Anticancer and Antifungal Compounds from Aspergillus, Penicillium and Other Filamentous Fungi

This review covers important anticancer and antifungal compounds reported from filamentous fungi and in particular from Aspergillus, Penicillium and Talaromyces. The taxonomy of these fungi is not trivial, so a focus of this review has been to report the correct identity of the producing organisms based on substantial previous in-house chemotaxonomic studies.


Introduction
Filamentous fungi such as Aspergillus, Penicillium and Talaromyces are some of the most incredible chemical factories known today. Accordingly, numerous bioactives such as mycotoxins, antifungal and anticancer agents have been reported in the literature within the last more than 100 years [1]. Despite this many new compounds revealing remarkable new bioactivities are still being discovered, including well-known metabolites such as griseofulvin [2][3][4][5]. This, combined with the fact that large combinatorial libraries have not provided the anticipated number of new chemical entities explains why the field of natural products is currently assuming new prominence. Thus, natural products are OPEN ACCESS still used as scaffolds for synthetic organic chemistry, although nowadays primarily for designing smaller and more focused diversity oriented synthesis-derived libraries [6].
It has been estimated that approximately 1.5 million or likely as many as 3 million fungal species exist on Earth, of which only around 100,000 species have been described so far [7]. A multitude of new species are likely to be discovered from diverse habitats, such as tropical forest plants and soils, associated to insects and in the marine environment. In addition to untapped biodiversity recent sequencing of complete fungal genomes has revealed that many gene clusters are silent, suggesting the possibility for many more compounds [8]. Despite several efforts to stimulate such pathways using epigenetic modifiers [9,10], it is evident that we still do not know the full biosynthetic potential even of well studied model organisms such as Aspergillus nidulans, altogether strongly indicating Nature's potential as source of new promising bioactive small molecule scaffolds.
The renewed interest in natural product discovery is further enhanced with the new strategies and methodologies for fast dereplication that have been developed within the last decade. Thus chromatographic and spectroscopic methods are combined with database searching in for example Antibase, which is a comprehensive database for natural products from microorganisms [11]. The performance of mass spectrometers is continuously improving, including easy access to both positive and negative ionization spectra even during fast ultra-high-performance liquid chromatography (UHPLC). Altogether the use of accurate mass measurements for dereplication of unknown compounds reduces the number of predicted elemental compositions, ensuring that database searches are conducted with the fewest possible candidates [12]. The use of UV spectral information is often important in the dereplication process for prioritizing between the MS-generated candidates, as well as for UV-guided discovery of novel compounds [13,14].
This article reviews anticancer and often also antifungal natural products primarily produced by Aspergillus, Penicillium and Talaromyces. For practical reasons the compounds have been grouped into major biosynthetic classes, according to the biosynthetic origin of the core part of the structures, despite that many compounds are actually of mixed biosynthetic origin (e.g., prenylated). This review is broader in scope than other recent reviews only focusing on endophytes [15,16], and a strong focus has been on giving the correct identity of the species reported in the literature, where many may previously have been misidentified or names have been changed according to the 2011 Amsterdam Declaration on Fungal Nomenclature [17].

Polyketide-Derived Anticancer Compounds
Polyketides represents one of the major classes of natural products of which many are biological active [1]. Today much is known about the enzymes involved in the biosynthesis of a huge diversity of both non-reduced (aromatic), partly reduced and highly reduced polyketides [8].
