Antiprotozoal and Antitumor Activity of Natural Polycyclic Endoperoxides: Origin, Structures and Biological Activity

Polycyclic endoperoxides are rare natural metabolites found and isolated in plants, fungi, and marine invertebrates. The purpose of this review is a comparative analysis of the pharmacological potential of these natural products. According to PASS (Prediction of Activity Spectra for Substances) estimates, they are more likely to exhibit antiprotozoal and antitumor properties. Some of them are now widely used in clinical medicine. All polycyclic endoperoxides presented in this article demonstrate antiprotozoal activity and can be divided into three groups. The third group includes endoperoxides, which show weak antiprotozoal activity with a reliability of up to 70%, and this group includes only 1.1% of metabolites. The second group includes the largest number of endoperoxides, which are 65% and show average antiprotozoal activity with a confidence level of 70 to 90%. Lastly, the third group includes endoperoxides, which are 33.9% and show strong antiprotozoal activity with a confidence level of 90 to 99.6%. Interestingly, artemisinin and its analogs show strong antiprotozoal activity with 79 to 99.6% confidence against obligate intracellular parasites which belong to the genera Plasmodium, Toxoplasma, Leishmania, and Coccidia. In addition to antiprotozoal activities, polycyclic endoperoxides show antitumor activity in the proportion: 4.6% show weak activity with a reliability of up to 70%, 65.6% show an average activity with a reliability of 70 to 90%, and 29.8% show strong activity with a reliability of 90 to 98.3%. It should also be noted that some polycyclic endoperoxides, in addition to antiprotozoal and antitumor properties, show other strong activities with a confidence level of 90 to 97%. These include antifungal activity against the genera Aspergillus, Candida, and Cryptococcus, as well as anti-inflammatory activity. This review provides insights on further utilization of polycyclic endoperoxides by medicinal chemists, pharmacologists, and the pharmaceutical industry.

In fungi, both cultivated and wild, polycyclic endoperoxides are found in small quantities, but ergosterol peroxide (33) is the most abundant [5,6]. Below, we present data on the distribution of this steroid and other polycyclic endoperoxides in fungi, fungal endophytes and lichens.
Endoperoxide (42), bearing a keto group at the 12 position, has been isolated from the fungus Fusarium monilforme [129]. Endoperoxy glycoside (43) was detected in ethanol extract of the fungus Lactarius volemus, which demonstrated anticancer activity [130,131]. Ergosterol peroxide (33) and unusual steroid called asperversin A (44) have been isolated from endophytic fungus of Aspergillus versicolor that was isolated from the seaweed Sargassum thunbergii. Both steroid antibiotics showed antibacterial activity against Escherichia coli and Staphylococcus aureus [132], and another steroid named fuscoporianol D (45) was found in a MeOH extract of in field-grown mycelia of Inonotus obliquus [133].
It is known that natural hypocrellin is a dark red dye with photodynamic activity against several microorganisms was isolated from the fungus Hypocrella bambusae, and its photooxidation product called peroxyhypocrellin (68) has an anthracene endoperoxide arrangement within the perylene quinone structure [149]. Structures  can be seen in Figures 3 and 4, and their biological activity is presented in Tables 3 and 4. s 2021, 26, x FOR PEER REVIEW 11 of 39 arrangement within the perylene quinone structure [149]. Structures  can be seen in Figures 3 and 4, and their biological activity is presented in Tables 3 and 4.
A peroxide-sesquetepene, called nardosaldehyde (69) was isolated from the roots of Nardostachys chinensis, and biological activity was not determined [155]. Structures  can be seen in Figure 5, and their biological activity is presented in Tables 5 and 6. Peroxygibberol (16) is marine peroxide (5.9%) was also found in Agarwood oil obtained from highly infected Aquilaria malaccensis wood [156].

Comparison of Biological Activities of Natural Polycyclic Endoperoxides
It is currently accepted that the biological activity of both natural and synthetic compounds depends on their structure [33,251,252]. Despite the activity cliffs observed for some drug-like compounds [253], which can be considered as a violation of this rule, structure-activity relationships (SAR) are widely used in medicinal chemistry for finding and optimization new pharmacological agents [254].
