Since plant polysaccharides are considered a structural constituent of the cell wall, the appropriate extraction methods for botanical polysaccharides require efficiently breaking the cell wall and accelerating the release of intracellular polysaccharides (Lin et al., 2017). Classical hot water extraction is the convenient approach for the extraction of polysaccharides (Zhang et al., 2010a; Xu et al., 2012), but in addition to this traditional method, other methods for the extraction of SPS have also been reported in recent years (Yang et al., 2014; Yuan et al., 2019). Generally, the dried flowers of safflower are ground into a powder before extraction with distilled water, either as a decoction or ultrasonically. After that, the extraction solutions are centrifuged, concentrated, and precipitated with ethanol to obtain crude SPS. Subsequently, the crude SPS is deproteinized with Sevag reagent, followed by successive washes with ethanol, acetone and ether to remove the fat-soluble ingredients. The final refined SPS can then be obtained by drying the separated precipitate (Xu et al., 2012). The extraction process of SPS is schematically shown in Figure 2. A review of the literature indicates that the yield of crude SPS varies widely under different extraction conditions, ranging from 2.75 to 20.35% (Zhang et al., 2010a; Wang and Wang, 2010; Zou, 2011; Xu et al., 2012; Yang et al., 2012; Deng et al., 2013; Yang et al., 2014; Yuan et al., 2019; Hu, 2020; Wang R. et al., 2021). The detailed information of various methods for the extraction of SPS is listed in Table 1.

Extraction conditions including particle size of the raw material powder, extraction method, ratio of raw material to solvent, extraction temperature and time all have great influence on the extraction rate of SPS. Yang et al. (2014) compared the extraction efficiency of safflower powder with different particle sizes, and found that the SPS yield increased 25% when using superfine powder with a particle size of 5–10 μm. The yield of SPS can be improved by different methods. Wang et al. investigated the effects of the solid-liquid ratio, temperature, and time in single factor and orthogonal experiments, which finally resulted in the highest yield of SPS based on hot water decoction reported to date. Under the optimized conditions, including an extraction time of 1.5 h, extraction temperature of 80°C, water: material ratio of 20:1, and three rounds of extraction, the yield of SPS reached up to 10.19% (Wang and Wang, 2010). In a different approach, it was found that ultrasonic-assisted extraction, which has the advantages of convenient operation, short extraction time, and low cost, can break the cell walls more effectively and increase the yield of SPS compared to other methods (Yang et al., 2012; Yuan et al., 2019). The optimal conditions for ultrasonic-assisted extraction were found to encompass a water/raw material ratio of 25:1 (V:m), temperature of 65°C, and ultrasonication time of 55 min, which resulted in a yield of 20.35% (Yang et al., 2012). These various extraction methods are all dedicated to improving the extraction efficiency of SPS, and aim to establish an economical and environmental strategy.

Natural polysaccharides often contain a lot of pigments and impurities, which not only reduces the purity of polysaccharide, but also affects the qualitative and quantitative analysis (Zheng et al., 2017). Therefore, efficient decolorization and purification of SPS is essential. After extraction, the crude SPS can be further purified by a combination of techniques, including deproteinization, decolorization, ion-exchange chromatography, and gel filtration chromatography, as shown in Figure 2 (Zou, 2011; Ren et al., 2013; Hu, 2020). Zou et al. investigated various approaches of the deproteinization and decolorization of SPS, and the optimized method based on a combined enzyme-Sevag method, and HPD-100 macroporous adsorption resin afforded an optimal decolorization rate and retention rate of SPS (Zou et al., 2011). The refined SPS was further separated and purified mainly via ion-exchange chromatography (DEAE-Sepharose, DEAE-Cellulose, DEAE-Sephadex) and gel filtration chromatography (Sephadex G, Sephacryl S, Sepharose CL) (Zhu et al., 2018). Five polysaccharide fractions (SPS1, SPS2, SPS3, CTLP1, CTLP2) were isolated and purified using a DEAE-52 cellulose column and Sephadex G-100 column (Zou, 2011; Ren et al., 2013; Hu, 2020). After purification on a DEAE-Sephadex A-25 column and size-exclusion column (Sephadex G-50), a single polysaccharide fraction named CTP was obtained (Huo et al., 2005a). The purification of safflower polysaccharides (SF1 and SF2) was directly conducted using a Sepharose CL-2B column (Wakabayashi et al., 1997). DEAE cellulose column chromatography and a Sephacryl S-200HR column were used to obtain the polysaccharide fraction HH1-1 (Yao et al., 2018). SPAW and SPSa were isolated and purified using a DEAE-cellulose column and a Sepharose CL-6B column (Cui et al., 2019; Cui et al., 2020). Appropriate separation and purification methods should be chosen according to the characteristics of the polysaccharides, while optimizing the convenience and efficiency. Up to now, there is no report on the industrialized production of safflower polysaccharides.