Some famous fungal polyketides with anticancer activity belong to the statin family. The statins are well known cholesterol synthesis inhibitors that are used in clinical treatment of hypercholesterolemia and cardiovascular diseases [18][19][20][21]. Moreover, members the statin family are known to have antifungal properties against Aspergillus spp. and Candida spp. [18,[22][23][24]. The statin family includes a long list of both natural and synthetic compounds, for example the naturally derived compactin, lovastatin and pravastatin. The statin structure is based on a dicyclohexene ring system connected to a dicyclohexene ring system connected to a side chain with a closed lactone ring or an open acid form [21]. The compactins are primarily produced by P. solitum and P. hirsutum [20,25] (first misidentified as P. brevicompactum [18] and a fungus identified as P. citrinum [19]). Another group of statins that has an extra methyl group attached on the dicyclohexene ring system are produced by A. terreus [26] and Monascus spp. [27]. Several in vitro activities of the statins have been published throughout the years. Reports showed that compactin (Figure 1a) inhibited acute myeloid leukemia (AML) cells with a full inhibitory concentration (IC 100 ) of 2.6 µM [28]. The analogs lovastatin ( Figure 1b) and simvastatin ( Figure 1c) have been shown to be even more potent. Lovastatin and the synthetic simvastatin selectively inhibited colony growth of primary AML cells with 75%-95% effectiveness. No effect was seen on normal bone marrow [29]. The more recent reported activities includes reduction of proliferation by lovastatin in four lung cancer cell lines with median inhibitory concentration (IC 50 ) values between 1.5 and 30 µM [30]. In 2010 it was shown that lovastatin induced apoptosis in ten ovarian cancer cell lines tested, with IC 50 values between 2 and 39 µM [31], and recently lovastatin was found to inhibit breast cancer MCF-7, liver cancer HepG2, and cervical cancer HeLa cell lines with IC 50 values of 0.7, 1.1 and 0.6 µg/mL, respectively [32]. Simvastatin inhibited two lung cancer, three melanoma, and four breast cancer cell lines with IC 50 values between 0.8 and 5.4 µM and induced apoptosis with reduced tumor growth in hepatic cancer cells [33,34]. Encouraged by these results simvastatin has entered clinical trials as an anticancer drug [35].
Asperlin (Figure 2c) was isolated from A. nidulans in 1966 [46]. In 2011 it was discovered that asperlin reduces cell proliferation and induce G 2 /M cell cycle arrest in the human cervical carcinoma HeLa cell line [47].
Filamentous fungi are known to produce anticancer polyketides with different spiro ring structures. One of the more well-known is the antifungal [48,49] compound griseofulvin ( Figure 3a) from P. griseofulvum [50]. Griseofulvin was introduced commercially in 1965 and first considered for cancer treatment in 1973 [49,51]. Later it was shown to induce cell proliferation and mitosis in the human cervical cancer cell line HeLa with an IC 50 value of 20 µM, as well as it inhibits centrosomal clustering in human squamous cancer SCC-114 cell line with an IC 50 value of 35 µM [2,3]. The synthetic analog GF-15 increased the inhibitory effect of centrosomal clustering in SCC-114 cells 25-fold with an IC 50 value of 0.9 µM [5]. It was further shown that combined treatment of griseofulvin and the cancer chemotherapeutic agent nocodazole in vivo improved the effect of nocodazole and arrested tumor growth in mice infected with COLO 205 tumors [4].
The two epimers chloctanspirone A and B (Figure 7a) are probably produced by P. chrysogenum or P. rubens [73,74] (originally published incorrectly as P. terrestre [75]). Chloctanspirone A and B were the first chlorinated sorbicillinoids isolated from a natural source and the first to be identified with their very unique ring structure. Chloctanspirone A was the more active analog and inhibited human leukemia HL-60 and lung cancer cell line A-549 cell lines with IC 50 values of 9.2 and 39.7 µM, respectively [76]. Chloctanspirone B however, only showed moderately or no activity against the same cell lines. Interestingly, the two precursors terrestrols K and L (Figure 7b) that contain the epicenter, which dissociates chloctanspirone A from B were inactive. This points to the conclusion that the cyclohexenone moiety has an impact on the activity though it is not the pharmacophore [76]. A group of around 20 p-terphenyls were isolated from A. candidus and found to be cytotoxic against the human cervical cancer cell line HeLa [77]. Thirty years later a new group of prenylated-p-terphenyls, called prenylterphenyllins (Figure 8a), terprenins (Figure 8b), and prenylcandidusins (Figure 8c) were isolated from A. candidus and A. taichungensis [78,79]. The activity of these compounds were studied against lung cancer cell line A-549, human leukemia cell lines HL-60 and P-388, and human epidermoid cancer cell lines KB3-1 [78,79]. Prenylterphenyllin A was found to be the more active against A-549 and HL-60 with IC 50 values of 8.3 and 1.5 µM, respectively [79], whereas 4′′-deoxyisoterprenin displayed higher activity towards KB3-1 with an IC 50 value of 6.2 µM [78], and finally prenylcandidusin B showed higher activity against P-388 with an IC 50 value of 1.6 µM [79]. In 2011 five new di-and tricitrinols were added to the known citrinin family from P. citrinum [80]. All five of them showed cytotoxic activity