PASS is the first software for in silico estimation of biological activity profiles [33,255], of which the development has been started more than 30 years ago [256]. Its current implementation predicts about 8000 pharmacological effects, molecular mechanisms of action, pharmacological effects, toxicity, side effects, anti-targets, transporters-related interactions, gene expression regulation, and metabolic terms [31]. Due to the utilization of chemical descriptors that reflect the essential features of ligand-target interactions and a robust mathematical approach for analysis of structure-activity relationships, the average accuracy of PASS predictions was 96% [31,252,257,258]. Based on the PASS predictions provided by the appropriate web-service [259], over 29,000 researchers from 104 countries selected the most promising virtually designed molecules for synthesis and determined the optimal directions for testing their biological activity [260][261][262][263][264].
In this study, PASS predictions were used to estimate the general pharmacological potential for the analyzed natural polycyclic endoperoxides. For about 8000 pharmacological effects and molecular mechanisms of action, probabilities of belonging to the class of "actives" Pa, varied from zero to one, were estimated. The higher the Pa value is, the higher the probability of confirming the predicted activity in the experiment. On the other hand, estimated Pa values might be relatively small for some activities if the analyzed molecule is not like the active compounds from the PASS training set. Thus, PASS prediction interpretation requires considering two contradictory issues high probability of activity vs. high structural novelty. The researcher decides which issue is more critical, depending on the task or the project [18,31,35,257,258].
Analyzing the data obtained with PASS of natural polycyclic endoperoxides and artemisinin and its analogs currently used in medicine, it can be stated that for all polycyclic endoperoxides, antiprotozoal activity is estimated with a Pa from 70 to 99.6%. For some compounds, antiparasitic activity is also estimated, with a Pa from 50 to 88.3%. The antiprotozoal and antiparasitic activities predicted using the PASS are shown in Tables 1-11, and the chemical structures are shown in Figures 1-11. A 3D graph of the predicted pharmacological activities of artemisinin (86) and its analogs is shown in Figure 12.
Artemisinin and its analogs (both natural and synthetic) are widely used in medical practice and are essential antimalarial treatment components. Figure 12 shows the predicted pharmacological activities of artemisinin and its analogs using PASS, and Figure 13 demonstrates the predicted pharmacological activities of artemisinin.
More than one million articles and reviews have been devoted to various antitumor and related activities of both natural and synthetic compounds. In an earlier section, we presented and discussed the antitumor activity of polycyclic endoperoxides isolated from various terrestrial and aquatic organisms computed using PASS.
According to the PASS estimates presented in Tables 1-11, many endoperoxides demonstrate antitumor and related activities to varying degrees. However, we are interested in compounds for which such activity is estimated with more than 95% probability. Figure 14 demonstrates natural compounds and their predicted antitumor activity with Pa > 95%. Figure 12. The 3D graph shows the predicted and calculated pharmacological activities of artemisinin (86) and its analogs, such as 12α-OH-artemisinin (87), 12β-OH-artemisinin (88), artemether (89), arteether (90), artelinate (91), and artesunic acid (92). According to the PASS data, artemisinin and its analogs (86)(87)(88)(89)(90)(91)(92) show selective activity against obligate intracellular protozoan parasites belonging to the genera Plasmodium, Toxoplasma, Leishmania, and Coccidia, which is the main pharmacological activity with a confidence level of more than 90%. In addition, all these endoperoxides show antifungal activity against the opportunistic pathogenic yeasts Candida and Cryptococcus, as well as anticancer activity for some compounds; the confidence level exceeds 90 percent.