There are three main approaches for the extraction of PBPC, including water extraction, dilute alkali extraction and microwave-assisted methods (Zuo and Qian, 2012; Zuo and Qian, 2013; Li et al., 2017c; Shi et al., 2020). Interestingly, traditional hot water extraction remains the method with the best extraction efficiency to date. A schematic representation of the extraction methods of PBPC is shown in Figure 2. In these approaches, the walls of safflower bee pollen are broken and the material is refluxed with petroleum ether (2 h) and 95% ethanol (two times) successively. After filtration, the residues are dried and extracted at 80°C over 12 h with hot water (1:20, w/v) for the first time, and then at the same temperature over 8 h with hot water (1:10, w/v) for the second time. The hot water extraction solutions are then centrifuged, concentrated, and precipitated with ethanol. The resulting precipitate is further processed through a series of treatments including deproteinization with Sevag reagent, followed by washing with ethanol, acetone and ether (three times) to finally obtain the crude PBPC (Zuo and Qian, 2012; Zuo and Qian, 2013). Li et al. (2017b) investigated various approaches for the deproteinization of PBPC using response surface methodolog, and the optimized method was enzyme-Sevag treatment. A DEAE-52 cellulose column and Sephadex G-75 column were used for the separation and purification of polysaccharides, which yielded six polysaccharide fractions designated as PBPC-Ia, PBPC-Ib, PBPC-II, PBPC-III, PBPC-IV, and PBPC-V (Li et al., 2020).

Unlike water extraction, dilute alkali extraction is better for isolating the cell wall bound or intracellular polysaccharides, which may have better antioxidant and free radical scavenging activities (Jiao et al., 2015). Compared with the preparation of water-soluble polysaccharides, alkali-soluble polysaccharide preparation includes an additional extraction with 0.05 M NaOH after water extraction. The alkali-soluble polysaccharides from bee pollen of safflower was then purified using a Sephadex G-75 column and named APBPC, which was further separated on a DEAE-52 cellulose column into the five fractions APBPC-1, APBPC-2, APBPC-3, APBPC-4, and APBPC-5 (Shi, 2021). Compared with these traditional approaches, there was no obvious advantage of microwave-assisted methods (Li et al., 2017c).

The average Mw of polysaccharides is mainly determined by high performance liquid chromatography (HPLC), gel chromatography, high performance size-exclusion chromatography or high performance gel permeation chromatography (HPGPC) (Huo, 2005; Zou, 2011; Hu, 2020; Shi, 2021). The Mw of isolated SPS ranges from 4 × 10³ Da (CTP) to >100 × 10³ Da (SF1 and SF2). The Mws of SPS derived from the same crude polysaccharide mixtures were in the same order of magnitude, but an analysis of PBPC from the same source on a Sephadex G-150 gel column, revealed a large variation in Mw, ranging between 3.77 × 10³ Da (PBPC-II) and 2.575 × 10⁵ Da (PBPC-Ib) (Li et al., 2017a).

