This was associated with a decrease in triglycerides and LDL-cholesterol, and an increase in HDL-cholesterol, mainly in the 150 mg/kg group. Previous studies have shown that the antioxidant capacity of some substances can modify lipid metabolism and reduce inflammation, suggesting positive effects on cardiovascular diseases mainly by modulating oxidative stress. Furthermore, the high plasma level of the atherogenic index is related to small LDL-cholesterol particles. This is a predictor of conditions such as obesity, insulin resistance and inflammation, and consequently coronary artery disease, diabetes mellitus, and metabolic syndrome [48,49].

In our study, there were no significant effects of aqueous graviola leaf extract on capillary fasting glycemia evaluated in the oral glucose tolerance test performed at the end of treatment, and in the serum concentration of fasting glucose. When we calculated the area under the curve at the end of the experiment, we did not observe a significant difference in the comparison between the groups. However, there was a reduction in blood glucose levels at 15 min according to the oral glucose tolerance test in group 150 mg/kg. Some studies have demonstrated a significant decrease in plasma glucose concentrations after treatment with graviola extract in diabetic animals induced by streptozotocin or monohydrate aloxane [14,47,50,51].

Thus, the results found in our study do not indicate the effectiveness of aqueous graviola leaf extract on insulin resistance and diabetes mellitus type 2, in relation to the intake of a high calorie diet, high in saturated fat and simple carbohydrates and low in dietary fiber associated with sedentary lifestyle. However, further studies with AGE concentrations above 150 mg/kg may prove effective in reducing blood glucose, and therefore should be conducted.

Evidence shows that a greater fluctuation of glycemia induces endothelial dysfunction in diabetic or non-diabetic individuals, through oxidative stress resulting from an increase in free radicals [52,53].

Pro-inflammatory cytokines and chemokines, such as TNF-α, IL-6 and MCP-1, are required to initiate an inflammatory response. TNF-α is a cytokine that initiates the inflammatory response since it triggers the production of other cytokines, such as IL-6. On the other hand, anti-inflammatory cytokines, such as IL-10, are required to inhibit the synthesis of proinflammatory cytokines [54].

Previous studies have demonstrated that secondary metabolites present in plants, such as triterpenes, flavonoids and steroids, can modulate the inflammation and metabolic dysfunctions associated with obesity [55]. In our study, AGE did not change the levels of the inflammatory markers TNF-α, IL-6 and MCP-1 in adipose tissue. On the other hand, the AGE showed an anti-inflammatory effect due to a significant increase in IL-10 levels at the AGE doses of 50 and 100 mg/kg. In this study, the increased doses of AGE did not significantly interfere with TNF-α, IL-6 and MCP-1 levels. However, recent studies have demonstrated that IL-10 can exert anti-inflammatory effects via Janus

kinase (JAK) signal transducer of activation 3 (JAK-STAT3), by binding IL-10 to the receptor on the target of the cell membrane—tyrosine kinase 2—leading to activation of the signal transducer and activator of transcription 3 (STAT3). However, further studies are needed to evaluate the possible effects of AGE on this pathway [56].

Adiponectin is a protein secreted by adipocytes. It exerts anti-diabetic, anti-atherogenic and anti-inflammatory effects directly. An increased expression may prevent and/or assist in the treatment of metabolic diseases related to obesity [57]. In our study, no significant effects of AGE were observed on adiponectin in adipose tissue. However, an increase of this protein was noticed in the group treated with AGE 50 mg/kg in relation to the other groups treated with AGE. Furthermore, the HFD SALINE group presented the lowest levels of adiponectin among the groups in our study.

Although the aqueous graviola leaf extract is able to induce a significant reduction in body weight according to the experimental model studied, and although there was a decrease in the weight percentage of all fat pads evaluated without significant differences in the comparison among groups, no decrease of visceral adiposity was observed at the end of the experiment when analyzing the weight of fat pads and the adiposity index. It is also possible to observe a significant decrease in the epididymal adipocyte area in the animals treated with AGE. Therefore, AGE attenuates the accumulation of lipids in mice, as was reported by another study after administration of blueberry and mulberry juice to C57BL/6 mice fed on a hyperlipidic diet for 12 weeks [58]. It should be noted that epididymal adipose tissue in mice is one of the major deposit areas of visceral fat [44].

In our experimental model, aqueous graviola leaf extract at the doses studied is not sufficient to prevent accumulation of liver fat and lesions to hepatocytes, as well as lesions to the pancreas. However, the ballooning of hepatocytes is less frequent in animals receiving treatment with the extract, as well as necrosis/atrophy of pancreatic acini. Thus, treatment with AGE is not able to avoid hepatic changes. However, it seems to protect the hepatocytes from morphological changes.

This may be related to a decrease in oxidative stress. In yet another study, the aqueous graviola leaf extract of *Annona muricata* Linn was able to protect pancreatic β-cells, and hence improve glucose metabolism, which was not visualized in our results [14].

#### **5. Conclusions**

In conclusion, no neurotoxic, behavioral, or mortality effects are produced by AGE in the acute toxicity test immediately after or during the post-treatment period. In addition, this study confirms the popular knowledge that graviola leaf tea reduces body weight and may also reduce cardiovascular risks, due to its beneficial effects in reducing plasma concentrations of LDL-cholesterol, VLDL-cholesterol, triglycerides, and the atherogenic index, while also attenuating the accumulation of body fat. In addition, in our experimental model, the results found do not indicate the effectiveness of aqueous graviola leaf extract on insulin resistance and diabetes mellitus type 2. However, the extract was effective in improving glucose tolerance in the higher concentration of the AGE. Furthermore, AGE has anti-inflammatory activity due to the increase in IL-10. However, it does not inhibit the expression of TNF-α, IL-6 and MCP-1. These data support the utility of conducting further studies aimed at identifying the active compounds of the aqueous extract of the aqueous graviola leaf extract, and at clarifying its mechanism of action.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6643/11/7/1509/s1, Figure S1: Body weight, weight of organs, food intake and water intake of animals in the control group and animals treated with AGE during the acute toxicity test.

**Author Contributions:** Conceptualization: S.S., K.d.C.F.; methodology: S.S., P.C.S.e.S., L.F.S., C.A.L.C., L.C.P., B.B.d.F., L.S.S., L.M.B., A.R.C.M.-C., A.F.d.S.; F.M.A.; R.d.C.A.G., K.d.C.F.; validation: S.S., RCAG KCF; formal analysis: S.S., P.C.S.e.S., L.F.S., C.A.L.C., L.C.P., R.d.C.A.G., K.d.C.F.; writing—review and editing: S.S., L.F.S., C.A.L.C., F.M.A.; L.C.P., B.B.d.F., A.F.d.S.; A.R.C.M.-C., R.d.C.A.G., K.d.C.F.; data curation: S.S., R.d.C.A.G., KCF; project administration S.S., R.d.C.A.G., K.d.C.F.; software: KCF R.d.C.A.G.; visualization: S.S., R.d.C.A.G., KCF; supervision: S.S., L.F.S., C.A.L.C., R.d.C.A.G., K.d.C.F.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors would like to thank CAPES for the scholarship awarded.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

#### *Article*

### **Polyphenol-Enriched Plum Extract Enhances Myotubule Formation and Anabolism while Attenuating Colon Cancer-induced Cellular Damage in C2C12 Cells**

#### **Faten A. Alsolmei 1,2, Haiwen Li 1, Suzette L. Pereira 3, Padmavathy Krishnan 4, Paul W. Johns <sup>3</sup> and Rafat A. Siddiqui 1,\***


Received: 5 April 2019; Accepted: 11 May 2019; Published: 15 May 2019

**Abstract:** Preventing muscle wasting in certain chronic diseases including cancer is an ongoing challenge. Studies have shown that polyphenols derived from fruits and vegetables shows promise in reducing muscle loss in cellular and animal models of muscle wasting. We hypothesized that polyphenols derived from plums (*Prunus domestica*) could have anabolic and anti-catabolic benefits on skeletal muscle. The effects of a polyphenol-enriched plum extract (PE60) were evaluated in vitro on C2C12 and Colon-26 cancer cells. Data were analyzed using a one-way ANOVA and we found that treatment of myocytes with plum extract increased the cell size by ~3-fold (*p* < 0.05) and stimulated myoblast differentiation by ~2-fold (*p* < 0.05). Plum extract induced total protein synthesis by ~50% (*p* < 0.05), reduced serum deprivation-induced total protein degradation by ~30% (*p* < 0.05), and increased expression of Insulin-Like Growth Factor-1 (IGF-1) by ~2-fold (*p* < 0.05). Plum extract also reduced tumor necrosis factor α (TNFα)-induced nuclear factor κB (NFκB) activation by 80% (*p* < 0.05) in A549/NF-κB-luc cells. In addition, plum extract inhibited the growth of Colon-26 cancer cells, and attenuated cytotoxicity in C2C12 myoblasts induced by soluble factors released from Colon-26 cells. In conclusion, our data suggests that plum extract may have pluripotent health benefits on muscle, due to its demonstrated ability to promote myogenesis, stimulate muscle protein synthesis, and inhibit protein degradation. It also appears to protect muscle cell from tumor-induced cytotoxicity.

**Keywords:** cachexia; plum; cancer; muscle wasting; myoblasts; protein synthesis

#### **1. Introduction**

Skeletal muscle weakness and wasting, which is also referred as cachexia, is a major clinical problem for advanced cancer patients [1]. In 1932, Warren described cachexia as the most common cause of death across a variety of cancers in a post mortem study of 500 patients [2]. The term "Cachexia" is derived from the Greek words "kakos" and "hexis," meaning "bad condition." It is a multi-organ syndrome associated with and characterized by at least 5% body weight loss due to muscle and adipose tissue wasting [3]. Cancer cachexia is a multifactorial syndrome that is common in advanced malignancy occurring in 80% of patients, which cannot be reversed by nutritional support and leads

to significant function deficits [4], and which is responsible for an estimated 20% of cancer-related deaths [5].

Colorectal cancer (CRC) patients are often presented with cachexia syndrome, which is a major contributor to colorectal cancer-related morbidity and mortality [4–8]. About 35 to 60% of CRC patients show some degree of muscle wasting and 28% lose >5% of their body weight in the six months preceding diagnosis [9]. Blocking muscle wasting can prolong life even in the absence of effects on tumor growth [10].

Oxidative stress through activating initial steps in protein degradation via the ubiquitin-proteasome pathway and the activation of caspases contributes to muscular atrophy [11–13]. In addition, inflammation also leads to muscle atrophy and this is mediated through cytokine (e.g., tumor necrosis factor α (TNFα), interleukin-6 (IL-6), and interferon γ (IFNγ)) induced activation of the nuclear factor κB (NF-κB) pathway [14].

