North African Medicinal Plants Traditionally Used in Cancer Therapy

Background: Cancer is a major cause of mortality worldwide with increasing numbers by the years. In North Africa, the number of cancer patients is alarming. Also shocking is that a huge number of cancer patients only have access to traditional medicines due to several factors, e.g., economic difficulties. In fact, medicinal plants are widely used for the treatment of several pathologies, including cancer. Truthfully, herbalists and botanists in North African countries prescribe several plants for cancer treatment. Despite the popularity and the potential of medicinal plants for the treatment of cancer, scientific evidence on their anticancer effects are still scarce for most of the described plants. Objective: Bearing in mind the lack of comprehensive and systematic studies, the aim of this review is to give an overview of studies, namely ethnobotanical surveys and experimental evidence of anticancer effects regarding medicinal plants used in North Africa for cancer therapy. Method: The research was conducted on several popular search engines including PubMed, Science Direct, Scopus and Web of Science. The research focused primarily on English written papers published between the years 2000 and 2016. Results: This review on plants traditionally used by herbalists in North Africa highlights that Morocco and Algeria are the countries with most surveys on the use of medicinal plants in folk medicine. Among the plethora of plants used, Nigella sativa and Trigonella foenum-graecum are the most referred ones by herbalists for the treatment of cancer. Moreover, a plethora of scientific evidence qualifies them as candidates for further drug development. Furthermore, we report on the underlying cellular and molecular mechanisms. Conclusion: Overall, this review highlights the therapeutic potential of some medicinal plants as anticancer agents. The North African flora offers a rich source of medicinal plants for a wide array of diseases, including cancer. The elucidation of their modes of action represents an indispensable condition for the rational development of new drugs for cancer treatment. Furthermore, testing the anticancer activity in vivo and in clinical trials are warranted to explore the full therapeutic potential of North African plants for cancer therapy.


INTRODUCTION
According to the World Health Organization (WHO) cancer represents a major cause for morbidity and mortality with ∼14 million new cases in 2012 and 8 million cancer-related deaths (Forman and Ferlay, 2014). However, this burden is expected to even increase to 75 M. prevalent cases, 27 M. incident cases and 17 M. cancer-related deaths by 2030 (Adeloye et al., 2016). The cancer prevalence is gender-dependent with men presenting higher incidence rates for tumors of the lung (16.7%), prostate (15.0%), colorectum (10.0%), stomach (8.5%) and liver (7.5%), while women reveal more cases of breast (25.2%), colorectum (9.2%), lung (8.7%), cervix (7.9%), and stomach (4.8%). North Africa and Middle-East acquaint worldwide for 3.8% of the new cancer cases and for 4.1% for cancer-related deaths. According to WHO, the raw incidence of cancer in Morocco was of 123.1 per 100,000 habitants in men and of 77.5 per 100,000 in women (Observatory) 1 in 2012. Throughout all North African countries, the prevalence in men are higher than in women, with Egypt as country with the highest prevalence (145.9 for men and 100.5 for women) and Tunisia with the lowest prevalence (96.6 for men and 52.9 for women).
Since immemorial times, human beings acquired knowledge on the medicinal use of plants (El-Seedi et al., 2013;Ouelbani et al., 2016). Those plants have been extensively applied in folk medicine to treat ailments and diseases (El-Seedi et al., 2013) and are still used in the rural areas of developing countries (El-Seedi et al., 2013;Ouelbani et al., 2016). In fact, World Health Organization (WHO) reported that around 80% of the world population still relies on plants as source for primary health care (Cordell, 1995), while traditional medicine is the only health source available for 60% of the global population (El-Seedi et al., 2013). Medicinal plants are frequently the only form of cancer treatment for many people in North Africa, either due to low income or spatial distance from the urban treatment centers (Kabbaj et al., 2012).
The aim of this review is to compile data about the anticancer potential of plants found in ethnobotanical surveys of North African countries. The information was collected in several scientific research engines, PubMed, Science Direct, Scopus, Web of Science, and Google Scholar comprising studies conducted between 2000 and 2016.
In addition to ethnobotanical surveys, this review also includes experimental evidence on the cytotoxic effects of medicinal plants as well as their cellular and molecular mechanisms in cancer cells.