The composition of monosaccharides was mostly determined using the 1-phenyl-3-methyl5-pyrazolonde (PMP)-derivatization method or the derivatization with saccharin acetyl (Guo et al., 2021). Both methods begin with acid hydrolysis, followed by derivatization, and finally analysis of the derivatized product. The products of the PMP derivatization method are detected by HPLC, while the saccharin acetyl method yields adducts that are detected by GC (Fan et al., 2013). Because each polysaccharide has a different monosaccharide composition or molar ratio, the structures of polysaccharides are diverse. The heteropolysaccharides, isolated the flowers of safflower, mainly consist of rhamnose, arabinose, mannose, xylose, glucose, and galactose with different molar ratios. Among the SPS fractions, there were eight heteropolysaccharides with different monosaccharide compositions and molar ratios, as well as two homopolysaccharides containing glucose. Notably, the SPS2, SPS3, PBPC-II and APBPC-2 fractions, which were isolated from safflower and bee pollen of safflower contain uronic acid. The contents of uronic acid were in the order of 7.06, 6.84 and 11.78%, while the content of uronic acid in PBPC-II was not mentioned ( Zou, 2011; Wang et al., 2019b; Shi, 2021). The differences of monosaccharide composition among these polysaccharides may be caused by different sources of raw materials or different separation and purification methods.

Numerous technical methods have been used to demonstrate the structural features of polysaccharides. However, studies on the structural identification of SPS and PBPC are still rare in the literature, and only nine obtained polysaccharides were structurally characterized in detail. The structural characteristics of these nine polysaccharides are listed in details below. The main chain of CTLP-1 and CTLP-2 could be determined by periodate oxidation, Fourier transform infrared (FT-IR) spectroscopy, and nuclear magnetic resonance (NMR) spectroscopy (Hu, 2020). The main chain of CTLP-1 was composed of →1)-GalAp-(1→4)-Arap-(1→2)-Glcp-(4→, and the main chain of CTLP-2 was composed of →1)-Galp-(2,6→1)-Arap-(4,6 →1)-Glup-(3→ repeats (Hu, 2020). According to the results of FT-IR spectroscopy, CTLP-1, CTLP-2 and SPS3 were all composed of a-D-pyranose units, while SPS2 was a β-D-(1→3)-glucan (Ren et al., 2013; Hu, 2020). In addition, Congo red staining showed that both CTLP-2 and SPS2 had a multi-stranded helical structure, which may have a great influence on their biological activity (Ren et al., 2013; Hu, 2020). Based on the results of periodate oxidation, Smith degradation and methylation, the main chain of CTP was deduced to be composed of (1→4)-linked glucose and five side chains (Huo et al., 2005b). The backbone structure of HH1-1, a neutral arabinogalactan, was found to consist of 1,6-linked galactopyranosyl residues with branched chains at C-3 according to NMR and methylation analysis (Yao et al., 2018). SPSa and SPAW are both homogeneous polysaccharides composed of β-Glc. However, the glycosidic bonds of their main chains are different, in the order of 1,4,6-β-Glcp and (1→3)-linked β-D-Glcp (Cui et al., 2019; 2020). According to the FT-IR spectroscopy, β-elimination and iodine reaction, it is presumed that PBPC-Ⅱ is a β-pyranosyl polymer with uronic acid units and O-glycosidic bonds, and its main chain may be composed of galactose with long side chains and branches (Wang et al., 2019b).

Immunomodulation is one of the most important biological activities of SPS and PBPC. Moreover, SPS can also act as a biological response modifier by enhancing antitumor effects through immunomodulation. It is known that the immune system is composed of immune organs, immune cells, and immune active substances, which all play important roles in immune surveillance, defense, and regulation (Wang and Hu, 2008). Thus, the immunomodulatory activity of SPS and PBPC is illustrated from the three aspects of organs, cells and cytokines.

The immune organ index is a preliminary tool for the assessment of immune capability (Wang et al., 2010). SPS have been proved to improve immune function by increasing the immune organ index. Compared with the control, SPS were found to significantly increase the indexes of the thymus and spleen in H22 tumor-bearing mice and thus enhance the immune function of the organism (He et al., 2009). Huang et al. (1984) found that SPS could effectively counteract the immunosuppressive effect of prednisolone injection, resulting in a significant increase in the weight and cell number of the shrunken spleen, indicating a positive immunomodulatory effect. Moreover, PBPC were found to increase the thymus and spleen index of immunosuppressed mice, antagonize the atrophy of the thymus and spleen caused by Cy, and improve non-specific immunity in mice (Wang L. F. et al., 2020).