Recent studies have shown that polyphenol-rich plant extracts prevent oxidative stress, reduce inflammation, and help reduce muscle atrophy. We have previously shown that curcumin treatment attenuated muscle wasting in cancer cachectic mice [15]. Supplementation with red grape polyphenols mitigated muscular atrophy in transforming growth factor (TGF) mice, a model of chronic inflammation, by reducing mitochondrial oxidative stress and by inhibiting caspase activation [16]. Grape seed extract supplementation effectively prevented muscle wasting in IL10-knock out mice [17]. Green tea polyphenol, catechins, protected normal and dystrophic muscle cells from oxidative damage [18]. Epigallocatechin-3-gallate (EGCG) supplementation preserved muscle in sarcopenic rats [19] and attenuated skeletal muscle atrophy caused by experimentally induced cancer cachexia in in mice [20]. More recently, ursolic acid—a polyphenol present in apple peels, basil leaves, prunes, and cranberries [21]—has been shown to increase muscle mass in mice exhibiting fasting-induced muscle atrophy [22]; it has also increased muscle mass, fast and slow fiber size, grip strength, and exercise capacity in mice with diet-induced obesity [23]. These observations clearly suggest that intake of polyphenols can be beneficial in preserving muscle mass.

The common plum (*Prunus domestica*) is well known to be rich in polyphenols and contains unique phytonutrients called neochlorogenic and chlorogenic acid which have high antioxidant activities. Among functional foods, plums are also considered "super foods" since their consumption has been associated with the decrease in chronic degenerative diseases and circulatory and digestive issues [24]. Dried plums have been shown to reduce symptoms of arthritis in an inflammation model [25]. These effects are attributed to their high polyphenolic composition and related high antioxidant activity [26]. Plums have several health benefits and studies have found that plums also initiate anti-cancer mechanisms that may help prevent the growth of cancerous cells and tumors [27–29].

In addition, plums have been extensively studied for their effects on bone health [30,31]. Plums contain caffeic acid (the polyphenol component of neochlorogenic and chlorogenic acids) and rutin, which have been shown to inhibit the deterioration of bone tissues and prevent diseases such as osteoporosis in postmenopausal women [32]. Research has also shown that regular consumption of dried plums helps in the restoration of bone density lost to aging [33].

Formation of bone and much of the skeletal tissues is derived from the proliferation and differentiation of skeletal stem cells. As dried plum was found to be a potent regulator of bone health, it is possible that plum and its associated polyphenols may have benefits on other cells of musculoskeletal system. Thus, in the present study, we sought to investigate the effect of a polyphenol-enriched plum extract on muscle cell growth and differentiation, and on muscle protein synthesis and degradation *in vitro*. In addition, we explored the effect of plum extract on inflammation as well as studied its effect on colon cancer cells.

#### **2. Materials and Methods**

#### *2.1. Materials*

Dulbecco's modified Eagle's medium (DMEM), fetal bovine serum (FBS), horse serum, and Penicillin-Streptomycin solution were purchased from Gibco-Thermo-Fisher Scientific (Grand Island, NY, USA). L-[2,3,4,5,6-3H] Phenylalanine and L-[Ring-3, 5-3H]-Tyrosine was purchased from Perkin-Elmer (Waltham, MA, USA), Prune extract-60% enriched polyphenol extract (PE60) was purchased from PL Thomas (Morristown, NJ, USA). All other chemicals were of reagent grade, and were purchased from Sigma Chemical Co. (St. Louis, MO, USA).

#### *2.2. Composition of the PE60-Plum Extract*

Free gallic acid, 3-cholorogenic acid, rutin, free quercetin, and proanthocyanidins were determined with an Agilent Technologies (Wilmington, DE, USA) Model 1200 HPLC System equipped with a Model G1311A quaternary pump, Model G1322A vacuum degasser, Model G1329A autosampler, Model G1316A thermostatted column compartment, a Model G1315B diode array detector, and a Chem Station data processor. The separations were performed with a YMC-Pack ODS-AQ analytical column (4.6 × 250 mm, 5 μm, P/N AQ12S05-2546WT, Waters Corporation, Milford, MA, USA), using mobile phase A = 1000/100 (*v*/*v*) 0.05 M KH2PO4, pH 2.9/acetonitrile, and mobile phase B = 200/800 (*v*/*v*) Milli-Q Plus water/acetonitrile, a column temperature of 40 ◦C, an injection volume of 5 μL, and the analytes were quantified at signals of 280 nm/590 nm (for gallic acid and the proanthocyanidins), 330 nm/590 nm (for 3-chlorogenic acid), and 375 nm/590 nm (for rutin and quercetin). The elution program was 0% mobile phase B from 0 to 5 min, 0 to 60% (linear gradient) mobile phase B from 5 to 35 min, 100% mobile phase B from 35 to 40 min, and 0% mobile phase B from 40 to 55 min (end). The PE60 extract was prepared for analysis by stirring (at room temperature for 15 min) 0.250 g in 100 mL of 50/50 (*v*/*v*) 0.05 M citric acid/methanol. The determinations were calibrated with standard solutions of gallic acid, 3-chlorogenic acid, rutin hydrate, and quercetin dihydrate (all obtained from Sigma-Aldrich, St. Louis, MO, USA), also prepared in the citric acid/methanol medium. The proanthocyanidin content was estimated by peak area proportionation vs. the corresponding peak areas (at 280 nm/590 nm) of grapeseed extracts (from Kikkoman, Polyphenolics, and Seppic) of known (i.e., label claim) proanthocyanidin content, included in the analysis. The anthocyanin concentration was estimated by a published colorimetric method [34]. During present investigation, minor isomers of chlorogenic acid (4-chlorogenic acid, 5-chlorogenic acid) were not determined.

#### *2.3. Characterization of Anti-Oxidation Capacity of the Plum Extract*

The PE60 (Lot PE6009-1601) extract was dissolved in water (10 mg/mL) and then centrifuged at 1500 × *g* for 10 min to remove any insoluble material. The dissolved material was sterile filtered and the filtrate was assayed for total polyphenols by the Folin Ciocalteu method [35], for total flavonoids by the AlCl3 complexation method [36], for anti-oxidant activity by the DPPH assay [37], and for oxygen scavenging activity by the ABTS assay [38], as described.

#### *2.4. Cell Culture*

C2C12 cell line (mouse myoblasts) were obtained from American Type Culture Collection (Manassas, VA, USA). The undifferentiated cells were grown in complete media consisting of Dulbecco's modified Eagle's medium (DMEM, 4.5 mg/mL glucose) supplemented with heat-inactivated fetal calf serum (10%), penicillin (100 units/mL), and streptomycin (100 μg/mL) at 37 ◦C in the presence of 5% CO2. The myoblasts were differentiated into myotubes by culturing them into differentiation medium, consisting of DMEM supplemented with heat-inactivated horse serum (5%), penicillin (100 units/mL), and streptomycin (100 μg/mL) for five days.

#### *2.5. Determination of C2C12 Myoblast Cell Size*

Muscles cells were grown in a 96-well plate for 24 h in 100 μL complete media. Cells were then treated with 0, 50, 100, 150, and 200 μg/mL of extract for 48 h to evaluate a dose-response effect of plum extract. After incubations, the cells were observed under a microscope and pictures (100 × magnification) were taken using a Nikon microscope with calibrated objectives. The size of cells was determined using Element-BR software (Nikon Instruments Inc, Melville, NY, USA).

#### *2.6. Assaying C2C12 Myoblast Di*ff*erentiation*

Muscle cells were initially cultured in a 96-well plate for 24 h in 100 μL complete media. Cells were then incubated with 0, 50, 100, and 200 μg/mL plum extract for five days and the medium containing corresponding concentration of plum extract was changed every 24 h. After treatment, the cells were washed once with PBS, and then fixed with cold 4% paraformaldehyde for 10 min on ice. The cells were washed three times with PBS and the monolayer was treated with blocking solution containing 2% albumin. The cells were then incubated with anti-myosin antibody at room temperature for 2 h. Cell were washed again and then incubated with anti-mouse Alexa-488 antibody (Abcam, Cambridge, MA) for two hours. Cells were washed again three times with PBS and the nuclei were stained briefly with Hoechst 33342 dye (1:2000 dilution). Pictures were taken at 200 × magnification using a Nikon Fluorescent Microscope (Nikon Instruments Inc, Melville, NY 11747, USA). Myotubes were defined as myosin positive cells with 2 or more fused nuclei.

#### *2.7. Protein Synthesis in Cultured C2C12 Myotubules*

C2C12 cells (375,000) were initially plated on a 12-well tissue culture plate that was initially coated with 2% gelatin. Cells were differentiated for five days in 5% horse serum (media was changed every two days) and then starved for 30 min by replacing the media with 1 ml PBS. The cells were then treated with 0, 50, 100, and 200 μg/mL of plum extract in PBS, spiked with [3H] phenylalanine (1μCi/well), and incubated for 2 h at 37 ◦C. The reaction was stopped by placing the plates on ice. Wells were washed two times with DPBS-media containing 2 mM cold phenylalanine. Further, 1 mL of 20% cold trichloroacetic acid (TCA) solution was added to each well and plates were incubated on ice for 1 h for protein precipitation. Wells were washed two times with cold TCA and then the precipitated proteins were dissolved in 0.5 mL of 0.5N NaOH containing 0.2% Triton X-100 overnight in a refrigerator. An aliquot (5 μL) of the NaOH solubilized material was used for protein determination and the rest of the dissolved proteins were mixed with scintillation fluid and counted. Data is computed as cpm/mg of proteins and then % change over control is calculated.

#### *2.8. Protein Degradation in C1C12 Myotubules*

C2C12 myoblasts were cultured and differentiated as described above. Cells were then labelled with [3H] Tyrosine 1 μCi/1 mL in serum free-DMEM (SF-DMEM) for 24 h. The unincorporated [3H] Tyrosine was removed by washing the cell monolayer three times with SF-DMEM containing 50 μM cycloheximide (protein synthesis inhibitor) and 2 mM non-labelled Tyrosine. Proteolysis was induced by serum deprivation for 48 h in the presence or absence of 50, 100, 200 μg/mL of plum extract in serum-free DMEM containing 50 μM cycloheximide. The extent of protein degradation was assayed by monitoring release of radioactive tyrosine in the media after 48 h of incubation and was expressed as protein degradation in comparison to control (normalized to 100%).

#### *2.9. Determination of Insulin-Like Growth Factor-1 (IGF-1) Expression*

Total RNA was extracted from C2C12 myotubules with RNeasy Plus Universal Mini Kit (Qiagen, Hilden, Germany), according to the manufacturer's instructions. The concentration and purity of RNA was determined by measuring the absorbance in a Nano drop spectrophotometer. RT2 First Strand Kit from Qiagen (Qiagen, Hilden, Germany) was used to synthesize first strand complementary DNA (cDNA). The gene expression levels were analyzed by Quantitative real-time RT-PCR conducted on the Bio-Rad CFX-96 Real-Time PCR System using RT2 SYBR Green Master mix (Bio-Rad Laboratories, Hercules, CA). The primers (*IGF*: forward primer GGACCAGAGACCCTTTGCGGGG and reverse primer, AGCTCAGTAACAGTCCGCCTAGA; *GAPDH*: forward primer ATCCCATCACCATCTTCCAG and reverse primer CCATCACGCCACAGTTTCC) were designed. Hot-Start DNA Taq Polymerase was activated by heating at 95 ◦C for 10 min and real time PCR was conducted for 40 cycles (15 s for 95 ◦C, 1 min for 60 ◦C). All results were obtained from at least three independent biological repeats. Data were analyzed using the ΔΔCT method. Glyceraldehyde-3-phosphate dehydrogenase (*GAPDH*) genes were used as house-keeping genes for expression calculation.