ETHNOBOTANICAL STUDIES
The ethnobotanical surveys were predominantly found in Morocco and Algeria. In other North African countries (Egypt, Tunisia, and Libya), less is known about the cytotoxic properties of medicinal plants against cancer cells. Table 1 compiles the botanical information, the geographical location, the type of therapy, the cancer types investigated, the plants' parts used (e.g., leaf, aerial parts, seeds), and the preparation method (e.g., infusion, decoction). In those cases, where information was obtained from either the general population or herbalists/botanists we considered cancer therapy as monotherapy (i.e., only medicinal plants). If information was gathered from patients in cancer treatment centers the therapy was considered to be a combination of complementary and standard chemo/radiotherapy (co-therapy). The most predominant botanical families used as anticancer agents were Lamiaceae (13 species), Apiaceae (9 species), Compositae (8 species), and Fabaceae (6 species; Figure 1). Two surveys conducted at the National Institute of Oncology in Rabat (Morocco) showed that the most used plants by the patients were  (Kabbaj et al., 2012;Chebat et al., 2014). Recently, a review was conducted on the anticancer potential of plants used in the Arabian and Islamic world (Ahmad et al., 2016) which included N. sativa. However, this work did not mention the countries in which they are used.
The plant parts used as well as the preparation methods of anticancer agents were as diverse as the plants themselves. The most common plant parts were the seeds (23) followed by the aerial parts (20) and leaves (19) (Figure 2). Regarding the preparation methods, the most described are decoction (25), grinded with honey (20) and infusion (17) (Figure 3).

SCIENTIFIC EVIDENCE AND MECHANISMS OF ACTION
Despite the widespread use of medicinal plants in North Africa, many species still lack scientific prove of their anticancer activity. In fact, Kabbaj et al. (2012) described that 55 plants are used by patients at the National Institute of Oncology at Rabat (Morocco) albeit only 28 have been previously described for their cytotoxic properties against tumor cells. This chapter focusses, first and foremost, on (1) in vitro cytotoxicity against cancer cell lines for both volatile and non-volatile extracts and isolated major phytochemicals and (2) in vivo assays in those cases, where such studies were carried out. Afterwards, potential mechanisms of action for both extracts and isolated compounds will be reviewed, mainly cell cycle arrest, cell death induction and signal transduction pathways as well as invasiveness and migration of cancer cells.

In vitro Cytotoxicity Assays
Of all used plants in North African folk medicine, N. sativa is one of the scientifically best analyzed. A plethora of cancer cell lines have been used for the determination of cytotoxicity of medicinal plants. Of those, breast cancer cell lines (MCF-7, MCF-7/Dox and MCF-7/Topo; MDA-MB-231), colon carcinoma (HCT 116), hepatocellular carcinoma (HepG2), cervix carcinoma (Hep-2),     prostate cancer (PC-3), and lung carcinoma (A549) are the most commonly used ones.
Trigonella Foenum-Graecum, Diosgenin, and Protodioscin T. foenum-graecum (Fabaceae) is known as "helba" in Arabic (Hammiche and Maiza, 2006) and fenugreek in English (Abdel- Barry et al., 1997). This plant have been described as possessing several pharmacological activities, including anticancer (Yadav and Baquer, 2014). In fact, a recent paper reviews the anticancer potential of both T. foenum-graecum extracts and isolated compounds (El Bairi et al., 2017) describing the effects on the major hallmarks of cancer, sustaining proliferative signaling, angiogenesis, cell death evasion, tumor promoting inflammation, invasion, and metastasis and genomic instability. A water extract obtained from the seeds was cytotoxic toward several cancer cell lines (T-cell lymphoma, B-cell lymphoma, thyroid papillary carcinoma, and breast cancer; Alsemari et al., 2014). A crude methanol extract from the seeds induced cell death in a dose-dependent manner (IC 50 = 1,000 µg/mL; Khalil et al., 2015), however, significant inhibition was achieved only at concentrations of 100 µg/mL. Seed extracts were cytotoxic to a plethora of breast, pancreatic and prostate cancer cell lines (Shabbeer et al., 2009). Crude extracts of fenugreek decreased the cell viability toward both drug-sensitive and drug-resistant cancer cell lines (Saeed et al., 2015). Methanol extracts from the whole plant induced cytotoxicity in MCF-7 breast cancer cells (IC 50 = 65 µg/mL; Alshatwi et al., 2013). In the same cell model, a chloroform extract also induced cytotoxicity (IC 50 = 41.6 µg/mL; Khoja et al., 2011). Both aqueous and ethanolic extracts exerted cytotoxic effects against MCF-7 cells in a doseand time-dependent manner (Sebastian and Thampan, 2007) Diosgenin (Figure 4, 2), the major compound of fenugreek, reduced cell proliferation in a time-and dose-dependent manner in HT-29 colon cancer cells (Raju et al., 2004). Furthermore, it induced cell death to several breast, pancreatic and prostate cancer cell lines (Shabbeer et al., 2009). In PC-3 prostate cancer cells, it inhibited the growth at 50 µM in a timedependent manner (Chen et al., 2011). Diosgenin (25 µM) induced cytotoxicity and inhibited cell proliferation in KBM-5 cells (Shishodia and Aggarwal, 2006). Furthermore, co-treatment with chemotherapeutic agents potentiated the cytotoxic effect of the latter. The saponin revealed inhibitory effects on several squamous cell carcinomas (Das et al., 2012). Diosgenin inhibited the growth of A549 lung cancer cells in a dose-and timedependent manner (47 and 43 µM after 72 h, respectively; Rahmati-Yamchi et al., 2014). Treatment with diosgenin induced cell death in HEL erythroleukemia cells (90% inhibition at 40 µM after 48 h; Leger et al., 2004). In a different leukemia model, K562, it also induced cytotoxicity (IC 50 = 15 µM; Liu et al., 2005). The 1547 osteosarcoma cell line is susceptible to this compound (86% inhibition at 40 µM for 24 h; Moalic et al., 2001).