Bioactive polysaccharides can directly or indirectly interact with the immune system, triggering several cellular/molecular events, that lead to immune system activation (Leung et al., 2006). Natural killer (NK) cells, T-cells, B-cells and macrophages are the main targets that were found to respond to these molecules (Ferreira et al., 2015). The modulatory effects of SPS are mainly reflected in the activation of immune cells, promotion of cell proliferation, and enhancement of cell killing activity. Both SF1 and SF2 can induce the proliferation of B-cells and IgM production, and also stimulate NO production by macrophages in different ways (Wakabayashi et al., 1997). By co-culturing human peripheral mononuclear cells (PBMC) with SPS in vitro, it was found that SPS could promote the proliferation of PBMC and CD8⁺ T-cells (Tao et al., 2011), as well as increase the killing activity of NK and LAK cells (Zhou et al., 2010). As a biological response modifier, SPS can partially reverse the inhibitory state of NK cells in a T739 mouse model of lung cancer, and significantly enhance the cytotoxicity of splenic CTL cells and NK cells in tumor-bearing mice (Shi et al., 2010b).

Similar to most immunostimulatory polysaccharides, SPS can also enhance the secretion of pro-inflammatory cytokines according to the in vitro and in vivo experiments. SPS could induce PBMCs to secrete interferon (IFN)-γ and interleukin (IL)-2 (Shi et al., 2010a). In S180 sarcoma mice, Ma et al. found that it could also up-regulate the cytokines IL-12 and tumor necrosis factor (TNF)-α, while down-regulating the cytokine of IL-10 (Ma et al., 2013b). The purified polysaccharides (HH1-1, SF1, SF2) obtained from safflower exert their immunomodulatory effects by activating NF-κB signaling. Among them, HH1-1 increased the expression of the cytokines TNF-α, IL-1β and iNOS in lymphocytes and macrophages (Yao et al., 2018). Similarly, SF1 and SF2 induced the production of IL-12 and IFN-γ by peritoneal macrophages in a dose-dependent manner (Ando et al., 2002). It was also reported that PBPC can promote humoral and cellular immunity in mice. Additionally, PBPC were found to promote the ConA-induced transformed proliferation of T lymphocytes in mice (Zuo et al., 2017). A schematic diagram illustrating the immunomodulatory mechanism of SPS and PBPC is shown in Figure 4.

It is generally believed that polysaccharides exert their antioxidant activity mainly by targeting the NF-κB and Nrf2/Keap1 ARE signaling pathways (Zhang et al., 2019). The intake of polysaccharides can activate pathways and up-regulate the expression of defensive genes, thereby increasing the levels of antioxidant enzymes and molecules such as superoxide dismutase (SOD), glutathione peroxidase (GSH-Px), catalase (CAT), glutathione (GSH), and glutathione reductase (GR), thereby protecting the cells from oxidative stress (Prestera et al., 1995). Reactive oxygen species (ROS) is a general term for oxygen-containing chemicals with strong oxidative capacity, which play an important role in cell signaling and homeostasis in vivo. When oxidative stress occurs in the body, the level of ROS will rise sharply, and the oxidative stress product malondialdehyde (MDA) will be produced, which can reflect the degree of oxidative damage (Wang L. et al., 2021; Guo et al., 2021).

The antioxidant activities of SPS and PBPC have been demonstrated in some studies. Wan investigated the antioxidant activity of SPS in H22 tumor-bearing mice, and the results showed that the activities of antioxidant enzymes (GR, GSH-PX, CAT and SOD) were increased, while the ROS and MDA content was decreased in the high-dose group (Wan, 2016). It is suggested that SPS can inhibit ROS production and slow down SOD depletion in the body to a certain extent, thus exerting a protective effect against lipid peroxidation. It was also found that APBPC could increase the activity of SOD and GSH-Px in the serum, liver and brain tissues of a D-galactose-induced mouse model of aging, thereby reducing the levels of the oxidative stress product MDA. This indicates that APBPC can enhance the function of free-radical scavenging, reduce the production of free radicals and exert anti-aging effects (Shi et al., 2020). The purified polysaccharides (SPS2, SPS3, CTLP-1, CTLP-2) have also shown antioxidant activity according to the vitro assays (Zou, 2011; Hu, 2020).