Artemisinin and its analogs (both natural and synthetic) are widely used in medical practice and are essential antimalarial treatment components. Figure 12 shows the predicted pharmacological activities of artemisinin and its analogs using PASS, and Figure  13 demonstrates the predicted pharmacological activities of artemisinin. Figure 13. The 3D graph shows the predicted and calculated pharmacological activities of artemisinin (86), which was found in 1979 in the extract of the Chinese herb Qinghaosu (Artemisia annua). According to PASS data, this endoperoxide demonstrated 16 different activities, with 5 activities having a found confidence of more than 90 percent. Antiprotozoal selective activity of artemisinin against obligate intracellular protozoan parasites belonging to the genera Plasmodium (99.5%), Toxoplasma (93%), Leishmania (92.3%), and Coccidia (78%) is the main pharmacological activity. In addition, artemisinin demonstrated strong anti-schistosomal activity (91.1%) against Schistosoma mansoni, a human blood fluke parasite. Additionally, artemisinin shows antifungal activity against an opportunistic pathogenic yeast Candida (91.5%) and Cryptococcus (85.3%), although anticancer activity is found at 80%.  (89), arteether (90), artelinate (91), and artesunic acid (92). According to the PASS data, artemisinin and its analogs (86)(87)(88)(89)(90)(91)(92) show selective activity against obligate intracellular protozoan parasites belonging to the genera Plasmodium, Toxoplasma, Leishmania, and Coccidia, which is the main pharmacological activity with a confidence level of more than 90%. In addition, all these endoperoxides show antifungal activity against the opportunistic pathogenic yeasts Candida and Cryptococcus, as well as anticancer activity for some compounds; the confidence level exceeds 90 percent. Figure 12. The 3D graph shows the predicted and calculated pharmacological activities of artemisinin (86) and its analogs, such as 12α-OH-artemisinin (87), 12β-OH-artemisinin (88), artemether (89), arteether (90), artelinate (91), and artesunic acid (92). According to the PASS data, artemisinin and its analogs (86)(87)(88)(89)(90)(91)(92) show selective activity against obligate intracellular protozoan parasites belonging to the genera Plasmodium, Toxoplasma, Leishmania, and Coccidia, which is the main pharmacological activity with a confidence level of more than 90%. In addition, all these endoperoxides show antifungal activity against the opportunistic pathogenic yeasts Candida and Cryptococcus, as well as anticancer activity for some compounds; the confidence level exceeds 90 percent.
Artemisinin and its analogs (both natural and synthetic) are widely used in medical practice and are essential antimalarial treatment components. Figure 12 shows the predicted pharmacological activities of artemisinin and its analogs using PASS, and Figure  13 demonstrates the predicted pharmacological activities of artemisinin. Figure 13. The 3D graph shows the predicted and calculated pharmacological activities of artemisinin (86), which was found in 1979 in the extract of the Chinese herb Qinghaosu (Artemisia annua). According to PASS data, this endoperoxide demonstrated 16 different activities, with 5 activities having a found confidence of more than 90 percent. Antiprotozoal selective activity of artemisinin against obligate intracellular protozoan parasites belonging to the genera Plasmodium (99.5%), Toxoplasma (93%), Leishmania (92.3%), and Coccidia (78%) is the main pharmacological activity. In addition, artemisinin demonstrated strong anti-schistosomal activity (91.1%) against Schistosoma mansoni, a human blood fluke parasite. Additionally, artemisinin shows antifungal activity against an opportunistic pathogenic yeast Candida (91.5%) and Cryptococcus (85.3%), although anticancer activity is found at 80%. Figure 13. The 3D graph shows the predicted and calculated pharmacological activities of artemisinin (86), which was found in 1979 in the extract of the Chinese herb Qinghaosu (Artemisia annua). According to PASS data, this endoperoxide demonstrated 16 different activities, with 5 activities having a found confidence of more than 90 percent. Antiprotozoal selective activity of artemisinin against obligate intracellular protozoan parasites belonging to the genera Plasmodium (99.5%), Toxoplasma (93%), Leishmania (92.3%), and Coccidia (78%) is the main pharmacological activity. In addition, artemisinin demonstrated strong anti-schistosomal activity (91.1%) against Schistosoma mansoni, a human blood fluke parasite. Additionally, artemisinin shows antifungal activity against an opportunistic pathogenic yeast Candida (91.5%) and Cryptococcus (85.3%), although anticancer activity is found at 80%.
presented and discussed the antitumor activity of polycyclic endoperoxides isolated from various terrestrial and aquatic organisms computed using PASS.