against human leukemia HL-60, colon cancer HCT-116, and cervical cancer KB cell lines. The more potent was tricitrinol B (Figure 9), with IC 50 values of 3.2, 4.8 and 3.9 µM, respectively [80]. The class of perylenequinones has been used for centuries in Chinese herbal medicine, but fungal production of several perylenequinones has likewise been reported [81]. The calphostins were isolated from Cladosporium cladosporioides and shown to inhibit cervical cancer HeLa-S 3 and breast cancer MCF-7 cell lines. Most potent was calphostin C (Figure 10a) with IC 50 values of 0.23 and 0.18 µM, respectively [82]. Furthermore, calphostin C induced apoptosis in acute lymphoblastic leukemia [83]. The hypocrellins, another branch, of the perylenequinone family was isolated from Hypocrella bambusae, originally published incorrectly as Shira bambusicola [84,85]. Hypocrellin D was one of the more potent analogs and inhibited liver (Bel-7721) and lung (A-549 and Anip-973) cancer cell lines, with IC 50 values, 1.8, 8.8, 38.4 µg/ml, respectively [85]. Additionally, the hypocrellins had a photodynamic effect on a wide range of tumor cell lines. Due to their insolubility in water several analogs have been synthesized to improve the pharmacokinetics and drug delivery. Two of the more successful analogs were carboxylate salt derivatives of hypocrellin B (Figure 10b) with increased water solubility and activity against human breast cancer MCF-7 when photodynamic therapy was used [81,86]. In 2012 a new group of perylenequinones called the phaeosphaerins were isolated from Phaeosphaeria sp. with phaeosphaerin B (Figure 10c) as the more potent against prostate PC-3, DU-145, LNCaP cancer cell lines with IC 50 values of 2.4 µM, 9.5 µM and 2.7 µM, respectively [87].

Nitrogen-Containing Anticancer Compounds
Fungal nitrogen containing compounds represent another large group of natural products, of which many have famous bioactivities. In general these types of compounds incorporate amino acid building blocks into often complex heteroaromatic compounds such as diketopiperazines, quinazolines and benzodiazepines [1]. Often these compounds are referred to as alkaloids, due to their basic nature, when containing primary, secondary or tertiary amine functionalities. However, compounds only containing amide bonds, that are essentially neutral, are often also referred to as alkaloids. In recent years it has become clear that many nitrogen containing compounds are biosynthesized by multifunctional enzymes (non-ribosomal peptide synthases, NRPS) with modular arrangements comparable to that seen for some polyketide synthases. In such contexts these compounds are usually referred to as non-ribosomal peptides (NRPs) even though they are also called alkaloids [88]. In the following section no attempt has been made to differentiate between the different naming of the numerous amino acid-derived metabolites.
Two other tryptophan/proline diketopiperazines are stephacidin A (Figure 13b) and B produced by A. ochraceus and A. westerdijkiae [106,107]. Stephacidin B is the dimeric form of stephacidin A and was approximate 10-fold more potent. Stephacidin B exhibited inhibitory effect on prostate, ovarian, colon, breast and lung cancer cell lines with IC 50 values between 0.06 and 0.4 µM [106].
The notoamides (Figure 13c) produced by A. amoenus are close analogs to the stephacidins, but with a spiro ring structure [115]. Contrary to the stephacidins the notoamides only showed low to moderate inhibition of two leukemic cell lines with IC 50 values 22-52 µg/mL [116].  [118]. A synthetic analog plinabulin (NPI-2358) displays very potent activity against human prostate carcinoma cell line DU-145 and has now entered phase II clinical trials [119]. Recently, more than 60 synthetic analogs of plinabulin have been designed and synthesized. The more active analog with a benzoyl group coupled on the phenylalanine unit was 10-fold more active than plinabulin with an IC 50 value of 1.4 nM [119]. The activity of phenylahistin was also compared to the diketopiperazine aurantiamine and other synthetic analogues [120], originally isolated by Larsen et al. [121]. Another group of diketopiperazines with anticancer activity contain a di-sulfide bridge in the diketopiperazine ring. One of them the antifungal, immunosuppressive and antimicrobial compound gliotoxin (Figure 15a) that was isolated from A. fumigatus, and D. cejpii [122][123][124]. Already in 1947 the anticancer activity of gliotoxin was suggested and in 2004 it was found that gliotoxin was a very potent inhibitor of six breast cancer cell lines with IC 50 values between 38 and 985 nM [125,126]. In 2012, gliotoxin was found active against human leukemia U-937 and human prostate cancer PC-3 cell lines with IC 50 values of 0.2 and 0.4 µM, respectively [112]. Another diketopiperazine with a di-sulfide bridge is the antifungal [127] compound emestrin A (Figure 15b) that was isolated from Emericella striata, now called A. striatus [128]. Emestrin A inhibits human leukemia (HL-60) with an IC 50 value of 83.5 nM [129]. Eight structural analogs of emestrin A were isolated from Cladorrhinum sp. and found to have strong antiproliferative effects on the human prostate DU-145 cancer cell line with an IC 50 value of 9.3 nM for the more potent emestrin C. Furthermore, it was proven that the activity decreased when the macro cyclic ring was opened and the polysulfide bridge in the diketopiperazine was absent [130].