Malignant tumors remain one of the leading causes of death worldwide (Dong, 2011). Modern pharmacological research has confirmed that the traditional Chinese medicine and its bioactive compounds can exert antitumor effects through multiple pathways and targets (Sun et al., 2021), which has become a new research hotspot in modern oncology. Due to the antitumor effect of safflower and other natural polysaccharides (Zhang et al., 2013; Niu et al., 2017), researchers have conducted a series of in-depth studies on the molecular mechanisms under the anticancer effect of SPS and PBPC. Both classes of polysaccharides have broad-spectrum antitumor activities, and exert their antitumor effects by improving the immune response, inducing the apoptosis of tumor cells and preventing the spread or migration of tumor cells. The detailed antitumor mechanisms of SPS and PBPC are summarized in below, as shown in Figure 5.

Cerebral ischemia-reperfusion injury (CIRI) is a complex temporal and spatial cascade of reactive pathophysiological processes that occur during the recovery of brain tissue from ischemia to perfusion (Ren and Tian, 2021), leading to a more severe pathological response than before perfusion. This disease is often associated with the increase of intracellular free radicals, overload of intracellular Ca2⁺, neurotoxicity of excitatory amino acids, overactive inflammatory response and excessive apoptosis (Yuan et al., 2021). SPS were able to significantly improve the neurological deficits as well as reduce the volume ratio and water content of the infarcted area in rats with cerebral ischemia-reperfusion injury. At the same time, the amounts of TNF-α, IL-1β and IL-10 in the brain tissue of modeled rats were also influenced by SPS. It can be concluded that the mechanism of the protective effects of SPS might be related to the inhibition of the synthesis of inflammatory factors (Ren et al., 2016). Apoptosis of neuronal cells in the ischemic area is one of the most important causes of CIRI. SPS can inhibit the expression of caspase-3 and RAPP in a dose-dependently manner, thus exerting anti-apoptotic effects on neuronal cells for the prevention of CIRI (Wang, 2011).

Steroid-induced avascular necrosis of the femoral head (SANFH) is a class of non-traumatic femoral head injury that can trigger osteoclast and bone marrow necrosis due to local blood circulation disorders (Wu et al., 2019). The precise mechanism of SANFH remains unclear and may be associated with various factors such as apoptosis and inflammation (Ye et al., 2019). In vitro experiment has shown that SPS can inhibit caspase-3 activation in the osteoblasts of rats and promote bone formation (Cui et al., 2018). Cui et al. also investigated the effects of two purified SPS (SPAW and SPSa) on femoral head necrosis induced by dexamethasone using in vivo assays. It was found that the abnormal histopathological changes in the treated group were significantly improved, and the HOM/HOP ratio, which reflects the metabolism of femoral head growth, was increased (Cui et al., 2019; Cui et al., 2020). SPAW was also able to significantly down-regulate the expression of Bax and caspase-3 proteins, while upregulating the expression of Bcl-2 protein at the same time. This implies that SPAW exerts an antiapoptotic effect in osteoclasts following SANFH treatment (Cui et al., 2020).

In addition to the above-mentioned activities, there are other biological activities observed in SPS and PBPC, such as anticoagulant and antibacterial effects. SPS was able to inhibit platelet aggregation induced by 4,5′-adenosine diphosphate disodium salt (ADP) in a concentration-dependent manner. The aggregation inhibition rate was higher than that of the aqueous and alcoholic extracts of safflower (Zhou et al., 2008). The bee pollen polysaccharide fraction PBPC-II exerted significant anticoagulant effects in vitro by mediating exogenous and endogenous coagulation pathways. It prolonged the activated partial thromboplastin time (APTT) and prothrombin time (PT) in human plasma as well. Compared with the anticoagulant heparin sodium, PBPC-II was more effective in prolonging PT (Li et al., 2017a). Moreover, PBPC-II was reported to have certain antibacterial effects on Escherichia coli and Staphylococcus aureus (Li et al., 2017a).