According to the PASS estimates presented in Tables 1-11, many endoperoxides demonstrate antitumor and related activities to varying degrees. However, we are interested in compounds for which such activity is estimated with more than 95% probability. Figure 14 demonstrates natural compounds and their predicted antitumor activity with Pa > 95%. Figure 14. The 3D graph shows the predicted and calculated antitumor activity of selected polycyclic endoperoxides (compound numbers: 11, 17, 30, 33, 142, 143, 164, and 165) showing the highest degree of confidence, more than 95%. These polycyclic endoperoxides can be used in clinical medicine as agents with strong antitumor activity. Some polycyclic endoperoxides, in addition to antiparasitic, antiprotozoal, and antitumor activities, demonstrate other activities with Pa > 90%, which should also be mentioned in this article. This is primarily anti-inflammatory activity. Figure 15 demonstrates such compounds as well as their predicted anti-inflammatory activity. It should also be noted that endoperoxide artemisinin (86) and its analogs, and some other compounds, show antifungal activity. Figure 16 demonstrates predicted antifungal activity with Pa> 90%. Figure 14. The 3D graph shows the predicted and calculated antitumor activity of selected polycyclic endoperoxides (compound numbers: 11, 17, 30, 33, 142, 143, 164, and 165) showing the highest degree of confidence, more than 95%. These polycyclic endoperoxides can be used in clinical medicine as agents with strong antitumor activity. Some polycyclic endoperoxides, in addition to antiparasitic, antiprotozoal, and antitumor activities, demonstrate other activities with Pa > 90%, which should also be mentioned in this article. This is primarily anti-inflammatory activity. Figure 15 demonstrates such compounds as well as their predicted anti-inflammatory activity. It should also be noted that endoperoxide artemisinin (86) and its analogs, and some other compounds, show antifungal activity. Figure 16 demonstrates predicted antifungal activity with Pa> 90%.  : 1, 8, 9, 68, 94, 95, 96, 97, 98, 100, and 113) showing the highest degree of confidence, i.e., more than 95%. These polycyclic endoperoxides can be used as potential agents with strong anti-inflammatory activity.  : 1, 8, 9, 68, 94, 95, 96, 97, 98, 100, and 113) showing the highest degree of confidence, i.e., more than 95%. These polycyclic endoperoxides can be used as potential agents with strong anti-inflammatory activity. Figure 15. The 3D graph shows the predicted and calculated anti-inflammatory activity of selected polycyclic endoperoxides (compound numbers: 1, 8, 9, 68, 94, 95, 96, 97, 98, 100, and 113) showing the highest degree of confidence, i.e., more than 95%. These polycyclic endoperoxides can be used as potential agents with strong anti-inflammatory activity.

Conclusions
In this review, we presented more than 170 polycyclic endoperoxides isolated from various sources and showed that all endoperoxides demonstrate antiprotozoal activity with varying degrees of reliability, and among them, the artemisinin group and some other compounds are significantly distinguished from of all endoperoxides presented and have a strong antiprotozoal activity. Our data only confirm that the artemisinin group has unique properties, which is why it has been used in medical practice for more than 50 years in the fight against malaria parasites. In addition, the artemisinin group has a high

Conclusions
In this review, we presented more than 170 polycyclic endoperoxides isolated from various sources and showed that all endoperoxides demonstrate antiprotozoal activity with varying degrees of reliability, and among them, the artemisinin group and some other compounds are significantly distinguished from of all endoperoxides presented and have a strong antiprotozoal activity. Our data only confirm that the artemisinin group has unique properties, which is why it has been used in medical practice for more than 50 years in the fight against malaria parasites. In addition, the artemisinin group has a high antifungal activity, while some other endoperoxides have a strictly strong anti-inflammatory activity.
Compounds such as (19), (23), and (25) exhibited anti-hypercholesterolemic action, and compounds (166) and (168) have a strong stimulating effect on the respiratory and vasomotor centers of the brain. However, to confirm the conclusions regarding the in silico estimations, more research is required.