#### *2.10. Determination of NFkB Activation*

A549/NFkB-luc cells (Panomics Catalog No. RC0002) at 3 <sup>×</sup> <sup>10</sup>5/well were seeded in 1 mL of Initial Growth Media (Dulbecco's Modified Eagle's medium containing 10% FBS and 1% Pen-Strep) in a 12-well plate. The cells were incubated in a humidified incubator at 37 ◦C and 5% CO2 for 24 h to allow cells to recover and attach. After washing the cells once with serum-free media containing penicillin (100 units/mL), and streptomycin (100 mg/mL), 1 mL of this media was added to each well. Cells were pretreated with varying concentrations of plum extracts for 1 h at 37 ◦C and 5% CO2, and then TNFα was added to achieve a final concentration of 2 ng/mL to all wells except control untreated cells. The cells were incubated in a humidified incubator at 37 ◦C and 5% CO2 for 6 h. After treatment, the media was carefully removed. Cells were washed with PBS once and then lysed by 100 μL of 1× lysis buffer. Assay for luciferase activity was performed according to assay manufacturer's (Promega P/N E1500) recommendations. The average relative luminescence units (RLU) were calculated and corrected for baseline quenching for each set of triplicate wells, using WinGlow software (PerkinElmer, Waltham, MA 02451, USA and Microsoft Excel (Microsoft Corporation, Redmond, WA 98073, USA). The data is reported as the relative percent inhibition of TNFα mediated NFκB activation on A549 cells.

#### *2.11. E*ff*ect of Plum Extract on Colon-26 Proliferation and its' Soluble Factor Induced Cytotoxicity on C2C12 Myotubules*

Colon 26 cells, a mouse colon carcinoma cell line, was obtained from American Type Culture Collection (Manassas, VA, USA). Effect of plum extract on Colon-26 cell proliferation was assayed using a Water-Soluble Tetrazolium-1 (WST-1) (Talkara, Shiga, Japan) assay as described previously [39]. To determine the effects of soluble factors released from Colon-26 on C2C12 myotubules, conditioned media from Colon-26 culture was collected after 24 h. of cultivation. The media was centrifuged at 2500 × *g* for 20 min to remove cellular material. The clear supernatant (conditioned media) was diluted 1:10 with normal complete media. The C2C12 differentiated myoblasts were then treated with normal complete medium or with Colon-26 conditioned medium with or without 50 μg/mL plum extract. A lower dose of plum extract (50 μg/mL) was used to avoid a direct effect of higher dose of plum extract (100 μg/mL or 200 μg/mL) on protein synthesis and degradation. The cell viability was assayed using a WST-1 assay. Control cells were subjected to equal amounts of non-conditioned media.

#### *2.12. Data Analysis*

The data is expressed as mean ± SD for at least three replicates. All comparisons were made by one-way ANOVA with Tukey's -HSD-post-hoc test using SPSS Statistics 20 software. All significant differences are reported at *p* < 0.05 and indicated by "\*".

#### **3. Results**

#### *3.1. Characterization of PE60 Plum Extract Composition and Anti-Oxidation Properties*

As shown in Table 1, the major components identified in the polyphenol-enriched PE60 plum extract are proanthocyanidins, along with minor components such as anthocyanidins, 3-chlorogenic acid, rutin, quercetin (free), and gallic acid (free).

The PE60 was also characterized by determining total phenolic content (TPC), total flavonoid content (TFC), anti-oxidant activity (DPPH assay), and oxygen scavenging activity (ABTS). The data in Table 2 shows that the content of TPC was in the same range as reported by the commercial vendor (60%). The data indicate that the PE60 contained TPC in range 525–575 mg/g of dry extract. The TFC was in 480–560 mg/g dry weight range. The anti-oxidation effects as determined by inhibition of DPPH oxidation and ABTS assay ranged from 3280–3460 and 4000–4500 μM Trolox equivalents/g, respectively.


**Table 1.** Characterization of composition of polyphenol-enriched plum extract (PE60).

Contents in PE60 plum extract were determined either using an Agilent Technologies Model 1200 HPLC System (Wilmington, DE, USA) or a colorimetric method as described in Section 2.2 in the text. Values are mean ± SD of three experiments. \* LC/UV = liquid chromatography/ultraviolet light detection


The anti-oxidation properties of PE60 plum extract were determined using specific assays (TPC: total phenolic content, TFC: total flavonoid content, DPPH: 2,2-diphenyl-1-picrylhydrazyl, ABTS: 2,2'-azino-bis{3-ethylbenzothiazoline-6-sulfonic acid}) as described in Section 2.3 in the text. Values are mean ± SD of three experiments.

#### *3.2. E*ff*ect of PE60 Plum Extract on C2C12 Myoblast Size and Di*ff*erentiation*

Plum extract had no cytotoxic effect on myoblast when used even at a high dose of 250 μg/mL (data not shown). It is evident from images that plum extract has some effect on cell proliferation; however, it was interesting to note that the plum extract increased the size of undifferentiated myoblasts cells in a dose-dependent manner (Figure 1a). The size of myoblast increased ~two-fold (*p* < 0.05) after treating cells with 50 μg/mL of plum extract when compared to that of untreated-control cells. Increase in myoblast size plateaued to a maximum increase of three-fold at 200 μg/mL concentration (Figure 1b). The effect of plum extract was also assessed on myoblast differentiation. Figure 2a indicates that the plum extract stimulated differentiation of myoblast in a dose-dependent manner using expression of myosin heavy chain as a marker for differentiation. The number of myotubes formed resulting from fusion of differentiated cells was increased by two-fold in cells treated with

100 μg/mL plum extract (*p* < 0.05) and by three-fold at 200 μg/mL (*p* < 0.05) compared to that of control cells (Figure 2b).

**Plum Extract Figure 1.** *Cont.*

**E**

**Figure 1.** The effect of plum extract on C2C12 myoblast cell size. (**a**) The representative pictures of myoblast after treatment with varying concentration of plum extract (100 × magnified images) taken by a Nikon Microscope. The bar represents a length of 500 μm. (**b**) The size of myoblast was determined using Element-BR software as described in "Materials and Methods". The data are expressed as mean ± SD for at least three experiments. All comparisons were made to control (untreated cells) using one-way ANOVA; significant differences are reported at \* *p* < 0.05.

(**a**)

 **0 ΐg/mL 50 ΐg/mL** 

 **100 ΐg/mL 200 ΐg/mL**

**Plum Extract** 

**Figure 2.** The effect of plum extract on C2C12 myoblast differentiation. (**a**) Images of differentiated cells after treatment with varying concentration of plum extract showing nuclei stained in blue (Hoechst 33342) and myofibers stained in green (Alexa 488). Pictures were taken at 200× magnification using a Nikon Fluorescent Microscope. The bar represents a length of 300 μm. (**b**) Fused cells from five random fields were counted manually under 200× as described in "Materials and Methods". The data are expressed as mean ± SD for at least three experiments. All comparisons were made to control (untreated cells) using one-way ANOVA; significant differences are reported at \* *p* < 0.05.

#### *3.3. E*ff*ect of PE60 Plum Extract on Myotubule Protein Synthesis*

Plum extract showed almost a linear increase in [3H] phenylalanine incorporation into proteins in a dose dependent manner in C2C12 myotubules (Figure 3). Doses of 100 μg/mL and 200 μg/mL of plum extract caused a significant increase in protein synthesis by 30% and 50%, respectively (*p* < 0.05).

#### *3.4. E*ff*ect of PE60 Plum Extract on Myotubules Protein Degradation*

We also examined if plum extract could reduce myotubule protein degradation induced by serum starvation. Figure 4 revealed that plum extract did inhibit protein degradation in a dose-dependent manner. Doses of 100 μg/mL and 200 μg/mL significantly inhibited protein degradation by 20% and 30%, respectively (*p* < 0.05).

μ **Figure 3.** The effect of plum extract on myotubule protein synthesis. Protein synthesis was measure by the incorporation of labeled phenylalanine into total myotubule proteins in response to various levels of plum extract. Data were computed as cpm/mg of proteins followed by calculation of % change over control. The data were expressed as mean ± SD for at least three experiments. All comparisons were made to control (untreated cells) using one-way ANOVA; significant differences are reported at \* *p* < 0.05.

**Figure 4.** The effect of plum extract on myotubule protein degradation. Proteolysis was induced by 48 h-serum starvation in the presence or absence of plum extract, and monitored by release of radioactive tyrosine from pre-labelled cells. Data were computed as cpm/mg of proteins and then % change over control was calculated. The data were expressed as mean ± SD for at least three experiments. All comparisons were made to control (untreated cells) using one-way ANOVA; significant differences are reported at \* *p* < 0.05.

#### *3.5. E*ff*ect of PE60 Plum Extract on IGF-1 Expression in Myotubules*

Expression of IGF-1 mRNA in C2C12 myotubules upon treatment with plum extract is shown in Figure 5. Compared to that of untreated cells, low concentration of plum extract (50 μg/mL) has no significant effect on IGF-1 mRNA expression; however, it significantly stimulated IGF-1 expression when cells were treated at a higher dose (100 or 200 μg/mL) plum extract.

μ **Figure 5.** The effect of plum extract of IGF-1 gene expression. Total RNA was extracted from C2C12 myotubules treated with various concentrations of plum extract and compared to untreated control. All results were obtained from at least three independent biological repeats. Data were analyzed using the ΔΔCT method. Glyceraldehyde-3-phosphate dehydrogenase (*GAPDH*) genes were used as house-keeping genes for expression calculation. All comparisons were made to control (untreated cells) using one-way ANOVA; significant differences are reported at \* *p* < 0.05.

#### *3.6. Anti-Inflammatory E*ff*ect of PE60 plum Extract in Vitro*

We evaluated the anti-inflammatory activity of plum extract by assessing its effect on TNF-α-induced NFkB activation where the activity was measured in terms of luciferase activity of NFkB reporter system assay. Plum extract inhibited NFkB activation in a dose dependent manner (Figure 6). A dose response assay indicated that ~40% inhibition (*p* < 0.05) of TNF-α-mediated NFkB activation was achieved at 25 μg/mL plum extract, and >80% inhibition (*p* < 0.05) of TNF-α-mediated NFkB activation was achieved at 50 μg/mL plum extract.

**Figure 6.** Effects of plum extract on NFkB activation. The effect of plum extract on TNFα-mediated NFkB activation was measured in the A549/NFκB-luc reporter stable cell line. Activity was measured in terms of luciferase activity. The data are reported as the relative percent inhibition of TNFα-mediated NFkB activation. The data are expressed as mean ± SD for at least three experiments. All comparisons were made to control (untreated cells) using one-way ANOVA; significant differences are reported at \* *p* < 0.05.