The peptide leucinostatin A (Figure 16) was first isolated in 1973 from Purpureocillium lilacinum (originally published as Penicillium lilacinum or Paecilomyces lilacinus) [132,133]. It was found active against a number of fungi and Gram-positive bacteria, as well as Ehrlich solid carcinoma of mice with an ED 50 value of 1.6 mg/kg [132,134]. Furthermore leucinostatin A inhibited a long range of breast, melanoma, lung, ovary, colon and laryngeal cancers as well as eight leukemia cell lines with IC 50 values between 4 nM and 12 µM [132,135,136]. Leucinostatin A inhibited growth of prostate cancer DU-145 cells in vitro and in vivo [137]. A natural analog leucinostatin A β di-O-glycoside displayed active against breast cancer BT-20 cell line, though not as potent as leucinostatin A [138].

Terpenoid-Derived Anticancer Compounds
Terpenoids form a third large and structurally diverse biosynthetic family of natural products derived from C 5 isoprene units. Until now active anticancer terpenoids mostly belong to the sesqui-or diterpenoids [15].
Taxol, also known as paclitaxel (Figure 17), is one of the best known anticancer drugs produced by fungi, though it originally was isolated from the bark of yew tree Taxus brevifolia [139]. Clinical development of taxol was delayed due to problems with production of large enough quantities of the compound. This problem was solved 20 years later when it was demonstrated that taxol was also produced by the fungus Taxomyces andreanea [140] and later including P. raistrickii [141]. Taxol was approved as anticancer drug against a wide range of tumors in the 1990s and is the first billion dollar drug against cancer [142,143]. Taxol functions by inducing cell cycle arrest in G 2 /M phase as well as apoptosis through a unique mode-of-action by promotion and stabilization of tubulin polymerization [144]. Today, taxol is routinely used to treat ovarian, breast and lung tumors as well as Kaposi's sarcoma [145]. Besides its anticancer activity taxol displays antifungal activity as well [146].
Many sesquiterpenes have been found active against cancer. Two of these are the drimane sesquiterpenes fudecadione A and B (Figure 18a) that were isolated in 2011 from Penicillium sp. BCC 17468. Fudecadione A was found active against human lung cancer NCI-H187, human breast cancer MCF-7, and human oral epidermoid carcinoma KB cell lines, with IC 50 values of 24.9, 12.6 and 22.6 µM, respectively. Fudecadione B, on the other hand, was inactive against all three cancer call lines [147]. The activity of the two compounds suggests that the pharmacophore is located around the carbon at position 13, where fudecadione B is more branched [147].
In 2013 a novel and as yet unnamed chlorotrinoreremophilane sesquiterpene (Figure 18c) was isolated from Penicillium sp. PR19N-1 [150]. This novel sesquiterpene exhibited cytotoxic activity against human leukemia HL-60 and lung cancer A-549 cell lines with IC 50 values of 11.8 and 12.2 µM, respectively [150].