In vivo Cytotoxicity
In vitro results can be taken as first clue for bioactivity, but they do not necessarily reflect activity in living organisms. Biotransformation in the liver degrades many natural products so that they finally lose activity in vivo (Reichling et al., 2009).

Nigella sativa and Thymoquinone
The essential oil of N. sativa reduced solid tumor volume in mice in a dose-dependent manner (2.5 cm 3 for untreated vs. 0.22 and 0.16 cm 3 for 30 and 50 µL/mouse, respectively). In addition, it reduced cancer metastasis (14 for untreated mice vs. 2 and 0 for 30 and 50 µL/mouse, respectively) and improved survival of mice (83% mortality in untreated group vs. 16.6 and 0% for 30 and 50 µL/mouse, respectively; Ait Mbarek et al., 2007). Thymoquinone significantly decreases the tumor size and mass in a sarcoma 180bearing mice xenograft model (Das et al., 2012). In addition, it inhibited angiogenesis as observed by a decrease of CD31 expression.

Trigonella Foenum-Graecum and Diosgenin
An aqueous extract from the seeds of fenugreek slowed the progression of DMBA-induced breast cancer (40 vs. 80% incidence; Amin et al., 2005). Furthermore, it decreased the mean tumor number (2.0 vs. 3.5) and weight (3.0 g vs. 5.0 g) in rats. The histology was also markedly different with the extract-treated rats bearing only mild-to-moderate hyperplasia. Supplementation with 0.01% fenugreek seed powder for 8 weeks decreased total colonic aberrant crypt foci, while continuously feeding rats with fenugreek seed powder decreased multicrypt foci in azoxymethane-induced aberrant crypt foci formation (Raju et al., 2004). These effects were observed both in the initiation and promotion stages of carcinogenesis. Seed extract was given at pre-initiation, post-initiation, promotion, and throughout all stages to a DMBA+TPA-induced mouse skin carcinogenesis model (Chatterjee et al., 2012). The treatment decreased the cumulative papilloma count compared to control. Furthermore, it decreased cancer incidence and tumor burden. In addition, it decreased the latent period. A methanol extract reduced the number of mice bearing tumors (13 vs. 18 for untreated control), the tumor incidence (65 vs. 90% for untreated), the average number of tumors (102 vs. 221 for untreated control mice) and tumor multiplicity (7.84 vs. 12.27 for untreated) in a DMBA-TPA-induced skin carcinogenesis model (Ali et al., 2014). Pretreatment of mice inoculated with Ehrlich ascites carcinoma (EAC) with an alcoholic extract of fenugreek decreased the number of EAC cell count per mouse (0.208 × 10 6 at 200 mg/kg i.p. vs. 4.42 × 10 6 ; Sur et al., 2001). Furthermore, post-treatment also decreased the EAC cell count per mouse (0.41 × 10 7 at 200 mg/kg i.p. vs. 1.68 × 10 7 ).
The supplementation with 0.1% diosgenin, the major compound of fenugreek, decreased both total colonic aberrant crypt foci and multicrypt foci in azoxymethane-induced aberrant crypt foci formation on initiation and promotion stages (Raju et al., 2004). Diosgenin is able to decrease tumor mass and size in mice bearing sarcoma 180 (Das et al., 2012). Furthermore, it decreased the expression of CD31, suggesting that diosgenin revealed antiangiogenic effects in vivo.