#### *3.7. E*ff*ect of PE60 Plum Extract on Colon-26 Mouse Adenocarcinoma Cell Line*

When Colon-26 cells were treated with plum extract, the cells viability was reduced in a dose-dependent manner reaching ~80% reduction (*p* < 0.05) at 150 μg/mL. Upon further increasing the concentration of plum extract, the cell viability was further reduced 90% (*p* < 0.05) at 200 μg/mL (Figure 7).

**Figure 7.** Effect of plum extract on Colon-26 adenocarcinoma cells. Data were calculated as % inhibition of cell growth in response to various concentrations of plum extract. The data are expressed as mean ± SD for at least three replicates. All comparisons were made to control (untreated cells) using one-way ANOVA; the significant differences are reported at \* *p* < 0.05.

#### *3.8. Effect of PE60 Plum Extract on C2C12 Cell Viability in Response to Colon-26 Cells Cytotoxicity-Inducing Factors*

μ

Mouse derived Colon-26 adenomacarcinoma cells are known to induce muscle wasting in rodents [40]. The effect of these circulating soluble factors released by Colon-26 was examined on growth of C2C12 myotubules in vitro in presence or absence of plum extract. Figure 8a,b shows that in the absence of plum extract, soluble factors released in media derived from Colon-26 cells caused a significant reduction of C2C12 cell viability by ~25% (*p* < 0.05). However, in the presence of plum extract, the negative effects of Colon-26-derived media on C2C12 viability was prevented and the cell viability was maintained to a similar level that was seen in the untreated cells.

**1RUPDO6HUXP &RORQ6HUXP 1RUPDO0HGLXP &RORQ&RQGLWLRQHG 0HGLXP**

**Figure 8.** The effect of plum extract on C2C12 viability in response to Colon-26-induced cytotoxicity. (**a**) Differentiated C2C12 myotubes were treated with normal medium (i & ii) or Colon-26-conditioned medium (iii & iv) in the absence (i & iii) or presence (ii & iv) of plum extract (50 μg/mL). (**b**) The viability of C2C12 myotubules were determined using WST-1 assay. The data is expressed as mean ± SD for at least three experiments. All comparisons were made to control (untreated cells) using one-way ANOVA; significant differences are reported at \* *p* < 0.05.

#### **4. Discussion**

In our study, we sought to investigate if plums had benefits on skeletal muscle. Specifically, we selected to use a plum extract that was enriched in polyphenols (~60% polyphenols) because the health benefits of plum have been partly attributed to its high polyphenol content [41–43]. Our data indicates that about 95% of total phenolic content in the plum extract used was present in the form of flavonoids. This data is not surprising as fruits are often reported to have phenolic compounds which are high in flavonoids with a range of 90–100% [44]. The anti-oxidant activity in the plum extract was found to be in range of 3–4 mM of Trolox equivalent/g, which is higher than that of turmeric (0.27–0.35 mM Trolox eq/g) and mulberry (1–2 mM Tolox eq/g), but lower than green tea (13–17 mM Trolox eq/g) and pomegranate (20–25 mM Trolox eq/g) [45–49].

Dried plum has previously been reported to have health benefits on bone. In rat models of osteoporosis, dried plum intake resulted in prevention and reversal of bone loss [50,51]. A three-month clinical intervention study showed that dried plum intake improved biomarkers of bone formation in postmenopausal women, whereas longer-term intake of dried plum resulted in mitigating loss of bone mineral density [31]. The present study was designed to analyze the effects of plum extract on muscle metabolism in C2C12 myotubules. In our initial experiments, the effect of plum extract was tested on myoblast viability. The data show that this plum extract has no toxicity on the muscle cells, even at very high doses. These results are consistent with prior literature on plum effects on non-diseased cells [52]. The maintenance of muscle mass is dependent on synthesis of new proteins and breakdown of old or damaged proteins. If these processes are balanced, the muscle mass is maintained; however, with aging and under certain catabolic condition including cancer, renal failure or trauma, muscle protein degradation exceeds the synthesis of new proteins, and results in muscle atrophy [53]. One interesting observation was that that plum extract increased the size of growing myoblast under un-differentiated conditions, suggestive of inducing increase in cytoplasmic volumes by stimulating protein synthesis. We also measured effect of plum extract on protein synthesis and degradation in differentiated myotubules. Our data clearly demonstrated that plum extract not only increased protein synthesis but also inhibited myotubules protein degradation in response to serum starvation, demonstrating both an anabolic and anti-catabolic effect.

The activity of the plum extract appears to be at least partly mediated through IGF-1 stimulation. Several studies have shown that IGFs stimulated both proliferation and differentiation of myoblasts, and also play a role in regenerating damaged skeletal muscle [54–58]. In line with our results, prior studies have also demonstrated that plums can increase IGF-1 levels in both humans [58] and animal models [51,59]. One of the manifestations of muscle loss is associated with decreased production of IGF-1 [60]. The signaling pathway IGF-1/PI3K/Akt (Insulin like growth factor -1/phosphatidyl inositol 3-kinase/protein kinase) is considered the main mediator of normal muscle development and one of the most studied signaling molecular systems involved in muscle metabolism [61]. Akt activation leads to activation of mTOR (mammalian target of rapamycin), which is responsible for promoting protein synthesis. The Akt-mTOR signaling pathway and its downstream components (p70s6k and 4E-BPI) are attenuated with muscle wasting [62]. Further studies need to be performed to confirm the if plum extract is indeed regulating Akt activity. The identification of compound or compounds in plum responsible for stimulating IGF-1 levels in myoblast was beyond the scope of the present study. As discussed earlier, ursolic acid has been shown to increase muscle mass in mice exhibiting fasting-induced muscle atrophy [22] or diet-induced obesity [23]. Interestingly, ursolic acid has also been shown to induce IGF-1 levels in the skeletal muscle of these mice with an increased Akt phosphorylation [22,23]. During present investigation, we were not able to detetect ursolic acid in PE60 extracts due to technical limitation for detecting all polyphenols; however, other studies have reported presence of ursolic acid in plums [21]. Therefore, it is possible that ursolic to some extent may have contributed in IGF-1 mediated muscle growth in our studies.

Studies have demonstrated the anti-inflammatory effect of dried-plum or plum juice in several cellular system including lipopolysaccharide-induced macrophages [63,64], splenocytes from ovariectomized mice [65], colorectal cells in azoxymethane-treated rats [66], heart tissues in obese rats [67] and joints of TNF-over expressing mice [25]. The antioxidation activities of plum appeared to be mediated through the inhibition of NFκB activation [25,66,67]. Based on these reported studies, we decided to test the effect of plum extract on NFκB activation, since oxidative stress and inflammatory responses through activation of NFκB play an important part in muscle atrophy. Activation of NFkB plays a central role in muscle atrophy in several catabolic situations including cancer cachexia [68,69]. We found that even a small dose of plum extract was able to almost completely inhibit (>80% inhibition) TNF-α-induced NFκB activity *in vitro*. It is likely that the proanthocyanidins, which comprise over 70% of the polyphenols, may be involved in suppressing the inflammatory cytokine (TNF)-induced activation of NFκB, although this has not been systematically tested with the individual components of the extract.

Cancer cachexia-related morbidity and mortality are often accompanied by whole body and muscle loss [4,7,8] and it is suggested that blocking muscle wasting can prolong life despite tumor growth [10]. The effect of plum extract on colon cancer cell viability, as well as its ability to protect muscle cells from colon-cancer cell induced cytotoxicity, were, therefore, also investigated. We used Colon-26 adenocarcinoma cells, which is a widely used preclinical model because it induces clinical cachexia, including its development as well as the resultant physiological and metabolic impairment [40,69,70]. Treating the Colon-26 colon cancer cells with plum extract caused a significant decrease in the Colon-26 cell's viability, indicating potential anti-tumor activity.

It is known that muscle wasting in cancer patients is mediated through factors released from tumor in circulation [71–74]. Studies have shown that elevated circulating levels of IL-6 mediated skeletal muscle cell death in severely cachectic mice with colon cancer [75]. Our studies found that plum extract can protect C2C12 myotubules from cytotoxicity induced by soluble factors released by the Colon-26 cells. The exact pathways leading to reduced cell viability in response to tumor induced soluble factors are not known, but it is possible that both atrophy and apoptosis may be attenuated by the plum extract. It is also possible that compound(s) in plum extract may directly affect colon cells to inhibit secretion of inflammatory cytokines. Future studies need to be conducted to elucidate the molecular mechanism involved in the anti-cytotoxic activity of the plum extract.

Our current studies have several limitations. The study was performed using an *in vitro* system that may not represent the complexities of an *in vivo* system. Furthermore, polyphenols in the plum extract can undergo biotransformation in vivo, which could either enhance or diminish the anabolic of plum extract on muscle as well as its anti-inflammatory benefit. However, previous human studies with dried plum still demonstrated its ability to activate IGF-1 as well as its anti-inflammatory benefits, indicating that biotransformation may not result in loss of these effects observed in our study.

#### **5. Conclusions**

In conclusion, the polyphenol-enriched plum extract has both anti-catabolic and anabolic effects on muscle cells, as well as myogenic potential. In addition, this plum extract exhibited anti-cytotoxic properties in response to soluble factors released from cancer cells. Thus, plum extract may be a useful intervention to be considered for cancer cachexia or other chronic disease-induced cachexia involving inflammation. These results need to be confirmed in an animal model of cachexia, followed by clinical translation.

**Author Contributions:** Conceptualization: R.A.S., and S.L.P.; Experimentation and Data collection: F.A.A., H.L., P.K., P.W.J.; Manuscript writing: R.A.S., P.W.J., and S.L.P.

**Funding:** The study was partly funded by a grant from Abbott-Nutrition, Columbus, OH, and by Evans-Allen grant from USDA.

**Acknowledgments:** The financial support to Faten A. Alsolmei was provided by Saudi Arabian Culture Mission, Kingdom of Saudi Arabia.

**Conflicts of Interest:** Suzette Pereira and Paul Johns are currently employed by Abbott. Padmavathy Krishnan is an ex-Abbott employee. Rafat Siddiqui, Haiwen Li, and Faten Alsolmei have no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

### **Rectal and Vaginal Eradication of** *Streptococcus agalactiae* **(GBS) in Pregnant Women by Using** *Lactobacillus salivarius* **CECT 9145, A Target-specific Probiotic Strain**

**Virginia Martín 1,†, Nivia Cárdenas 2,‡, Sara Ocaña 2,3 , María Marín 1, Rebeca Arroyo 1, David Beltrán 4, Carlos Badiola 5, Leónides Fernández <sup>2</sup> and Juan M. Rodríguez 1,\***


Received: 13 March 2019; Accepted: 5 April 2019; Published: 10 April 2019

**Abstract:** *Streptococcus agalactiae* (Group B Streptococci, GBS) can cause severe neonatal sepsis. The recto-vaginal GBS screening of pregnant women and intrapartum antibiotic prophylaxis (IAP) to positive ones is one of the main preventive options. However, such a strategy has some limitations and there is a need for alternative approaches. Initially, the vaginal microbiota of 30 non-pregnant and 24 pregnant women, including the assessment of GBS colonization, was studied. Among the *Lactobacillus* isolates, 10 *Lactobacillus salivarius* strains were selected for further characterization. In vitro characterization revealed that *L. salivarius* CECT 9145 was the best candidate for GBS eradication. Its efficacy to eradicate GBS from the intestinal and vaginal tracts of pregnant women was evaluated in a pilot trial involving 57 healthy pregnant women. All the volunteers in the probiotic group (*n* = 25) were GBS-positive and consumed ~9 log10 cfu of *L. salivarius* CECT 9145 daily from week 26 to week 38. At the end of the trial (week 38), 72% and 68% of the women in this group were GBS-negative in the rectal and vaginal samples, respectively. *L. salivarius* CECT 9145 seems to be an efficient method to reduce the number of GBS-positive women during pregnancy, decreasing the number of women receiving IAP during delivery.