A large group of mycotoxins called the trichothecenes cover more than 150 analogs and are mainly produced by a number of Fusarium spp. [151]. All the trichothecenes contain a sesquiterpenoid ring structure with an epoxide. The epoxide is often responsible for the cytotoxic activity by binding to the 60S ribosomal subunit in eukaryote cells thereby inhibiting protein synthesis [152,153]. Many of the trichothecenes exhibit cytotoxic activity against both fungi and cancer cell lines [154][155][156]. One of the more potent is AETD (Figure 19a) that inhibits HL-60, U-937, HeLa, MCF-7 and Hep-G2 cell lines with IC 50 values of 10, 22, 45, 53 and 170 nM [157]. Other bioactive groups are the roridins where a macrocyclic ring is connected to the sesquiterpenoid unit. One of the more active compounds of this group is 12′-hydroxyroridin E (Figure 19b), which inhibits leukemia L-1210 with an IC50 value of 0.2 µM [158]. Another trichothecene called anguidine even entered clinical trials against cancer, but did not progress beyond phase II due to a lack of therapeutic efficacy [152,159].  The antifungal compound wortmannin (Figure 20b) is produced by T. wortmannii originally published as P. wortmannii [162] or a related species. Wortmannin inhibit the activity of leukemia HL-60 and K-562 cell lines with IC 50 values of 30 and 25 nM, respectively [163,164], including the breast cancer MCF-7 cell line that was inhibited by 51.3% after 48 h with a concentration of 25 nM [165]. In human pancreatic cancer cells lines PK1 and PK8 wortmannin induces apoptosis both in vitro and in vivo [166,167]. Finally, wortmannin has been shown to inhibit proliferation in lung cancer cell lines KNS-62 and Colo-699 both in vitro and in vivo, with IC 50 values between 100 and 200 nM [168].

Anticancer Natural Products of Mixed or Unresolved Biosynthetic Origin
Filamentous fungi are capable of producing secondary metabolites of mixed biosynthetic origin. This includes both compounds such as meroterpenoids that comprise a huge class of compounds that integrate a polyketide part with a terpenoid part [182], in addition to the cytochalasins [183] and chaetoglobosins that are biosynthesized by incorporation of amino acids into a core polyketide part. The cytochalasins contains a phenylalanine coupled to the polyketide chain where the chaetoglobosins have an tryptophan moiety [183,184]. Both the cytochalasins and the chaetoglobosins exhibit antifungal activities against a broad range fungal species [56,185,186]. The cytochalasins are produced by many fungal genera including Aspergillus, Hypoxylon, Metarrhizium, Zygosporium, Hypocrella and Phoma [90,[187][188][189][190]. Many of the cytochalasins have shown inhibitory activities towards lung cancer A-549. One of the more potent is cytochalasin E (Figure 22a) which inhibited human ovarian A-2780S, human colon HCT-116 and SW-620, and lung A-549 cancer as well as human leukemia P-388 with IC 100 values of 0.02, 1.0, and 0.2 µg/ml and IC 50 values of 0.006 and 0.09 µM, respectively [191,192]. The chaetoglobosins were originally isolated from Chaetomium globosum in 1973 [184]. Several chaetoglobosins have been isolated over the years from P. discolor and P. expansum among others [1] and many of them showed activity against cancer cell lines. Of the more potent ones was chaetoglobosin B (Figure 22b) that inhibited human breast BC cancer cell line with an IC 50 value of 3.0 µM [193] and chaetoglobosin D that inhibited adenocarcinoma KKU-100 and KKU-OCA17 cancer cell lines with IC 50 values of 3.4 and 12.2 µM, respectively [193]. Chaetoglobosin U showed activity as well against KB tumor cell line with an IC 50 value of 16.0 µM [194], and chaetoglobosin X with activity against murine hepatic cancer H-22 cell line with an IC 50 value of 7.5 µM [195].

Conclusions
In this article we have reviewed 50 compounds or compound families with anticancer and often also antifungal activities, primarily produced by Aspergillus, Penicillium and Talaromyces. Mycologists predict that less than 10% of all fungal species have been isolated so far, indicating a huge potential for further discovery of novel bioactive chemical scaffolds, if these fungi can be cultured in the laboratory. New strategies such as the application of epigenetic modifiers may help to uncover the full biosynthetic potential of fungi and other microorganisms. Development and improvement of screening methods and assays will further assist revealing new bioactivities of already known compounds. Additionally, ongoing progress in fast dereplication, including improvement of the performance of mass spectrometers and high resolution of UHPLC chromatograms, will ensure that database searches will lead to fewer possible candidates thereby advancing the drug discovery process. Altogether the future seems promising for discovery of many more bioactive small molecules to be used either as scaffolds for: (i) diversity oriented synthesis, or (ii) as a starting point for cloning and engineering of whole biosynthetic gene clusters towards novel engineered bioactive natural products.