**Keywords:** Lactobacillus salivarius; Streptococcus agalactiae; GBS; probiotic; pregnancy

#### **1. Introduction**

Neonatal sepsis contributes substantially to neonatal morbidity and mortality and is a major global public health challenge worldwide. According to the age of onset, neonatal sepsis is divided into early-onset sepsis (EOS) and late-onset sepsis (LOS). EOS has been variably defined based on the age at onset, with bacteremia or bacterial meningitis occurring at ≤72 h in infants hospitalized in the neonatal intensive care unit versus <7 days in term infants, and usually reflects transplacental or ascending infections from the maternal genitourinary tract [1].

*Streptococcus agalactiae* (Group B Streptococci, GBS) is one of the microorganisms most frequently involved in severe neonatal EOS cases [2–4]. Women, men and children of all ages can be asymptomatically colonized with GBS, acting the gastrointestinal tract, vagina and urethra as reservoirs. A recent systematic review and meta-analyses found that adjusted estimate for maternal GBS colonization worldwide was 18% (95% confidence interval [CI], 17%–19%), with regional variations (11%–35%) [5]. GBS vaginal and/or intestinal colonization is considered as a risk factor for ascending infection during pregnancy [6].

The relevance of GBS as an agent of neonatal infections soon prompted the finding of strategies for its eradication from the intestinal and genitourinary mucosal surfaces of pregnant women [7], including the use of chlorhexidine, which showed no effect [8] and, particularly, the development of vaccines. Unfortunately, no GBS vaccine is available at present despite the strong research efforts made in the last decades [9]. At present, two main approaches have been recommended for the prevention of neonatal GBS infections in Western countries: (a) a risk-based strategy; and (b) a screening-based strategy [10]. The second approach, involving recto-vaginal GBS screening at week 35–38 of pregnancy and subsequent intrapartum antibiotic prophylaxis (IAP) to positive mothers, is the preventive option followed in the USA and some European countries.

However, such a strategy also faces some limitations: (a) it does not guarantee GBS eradication [11]; (b) it does not prevent GBS-related abortions, stillbirths and preterm births [4]; (c) it may lead to increasing rates of antibiotic resistance among GBS and other clinically relevant microorganisms [12–14]; and (d), it has a very negative impact on the acquisition, composition and development of the infant microbiota. Perinatal antibiotic use affects the gut microbiota development during the critical first weeks of life [15,16]. The composition of the gut microbiota of neonates whose mothers received IAP has been described as aberrant in comparison with that of non-treated neonates [17,18]. The detrimental impact of perinatal antibiotics, mainly IAP, on early life microbiota may have a lasting effect on the host's health [19]. Therefore, there is a need for alternative strategies to avoid GBS colonization during pregnancy.

In this context, the objective of this work was, first, the assessment of the presence of GBS in the vaginal exudate of healthy pregnant and non-pregnant women; and, second, the selection of a safe probiotic strain with the ability to eradicate GBS from the intestinal and genitourinary tracts of pregnant women.

#### **2. Material and Methods**

#### *2.1. Microbiological Analysis of Vaginal Swabs Obtained from Pregnant and Non-pregnant Women*

A total of 54 women (30 non-pregnant women and 24 pregnant women), aged 25–35, participated in this part of the study. In accordance with the Declaration of Helsinki, all volunteers gave written informed consent to the protocol, which had been approved (protocol 10/017-E) by the Ethical Committee of Clinical Research of the Hospital Clínico San Carlos Madrid (Spain). In relation to non-pregnant women, 4 vaginal exudates samples were collected within a menstrual cycle (days 0, 7, 14 and 21). Pregnant women provided a single sample in week 35–37 of pregnancy. All women claimed to be completely healthy.

Samples were diluted in peptone water and spread onto Columbia Nalidixic Acid (CNA), Mac Conkey (MCK), Sabouraud Dextrose Chloramphenicol (SDC), Gardnerella (GAR) and Mycoplasma agar plates (BioMerieux, Marcy l'Etoile, France) for selective isolation and quantification of the main agents involved in vaginal infections. They were also spread onto agar plates of MRS (Oxoid, Basingstoke, UK) supplemented with either L-cysteine (2.5 g/L) (MRS-C) or horse blood (5%) (MRS-B) for isolation of lactobacilli. All the plates were incubated for 48 h at 37 ◦C in aerobic conditions, with the exception of the MRS-C and MRS-B ones, which were incubated anaerobically (85% nitrogen, 10% hydrogen, 5% carbon dioxide) in an anaerobic workstation (DW Scientific, Shipley, UK). Parallel, all the

samples were submitted to an enrichment step in Todd Hewitt broth (Oxoid) to facilitate the isolation of *S. agalactiae* in CNA plates.

Initially, identification of the bacterial strains (at least one isolate of each colony morphology per medium and per sample) was performed by 16S rDNA sequencing using the primers and PCR conditions described by Kullen et al. [20]. Sequencing reactions were prepared using the ABI PRISM® BigDye™ Terminator Cycle Sequencing kit with AmpliTaq DNA polymerase according to the manufacturer's instructions (Applied Biosystems, Foster City, CA, USA) and were run on an ABI 377A automated sequencer (Applied Biosystems). The resulting sequences were used to search sequences deposited in the EMBL database using the BLAST algorithm. The identity of the strain was determined on the basis of the highest (>98%) scores.

Identification of yeasts and confirmation of the initial 16S rDNA-based bacterial identifications was performed by MALDI-TOF (VITEK MS, BioMerieux, Marcy-L'Étoile, France) [21]. Identification of *S. agalactiae* isolates was also confirmed by using a latex agglutination test (Streptococcal grouping kit, Oxoid, Basingstoke, UK), following the instructions of the manufacturer.

Those isolates identified as belonging to the genus *Lactobacillus* were preserved for further studies. For such a purpose, an MRS-C broth culture of each isolate was mixed with glycerol (30%, v/v) and kept at −80 ◦C until required. A total of 89 different *Lactobacillus* strains were isolated from the vaginal swabs and submitted to the Random Amplification of Polymorphic DNA (RAPD) genotyping as described [22] in order to avoid duplication of isolates. Among them, 10 *Lactobacillus salivarius* strains were selected for further characterization on the basis of the following criteria: (1) absence of *S. agalactiae*, *Gardnella vaginalis, Candida spp., Ureaplasma* spp. and *Mycoplasma spp* in the vaginal samples from which the lactobacilli were originally isolated; (2) Qualified Presumption of Safety (QPS) status conceded by EFSA; and (3) the ability of the strain to grow rapidly in MRS broth under aerobic conditions (≥<sup>1</sup> ×106 cfu/mL after 16 h at 37 ◦C).

#### *2.2. Antimicrobial Activity of the Lactobacilli Strains against GBS*

Initially, an overlay method [23] was used to determine the ability of the lactobacilli strains to inhibit the growth of 12 different *S. agalactiae* strains. Among them, 6 strains had been isolated from blood or cerebrospinal fluid in clinical cases of neonatal sepsis (Hospital Universitario Ramón y Cajal, Madrid, Spain) while the remaining 6 had been isolated from vaginal samples of pregnant women (our own collection). It was performed using MRS agar plates, on which the lactobacilli strains were inoculated as approximately 2 cm-long lines and incubated at 37 ◦C for 48 h. The plates were then overlaid with the indicator *S. agalactiae* strains vehiculated in 10 mL of Brain Heart Infusion (BHI, Oxoid) broth supplemented with soft agar (0.7%), at a concentration of ~10<sup>4</sup> colony-forming units (cfu)/mL. The overlaid plates were incubated at 37 ◦C for 48 h and, then, examined for clear zones of inhibition (>2 mm) around the lactobacilli streaks. All experiments assaying inhibitory activity were performed in triplicate.

#### *2.3. Production of Specific Antimicrobials (Bacteriocins, Lactic Acid, Hydrogen Peroxide) by the Lactobacilli Strains*

Bacteriocin production was assayed using an agar diffusion method as described by Dodd et al. [24] and modified by Martín et al. [25], using the *S. agalactiae* strains as the indicator bacteria employed for the overlay method. The lactobacilli strains were screened for hydrogen peroxide production following the procedure described by Song et al. [26]. In the case of positive strains, hydrogen peroxide production was also measured by the quantitative method of Yap and Gilliland [27]. The concentration of L- and D-lactic acid in the supernatants of MRS cultures of the lactobacilli strains was quantified using an enzymatic kit (Roche Diagnostics, Mannheim, Germany), following the manufacturer's instructions. The pH values of the supernatants were also measured. All these assays were performed in triplicate and the values were expressed as the mean ± SD.

#### *2.4. Coaggregation and Co-culture Assays*

The ability of the lactobacilli strains to aggregate with cells of the *S. agalactiae* strains was investigated following the procedure of Younes et al. [28]. The suspensions were observed under a phase-contrast microscope after Gram staining.

To test the anti*-S. agalactiae* activity of the lactobacilli in a broth assay format, tubes containing 20 mL of MRS broth were co-inoculated with 1 mL of a *Lactobacillus* strain culture (7 log10cfu/mL) and 1 mL of an *S. agalactiae* strain (7 log10 cfu/mL). Subsequently, the cultures were incubated for 6 h at 37 ◦C in aerobic conditions. Immediately after the co-inoculation and after the incubation period, aliquots were collected, serially diluted and plated on MRS-C plates and CHROMagar StrepB agar plates (CHROMagar, París, France) for the selective enumeration of lactobacilli and streptococci, respectively. Correct taxonomic assignment was confirmed by the MALDI-TOF analysis as described previously.

#### *2.5. Survival After In Vitro Exposure to Saliva and Gastrointestinal-Like Conditions*

The survival of the strain to conditions resembling those found in the human digestive tract (saliva, human stomach and small intestine) was assessed in the in vitro system described by Marteau et al. [29], with the modifications reported by Martín et al. [25]. For this purpose, the strain was vehiculated in UHT-treated milk (25 mL) at a concentration of 109 CFU/mL. The values of the pH curve in the stomach-like compartment were those recommended by Conway et al. [30]. Different fractions were taken at 20, 40, 60, and 80 min from this compartment, and exposed for 120 minutes to a solution with a composition similar to that of human duodenal juice [30]. The survival rate of the strain was determined by culturing the samples on MRS agar plates, which were incubated at 37 ◦C for 48 h.

#### *2.6. Adhesion to Caco-2, HT-29 and Vaginal Cells and to Mucin*

The ability of the strains to adhere to HT-29 and Caco-2 cells was evaluated as described by Coconnier et al. [31] with the modifications reported by Martín et al. [25]. HT-29 and Caco-2 were cultured to confluence in 2 mL of DMEM medium (PAA, Linz, Austria) containing 25 mM of glucose, 1 mM of sodium pyruvate and supplemented with 10% heat-inactivated fetal calf serum, 2 mM of L-glutamine and 1% non-essential amino acid preparation. At day 10 after confluence, 1 mL of the medium was replaced with 1 mL of DMEM containing 108 CFU/mL of the strains. Adherence was measured as the number of lactobacilli adhered to the cells in 20 random microscopic fields. The assay was performed by triplicate.

Adherence to vaginal epithelial cells collected from healthy premenopausal women was performed as described previously [32].

The adhesion of the lactobacilli strains to mucin was determined according to the method described by Cohen and Laux [33].

#### *2.7. Sensitivity to Antibiotics*

The sensitivity of the strains to antibiotics was tested using the lactic acid bacteria susceptibility test medium (LSM) [34] and the microtiter VetMIC plates for lactic acid bacteria (National Veterinary Institute of Sweden, Uppsala, Sweden), as described previously [35]. Parallel, minimum inhibitory concentrations (MICs) were also determined by the E-test [AB BIODISK, Solna, Sweden) following the instructions of the manufacturer. Results were compared to the cut-off levels proposed by the European Food Safety Authority [36].

#### *2.8. Hemolysis, Formation of Biogenic Amines and Degradation of Mucin*

For investigation of hemolysis, strains were streaked onto layered fresh horse blood agar plates and grown for 24 h at 37 ◦C. Zones of clearing around colonies indicated hemolysin production. The capacity of the strains to synthesize biogenic amines (tyramine, histamine, putrescine and cadaverine) from their respective precursor amino acids (tyrosine, histidine, ornithine and lysine; Sigma-Aldrich) was evaluated using the method described by Bover-Cid and Holzapfel [37]. The potential of the strains to degrade gastric mucine (HGM; Sigma) was evaluated in vitro as indicated by Zhou et al. [38].

#### *2.9. Acute and Repeated Dose (4-Weeks) Oral Toxicity Studies in a Rat Model*

Wistar male and female rats (Charles River Inc., Marget, Kent, UK) were used to study the acute and repeated dose (4-weeks) oral toxicity of *L. salivarius* CECT 9145 in a rat model. Acclimation, housing and management (including feeding) of the rats was performed as previously described [39]. The rats were 56-days old at the initiation of treatment. Acute (limit test) and repeated dose (4 weeks) studies were conducted in accordance with the European Union guidelines (EC Council Regulation No. 440), and authorized by the Ethical Committee on Animal Research of the Complutense University of Madrid (protocol 240111).

In the acute (limit test) study, 24 rats (12 males, 12 females) were distributed into two groups of 6 males and 6 females each. After an overnight of fasting, each rat received skim milk (500 μL) orally (control group or Group 1), or a single oral dose of 1 × <sup>10</sup><sup>10</sup> CFU of *L. salivarius* CECT 9145 dissolved in 500 μL of skim milk (treated group or Group 2). Doses of the test and control products were administered by gavage. At the end of a 14 days observation period, the rats were weighed, euthanized by CO2 inhalation, exsanguinated, and necropsied.

The repeated dose (4 weeks) (limit test) study was conducted in 48 rats (24 males, 24 females) divided in four groups of 6 males and 6 females each (control group or Group 3; treated group or Group 4; satellite control group or Group 5; and satellite treated group or Group 6). Rats received a daily oral dose of either skim milk (Groups 3 and 5) or 1 × <sup>10</sup><sup>9</sup> CFU of *L. salivarius* CECT 9145 dissolved in 500 μL of skim milk (Groups 4 and 6) for 4 weeks. All rats of Groups 3 and 4 were deprived of food for 18 h, weighed, euthanized by CO2 inhalation, exsanguinated, and necropsied on Day 29. All animals of the satellite groups (Groups 5 and 6) were kept a further 14 days without treatment to detect the delayed occurrence, persistence or recovery from potential toxic effects. All rats of the Groups 5 and 6 were deprived of food for 18 h, weighed, euthanized by CO2 inhalation, exsanguinated, and necropsied on day 42.

Behavior and clinical observations, blood biochemistry and hematology analysis, organ weight ratios and histopathological analysis were carried as described previously [39]. Bacterial translocation to blood, liver or spleen, and total liver glutathione (GSH) concentration was evaluated following the methods described by Lara-Villoslada et al. [40].

#### *2.10. Efficacy of L. salivarius CECT 9145 to Eradicate GBS from the Intestinal and Vaginal Tracts of Pregnant Women: A Pilot Clinical Trial*

In this prospective pilot clinical assay, 57 pregnant women (39 rectal and vaginal GBS-positive women; 18 rectal and vaginal GBS-negative women at the start of the intervention), aged 25–36, participated in this study. All met the following criteria: a normal pregnancy and a healthy status. Women ingesting probiotic supplements or receiving antibiotic treatment in the previous 30 days were excluded. Women with lactose intolerance or a cow's milk protein allergy were also excluded because of the excipient used to administer the strain. All volunteers gave written informed consent to the protocol (10/017-E), which had been approved by the Ethical Committee of Clinical Research of the Hospital Clínico San Carlos Madrid (Spain).

Volunteers were distributed into 3 groups (1 probiotic group and 2 placebo groups). All the volunteers in the probiotic group (*n* = 25) were GBS-positive and consumed a daily sachet with ~50 mg of freeze-dried probiotic (~9 log10 cfu of *L. salivarius* CECT 9145) from week 26 to week 38 of the pregnancy. Placebo subgroup 1 (*n* = 14) included GBS-positive women (pregnancy week ranging from 19 to 30) that were going to receive IAP because they had a previous baby that suffered a GBS sepsis. Placebo subgroup 2 (*n* = 18) included GBS-negative women (pregnancy week ranging from 14 to 26). Women in both placebo subgroups received a daily sachet containing 50 mg of the excipient used to carry the probiotic strain. In all cases, the intervention lasted from the start of the intervention to week

38. Probiotic- and excipient-containing sachets were kept at 4 ◦C throughout the study. All volunteers were provided with diaries to record compliance with the study product intake. Minimum compliance rate (% of the total treatment doses) was set at 86%.

Recto-vaginal GBS screening was performed at 28, 32 and 38 weeks. Rectal and vaginal exudates samples collected during the trial were serially diluted and plated on Granada (Biomerieux; isolation of hemolytic GBS, which appear as orange colonies), and CHROMagar StrepB (CHROMagar; for isolation of hemolytic and non- hemolytic GBS, which appear as purple colonies) agar plates. To avoid sensitivity-related problems, samples were submitted to a GBS enrichment step in Todd-Hewitt broth (Oxoid). After 24 h at 37 ◦C, the broth cultures were spread on CHROMagar agar plates. Correct taxonomic assignment was confirmed by MALDI-TOF and latex agglutination analyses, as described previously. At the last sampling time (week 38), recto-vaginal GBS screening was performed not only in our laboratory but also in those of the hospitals in which the respective women were going to deliver their babies.

Microbiological data were recorded as CFU/mL and transformed to logarithmic values before statistical analysis. Two-way ANOVA was used to investigate the effect of the individual (woman) and sampling time on the semiquantitative *S. agalactiae* counts in vaginal swabs. Statistical significance was set at *P* < 0.05. Statgraphics Centurion XVI version 17.0.16 (Statpoint Technologies Inc, Warrenton, Virginia) was used to carry out statistical analyses.

#### **3. Results**

#### *3.1. Microbiological Analysis of Vaginal Swabs Obtained from Pregnant and Non-Pregnant Women*

Bacterial growth was detected in all the samples when they were inoculated on MRS (2.70–8.08 log10 colony-forming units (cfu); mean 5.36 log10 cfu); CNA (3.00–7.92 log10 cfu; mean 5.13 log10 cfu) and GAR (2.70–8.10 log10 cfu; mean 5.24 log10 cfu) agar plates. Similar bacterial groups grew in these three media. Growth on MCK, SDC or Mycoplasma plates was only detected in a few percentages of samples (from 0% in Mycoplasma plates to ~40% in SDC plates).

*S. agalactiae* could be isolated from both non-pregnant (~25%) and pregnant (~19%) women. *Candida albicans* and other yeasts were isolated from approximately 7 and 36% of the non-pregnant and pregnant women, respectively. *Gardnerella vaginalis* was isolated in ~7% of the pregnant women. In both groups, *Lactobacillus* was the dominant genus since it was detected in ~93% of the participating women.

In relation to the samples provided by non-pregnant women, a total of 433 isolates (including at least one representative of each colony and cell morphology) were submitted to taxonomical analyses. The highest number of isolates corresponded to the genus *Lactobacillus* (28% of the total isolates), followed by *Staphylococcus* (17%)*, Enterococcus* (11%)*, Corynebacterium* (7%), and *Streptococcus* (4%). Among the *Lactobacillus* isolates, the main species were *L. gasseri* (24%), *L. crispatus* (23%), *L. salivarius* (21%), *L. vaginalis* (12%), *L. plantarum* (13%), *L. coleohominis* (5%), and *L. jensenii* (2%). Isolates belonging to the species *L. crispatus*, *L. gasseri*, *L. salivarius*, *L. vaginalis*, and *L. plantarum* could be isolated from the 4 phases of the menstrual cycle sampled in this study.

From the samples provided by pregnant women, 120 isolates were submitted to taxonomical analyses. Again, the genus *Lactobacillus* was associated to the highest number of isolates (17%), followed by *Staphylococcus* (15%)*, Streptococcus* (8%), yeasts (8%), *Enterococcus* (5%), *Bifidobacterium* (3%) and *Corynebacterium* (1%). Among the *Lactobacillus* isolates, the main species were *L. gasseri* (41%), *L. casei* (19%), *L. salivarius* (16%), *L. fermentum* (8%), *L. vaginalis* (6%), *L. reuteri* (5%) and *L. jensenii* (5%).

Among the *Lactobacillus* isolates obtained in this study, a few were selected to evaluate their potential as probiotics to control GBS populations on the basis of the following criteria: (1) absence of *S. agalactiae*, *Gardnella vaginalis, Candida* spp., *Ureaplasma* spp., and *Mycoplasma* spp. in the vaginal samples from which the lactobacilli were originally isolated; (2) Qualified Presumption of Safety (QPS) status (European Authority of Food Safety, EFSA); and (3) ability of the strain to grow rapidly in MRS broth under aerobic conditions (≥<sup>1</sup> ×106 cfu/mL after 16 h at 37 ◦C). In fact, only 10 strains (V3III-1, V4II-90, V7II-1, V7II-62, V7IV-1, V7IV 60, V8III-62, V11I-60, V11III-60 y, V11IV-60) met all the criteria and all of them belonged to the same species (*Lactobacillus salivarius*). These strains were then selected for further characterization. Later, *L. salivarius* V4II-90 was deposited in the Spanish Collection of Type Cultures (CECT) as *L. salivarius* CECT 9145 and, therefore, this is the name used for this strain in this article.

#### *3.2. Antimicrobial Activity of the Lactobacilli Strains Against GBS and the Production of Potential Antimicrobial Compounds*

Initially, the antimicrobial activity of the 10 selected lactobacilli against the *S. agalactiae* strains was determined by an overlay method. Clear inhibition zones (ranging from 2 to 20 mm) were observed around the lactobacilli streaks.

In relation to the antimicrobial compounds that may be responsible for such activity, the concentration of L- and D-lactic acid and the pH of the supernatants obtained from MRS cultures of the lactobacilli are shown in Table 1. The global concentration of L-lactic acid was similar (~10 mg/mL) in all the supernatants. In contrast, D-lactic acid was not detected in the supernatants of the tested strains. In addition, all the strains acidified the MRS-broth medium to a final pH of ~4 after 16 h of incubation; among them, *L. salivarius* V7IV-1 showed the highest acidifying capacity (final pH of 3.8). No bacteriocin-like activity could be detected against the tested *S. agalactiae* strains. Two strains (*L. salivarius* CECT 9145 and V7IV-1) were able to produce hydrogen peroxide (7.29 μg/mL ± 0.69 and 7.46 μg/mL ± 0.58, respectively) (Table 1).

**Table 1.** The pH and concentrations of L- and D-lactic acid (mg/mL; mean ± SD), and hydrogen peroxide (μg/mL; mean ± SD) in the supernatants obtained from the MRS cultures of the lactobacilli (*n* = 4).


The initial pH value of MRS broth was 6.2. Nd: not detectable.

The capacity of the lactobacilli strains to form large well-defined co-aggregates with *S. agalactiae* was strain-dependent. Strains V3III-1, V7IV-60 and V11IV-60 coaggregated with 5 *S. agalactiae* strains; strains V8III-62, V11I-60 and V11III-60 with 7; strain V7II-62 with 9 *S. agalactiae* strains; and strains CECT 9145, V7II-1 and V7IV-1 with 10 *S. agalactiae* strains (Figure 1). The ability of the lactobacilli strains to interfere or inhibit the growth of four *S. agalactiae* strains was evaluated using MRS broth co-cultures. Co-cultures with *S. agalactiae* seemed not to affect the growth of any of the *L. salivarius* strains (Table 2). In contrast, most of the *L. salivarius* strains were able to interfere at a higher or lower degree with the growth of the different *S. agalactiae* strains included in this assay. Among them, *L. salivarius* CECT 9145 showed the highest ability to inhibit the growth of *S. agalactiae* since the presence of two of the four *S. agalactiae* strains was not detectable in the co-cultures and the concentration of the other two showed a ~2.5 log10 decrease after an incubation period of only 6 h at 37 ºC (Table 2). Interestingly, no viable streptococci could be detected when the co-cultures were incubated for 24 h (Table 2).


**Table 2.** The bacterial counts (log10 cfu/mL) of the *S. agalactiae* strains when co-cultured with the *L. salivarius* strains in MRS broth for 0, 6 and 24 h at 37 ◦C.

Nd: *S. agalactiae* was not detected.

**Figure 1.** The strong co-aggregation between *L. salivarius* CECT 9145 (rods) and an *S. agalactiae* strain (cocci chains).

#### *3.3. Survival After In Vitro Exposure to Saliva and Gastrointestinal-Like Conditions*

The viability of the strains after exposition to conditions simulating those found in the gastrointestinal tract varied from ~64% (*L. reuteri* CR20, *L. salivarius* CECT 9145) to 30% (*L. salivarius* V3III-1) (Table 3).


**Table 3.** The percentage (%) of initial lactobacilli (9 log10 cfu/mL) that survived to conditions simulating those of the gastrointestinal tract.

*\*,* different letters mean statistically different values.

#### *3.4. Adhesion to Caco-2, HT-29 and Vaginal Cells and to Mucin*

In this study, the lactobacilli strains tested were strongly adhesive to both Caco-2 and HT-29 cells, with the exception of the negative control strain (*L. casei imunitas*) which showed a low adhesive potential (Table 4). In addition, all showed adhesion to vaginal epithelial cells. Among the *L. salivarius* strains, *L. salivarius* CECT 9145 globally displayed the highest ability to adhere to both intestinal and vaginal epithelial cells (Table 4). The lactobacilli strains tested showed a variable ability to adhere to porcine mucin (Table 4). *L. salivarius* CECT 9145 and *L. salivarius* V7IV-1 were the strains that showed the highest adherence ability.

#### *3.5. Sensitivity to Antibiotics*

The MIC values of the lactobacilli strains for 16 antibiotics assayed are shown in Table 5. All the strains were sensitive to most of the antibiotics tested, including those considered clinically relevant antibiotics such as gentamycin, tetracycline, clindamycin, chloramphenicol, and ampicillin, showing MICs equal to or lower than the breakpoints defined by EFSA (EFSA, 2018). All the strains were resistant to vancomycin and kanamycin, which is an intrinsic property of the *L. salivarius* at the species level.

**Table 4.** The ability of the lactobacilli to adhere to HT-29, Caco-2 and vaginal epithelial cells, and to porcine mucin.


<sup>a</sup> The adherent lactobacilli in 20 random microscopic fields were counted for each test (*n* = 4). <sup>b</sup> Semiquantitative scale: 0, no adhesion; +, low adhesion; ++, middle adhesion; +++, high adhesion. <sup>c</sup> Values are expressed as the percentage of the fluorescence retained in the wells after the washing steps of the assay.

#### *3.6. Hemolysis, the Formation of Biogenic Amines and the Degradation of Mucin*

The strains did not show the ability to produce biogenic amines, and they were neither hemolytic nor able to degrade gastric mucin *in vitro*.

#### *3.7. Acute and Repeated Dose (4 Weeks) Oral Toxicity Studies in a Rat Model*

All animals survived both oral toxicity trials. The development of the treated animals during the experimental periods corresponded to their species and age. There were no significant differences in body weight or body weight gain among groups treated with *L. salivarius* CECT 9145 (including the satellite ones) in comparison to the control groups at any time point of the experimental period. No abnormal clinical signs, behavioral changes, body weight changes, hematological and clinical chemistry parameters, macroscopic or histological findings, or organ weight changes were observed. There were no statistical differences in body weights among groups. Similarly, no statistically significant differences in body weight gain, food and water consumption were observed between the groups. No significant differences in liver GSH concentration were observed between the control and treated groups (9.54 ± 1.21 vs. 9.37 ± 1.39 mmol/g, *P* > 0.1). *L. salivarius* CECT 9145 could be isolated from colonic material and vaginal swabs samples of all the treated animals (probiotic groups) at the end of the treatment. The concentration oscillated between 5.39 and 8.85 log10 cfu/g of the colonic material, and between 3.34 and 6.14 log10 cfu/swab in the vaginal samples. The strain could not be detected in any sample from the placebo group.

#### *3.8. The Efficacy of L. salivarius CECT 9145 to Eradicate GBS from the Intestinal and Vaginal Tracts of Pregnant Women: A Pilot Clinical Trial*

At the inclusion in the study, GBS was detected in both rectal and vaginal swabs obtained from 39 women, out of a total of 57 participating women, while the rest of the women (*n* = 18) were GBS-negative (Table 6). This last group of GBS-negative women, who did not ingest the *L. salivarius* strain also had negative GBS cultures from rectal and vaginal swabs taken regularly at 28, 32 and 36–38 weeks (Table 6). A group of GBS-positive women at the start of the study (*n* = 14) did not receive the probiotic and the routine screening results for vaginal and rectal GBS at 28, 32 and 36–38 weeks were found to be all positive (Table 6).


**Table 5.** The minimal inhibitory concentration (MIC, mg/mL) values of 16 antibiotics a to the*L. salivarius* strains.

a Abbreviations: GEN, gentamycin; KAN, kanamycin; STP, streptomycin; NEO, neomycin; TET, tetracycline; ERY, erythromycin; CLI, clindamycin; CHL, chloramphenicol; AMP, ampicillin; PEN, penicillin; VAN, vancomycin; VIR, virginiamycin; LIN, linezolid; TRM, trimethoprim; CIP, ciprofloxacin; RIF, rifampicin; nr, not required by EFSA. R, the species *L. salivarius* is intrinsically resistant. b Breakpoint: microbiological breakpoints (mg/mL) that categorise *Lactobacillus salivarius* as resistant (microbiological breakpoints are defined as the MIC values that clearly deviate from those displayed by the normal susceptible populations; EFSA, 2018).



*Nutrients* **2019** , *11*, 810

Significantly, the group of GBS-positive women that started using the probiotic (9 log10 cfu/daily) since they were enrolled in this study (from 26 weeks) also tested positive for GBS at 28 weeks, but an increasing number of GBS-negative results appeared in the successive swabs collected until delivery (Table 6). At 30 weeks, the culture of rectal swabs taken from four women of this group rendered a negative result and the number of these samples increased to 18 (72% of the participants) at 38 weeks. Similar results were obtained by culturing vaginal swabs obtained from this group, although the proportion of women testing negative for GBS were always slightly higher when analyzing the rectal swabs than in vaginal swabs (Table 6).

The estimation of the concentration of GBS in vaginal swabs taken regularly up to the delivery from all participants is shown in Figure 2. There were no significant changes in both GBS-negative women (*n* = 18) and GBS-positive women (*n* = 14) without oral administration of *L. salivarius* CECT 9145 regarding the semiquantitative estimation of GBS. However, the number of vaginal swabs where GBS could not be detected increased in successive sampling times in the group that initially tested positive for GBS taking 9 log10cfu of *L. salivarius* CECT 9145 (*n* = 25). The mean value for *S. agalactiae* counts decreased significantly with the administration time *of L. salivarius* CECT 9145 (Figure 2) from a mean value of 5.14 cfu/mL at 26 weeks (*n* = 25) to 3.80 cfu/mL at 38 weeks (*n* = 9) (Figure 2).

**Figure 2.** The mean concentrations (CFU/mL) of *S. agalactiae* (GBS) in vaginal swabs sampled regularly up to the delivery from Group B Streptococci (GBS)-positive women taking 9 log10 cfu of *L. salivarius* CECT 9145 daily. Statistically significant differences between samples taking at different sampling times are indicated by letters (Bonferroni post-hoc test).

No adverse effects arising from the intake of *L. salivarius* CECT 9145 were reported by any of the women who participated in this study. The results of the GBS status obtained in our laboratory at week 38 were identical to those obtained in the hospitals were the recruited women were screened for GBS and, as a result, none of the women who became GBS-negative in this study received IAP.

#### **4. Discussion**

In this work, the GBS colonization rates were 25% and 20% among non-pregnant and pregnant women, respectively. In pregnant women, GBS colonization is found in up to 30% of rectovaginal samples [2,41] and stable colonization with the same clone for several years has been demonstrated [41]. Previous studies have shown that the presence of GBS is not linked to an abnormal microbiome or a reduction of the predominant *Lactobacillus* genus in the vaginal tract of the mother [42–44].

In contrast, a study involving a low number of participants found significant taxonomic differences in stools of 6-month infants, when mothers were GBS carriers, as compared to non-carriers [45]. Anyway, there is no epidemiological evidence for a correlation between neonatal colonization with GBS and specific shifts in the maternal intestinal or vaginal microbiome.

In the USA and many other countries (including Spain), women are routinely screened in the late third trimester (between 35 and 37 weeks' gestation) for GBS colonization by rectovaginal swabs and subsequent cultures. If the rectovaginal swab is culture-positive, or if the patient has GBS in the urine, or has a prior history of GBS perinatal infection, intrapartum prophylactic antibiotics are administered to prevent vertical transmission of GBS to the neonate during labor and delivery. Some European countries (e.g., UK) have not adopted the GBS screening program but, instead, administer antibiotics upon the development of a risk factor for GBS neonatal disease (e.g., prolonged rupture of membranes). However, none of these approaches has eliminated neonatal GBS infections. This is because these prevention strategies do not address the risk of ascending infection, which can potentially occur anytime during pregnancy, leading to preterm birth or stillbirth.

Overall, the prevention of GBS infection in pregnancy is still a complex question, with risk likely associated to several factors, including the pathogenicity of the GBS strain, host factors, influence of the vaginal/rectal microbiome, false-negative screening results, and/or changes in GBS antibiotic resistance [6,46]. Currently, strategies are mainly focused on the prevention of GBS transmission during labor and delivery through the use of antibiotics. This strategy does not fully capture the biology of the GBS infection, nor does it completely address the full burden of the GBS disease. Moreover, antibiotic resistance is increasing and the use of antibiotics during pregnancy has consequential effects for neonatal health that are only now being appreciated [47]. To successfully eradicate the burden of disease, interventions need to be specifically targeted while having minimal detrimental effects on the microbiome. Therefore, there is a need for alternatives that are respectful with the neonatal and infant microbiota, and that do not compromise the health of future generations. In this context, the final objective of this work was the selection of safe probiotic strains with the in vitro and in vivo ability to eradicate GBS from the intestinal and genitourinary tracts of pregnant women and/or their infants.

The genus *Lactobacillus* constitutes the dominant bacterial group of the vaginal tract in most healthy women, playing a key role in the genitourinary homeostasis [48–52]. In this study, all the vaginal isolates (from either pregnant or non-pregnant healthy women) that fulfilled the initial selection criteria belonged to the species *L. salivarius.* This species is part of the indigenous microbiota of the human gastrointestinal tract, oral cavity, genitourinary tract and milk, and some strains have been studied as probiotics because of their in vitro and in vivo antimicrobial, anti-inflammatory and immunomodulatory properties [53–64]. Previous studies have shown the ability of *L. salivarius* strains to inhibit the growth of vaginal pathogens, including *Gardnerella vaginalis* and *Candida albicans* and, therefore, we have suggested their potential to be used as probiotics for the treatment or prevention of vaginal infections [65,66].

Administration of probiotic bacteria benefits the host through a wide array of mechanisms that are increasingly recognized as being either species- and/or strain-specific [67]. A comparative genomics study that included 33 *L. salivarius* strains isolated from humans, animals or food revealed that this species displays a high level of genomic diversity [68]. Therefore, the selection of *L. salivarius* strains for probiotic use requires the experimental validation of target-tailored phenotypic traits. Some *L. salivarius* strains have shown to be efficient in preventing infectious diseases such as mastitis caused by staphylococci and streptococci, when administered during late pregnancy [69]. Moreover, the oral administration of *L. salivarius* strains is also a valid strategy for the treatment of such a condition during lactation and, in fact, one of the strains was more efficient than antibiotics for this target [70]. In this work, the target was the antagonism towards GBS and, as a consequence, properties such as antimicrobial activity against *S. agalactiae* strains or coaggregation with this species were considered particularly relevant.

The production of antagonistic substances such as bacteriocins, hydrogen peroxide or organic acids represents an important contribution to the defense mechanisms exerted by intestinal and vaginal lactobacilli [59,71]. Some *L. salivarius* strains produce bacteriocins or display bacteriocin-like activity against a variable spectrum of Gram-positive bacteria, including *S. agalactiae* strains [72]. However, none of the *L. salivarius* strains selected in our study displayed bacteriocin-like activity against *S. agalactiae* strains. Therefore, the antimicrobial activity that the selected *L. salivarius* strains exhibited against *S. agalactiae* must be related to the production of other antimicrobial compounds, such as organic acids. The ability of lactobacilli to acidify the vaginal milieu contributes to the displacement and inhibition of pathogens proliferation [73] and, more specifically, the acid production by lactobacilli has been directly correlated with the inhibition of GBS growth [74]. Another antimicrobial defense mechanism attributed to some intestinal or vaginal lactobacilli is the production of peroxide hydrogen, a compound that is toxic for catalase-negative bacteria, such as streptococci [75]. The production of this compound by *L. salivarius* has already been reported [59,76,77]. In our study, *L. salivarius* CECT 9145 (the strain that showed the highest anti-GBS activity) produced high amounts of lactic acid and, in addition, was able to produce peroxide hydrogen.

The ability to adhere to intestinal or vaginal epithelial cells or to mucin, and to co-aggregate with potential pathogens constitutes one of the main mechanisms for preventing their adhesion and colonization of mucosal surfaces. Therefore, it is not strange that such properties are considered relevant to the selection of probiotic strains [28]. The high adherence of *L. salivarius* strains to Caco-2 and HT-29 cells or to mucin has been previously observed [53,59,78]. Globally, *L. salivarius* CECT 9145 showed the best combination of adherence to epithelial cells, co-aggregation with *S. agalactiae* and the inhibition of *S. agalactiae* strains in broth co-cultures. This strain showed a high survival rate during transit through an in vitro gastrointestinal model and survival of lactobacilli when exposed to conditions found in the gastrointestinal tract seems to be a critical pre-requisite for a probiotic strain when its use as a food supplement is pursued, as it was the case.

Some vaginal strains of *L. gasseri* and *L reuteri* have also been reported to co-aggregate with GBS [78]. In contrast, no co-aggregation activity between *S. agalactiae* and other vaginal lactobacilli belonging to the species *L. acidophilus*, *L. gasseri* and *L. jensenii* was observed in another study [32], a fact suggesting that such property is a highly strain-specific trait. In relation to broth co-cultures, the capacity to antagonize the growth of *S. agalactiae* by lactobacilli strains belonging to different species, including *L. salivarius*, has been previously reported [79,80]. Similar to our results, this activity was strain-dependent [79].

One of the most important criteria for the selection of probiotic strains is the assessment of their safety, particularly to the target population. In this work, no adverse effect was reported by any of the women who participated in the clinical trial and ascribed to the probiotic group [thus, receiving *L. salivarius* CECT 9145 at 9 log10 cfu daily for several weeks). Previously, other *L. salivarius* strains have been shown to be well-tolerated and safe in animal models [40] and in human clinical assays [70,81–83], including one involving pregnant women [69].

The *L. salivarius* strains included in this study were very susceptible to most of the antimicrobials tested. In fact, their MICs were lower than the cut-offs established for lactobacilli to seven out of the eight antibiotics required for this species by the European Food Safety Authority [36]. The only exception was kanamycin. The intrinsic resistance of lactobacilli to kanamycin and other aminoglycosides (such as neomycin or streptomycin) has been repeatedly reported [84,85], and this is thought to be an *L. salivarius* species-specific trait due to the lack of cytochrome-mediated transport of this class of antibiotics [86]. The *L. salivarius* strains were also resistant to vancomycin but the assessment of vancomycin sensitivity is not required by EFSA in the case of homofermentative lactobacilli (including *L. salivarius*) since they are intrinsically resistant to this antibiotic probably due to the presence of D-Ala-D-lactate in their peptidoglycan structure [87]. Therefore, *L. salivarius* CECT 9145 and the other strains evaluated in this study can be considered as safe from this point of view.

Lactobacilli are among the Gram-positive bacteria with the potential to produce biogenic amines and these substances can cause several toxicological problems and/or may act as potential precursors of carcinogenic nitrosamines [37]. The screened *L. salivarius* strains neither produced histamine, tyramine, putrescine or cadaverine nor harbored the gene determinants required for their biosynthesis. Additionally, they were unable to degrade gastric mucin in vitro.

Some studies have been focused on the potential of different lactic acid bacteria strains or their metabolites to inhibit the growth of *S. agalactiae* in vitro or in murine models [74,80,88–95]. However, few studies have evaluated the efficacy of probiotic strains for the rectal and vaginal eradication of GBS in pregnant women. Ho et al. [96] examined the effect of *Lactobacillus rhamnosus* GR-1 and *Lactobacillus reuteri* RC-14 taken orally on GBS-positive pregnant women at 35–37 weeks of gestation, and found that GBS colonization changed from positive to negative in 42.9% of the women in the probiotic group. The rate of women that became GBS-negative was lower than in our study and this might be due to the fact that, in the cited study, the recruited women started the probiotic intake many weeks later. A second study using the same two strains (*L. rhamnosus* GR-1 and *L. reuteri* RC-14) provided non-conclusive results due to the low adherence to the probiotic treatment since only seven of 21 women in the intervention group completed the entire 21 days of probiotics [97].

It is important to highlight that nutrition may also play a key role in creating mucosal conditions favoring the action of bacterial strains that are able to improve the rectal and vaginal environments, as it is the case of *L. salivarius* CECT 9145. Such conditions may include the selective fermentation of dietary fiber, the production of relevant bioactive compounds, such as short-chain fatty acids [98], or the use of hyaluronic acid, which has been shown to be useful in the treatment of female recurrent genitourinary infections [99]. The impact of diet on the outcomes of clinical assays involving probiotic-interventions is often underrated and should be taken into account in future studies.

Our study includes the whole process from strain isolation to a pilot clinical study specifically targeting GBS eradication in pregnant women. The criteria followed for the selection of the best candidate for such a target (*L. salivarius* CECT 9145) allowed a notable reduction in the rate of GBS-colonized women and led to a reduction in the use of antibiotics during the peripartum period. As a conclusion, the administration of *L. salivarius* CECT 9145 to GBS-positive pregnant women is a safe and successful strategy to significantly decrease the rates of GBS colonization during pregnancy and, therefore, to reduce the exposure of pregnant women and their infants to intrapartum prophylaxis. Work is in progress to study the mechanisms involving GBS antagonism, including the study of the strain genome and to initiate a multicenter well-designed clinical trial involving a higher number of women.

**Author Contributions:** J.M.R., L.F., D.B. and C.B. designed the study; N.C., V.M., M.M. and R.A. characterized the strain and performed the microbial analyses. S.O. collected the biological samples and coordinated the pilot clinical study. L.F. did the statistical analysis. J.M.R. wrote, edited and revised the manuscript. All the authors approved the final version of the manuscript.

**Funding:** This work was supported by Laboratorios Casen Recordati SL. The funding agency had no role in study design, data collection and analysis.

**Acknowledgments:** We acknowledge all the women, gynaecologists and midwives that participated in this study.

**Conflicts of Interest:** NC is an employee of Probisearch SLU. CB is employee of Casen Recordati SL. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

#### *Article*