Cell Cycle Perturbations
The effect of cytotoxic compounds is frequently measured by detection of cellular DNA with propidium iodide and subsequent measurement by flow cytometry. The resulting DNA histograms show the distribution of cell populations in the cell cycle. Anticancer compounds at subtoxic concentrations frequent led to an arrest of the cell cycle in the G1, S, or G2/M phase resembling cytostatic effects. Cytotoxicity of compounds results in cell death, which appears as dead cell fragments below the G0/G1 fraction. Hence, flow cytometric measurements can deliver data on both cytostatic and cytotoxic effects of anticancer substances. Cell cycle is a highly-regulated process by the sophisticated activity of cyclin-dependent kinases (CDKs). This type of kinases are activated by the presence of cyclins and de-activated by the presence of CDK inhibitors (Bassermann et al., 2014). The type of CDK-cyclin complex is dependent on extracellular signals as well as intrinsic information. The CDK inhibitors belong to two different families, the INK4 and CIP/KIP proteins. The first group interacts with CDK4/6 kinases inhibiting the capacity to bind to D-type cyclins, e.g., p16 INK4a , while the latter binds to CDK-cyclin complexes, such p21 Cip1 and p57 Kip2 (Yun et al., 2013;Ruijtenberg and van den Heuvel, 2016). CDK inhibitors such as p21, p27, and p53 are tumor suppressor genes that block both cell proliferation and migration (Yun et al., 2013). Due to the importance of cell cycle progression for tumor development, several authors have described the cell cycle arrest capacity of plant extracts as well as isolated major compounds. In Figure 5 the effects of the extracts and major phytochemicals (thymoquinone, diosgenin and protodioscin) on cell cycle are represented.

Thymoquinone
To the best of our knowledge, no studies addressing the cell cycle arrest capacities of whole extracts or oils from N. sativa were conducted to the date. Thus, only studies with its major compound (thymoquinone) will be referred. Thymoquinone was able to cause cell cycle arrest at S phase (Khalife et al., 2016). By contrast, Woo et al. (2011) described that this compound caused accumulation of cells in the sub-G1 phase without effect on p53 activation. This compound caused cell cycle arrest at G1 after 24 h, but accumulated cells in the sub-G1 phase after 48 h treatment. Furthermore, thymoquinone induced the expression of p27 and p53, but downregulated the expression of cyclin D1 and E (Rajput et al., 2013). It caused the accumulation of sub-G1 cells (25.6 vs. 1.3% in untreated cells; Arafa et al., 2011). In p53-defective Jurkat lymphoblastic leukemia cells, the compound caused cell accumulation in sub-G1 (Alhosin et al., 2010). Gurung et al. (2010) described an accumulation in the sub-G1 phase in DNA-PKcs-normal (M059K) and -defective (M059J) glioblastoma cell lines. Thymoquinone also increased the sub-G1 fraction in U266 multiple myeloma cells (Li et al., 2010).

Trigonella Foenum-Graecum and Diosgenin
A crude methanol extract of fenugreek seeds drove cells into the sub-G1 (∼49.1%) and G1 phase (65% cells) in HepG2 (Khalil et al., 2015). In pancreatic cancer cells, fenugreek exerted two different effects. In both LNCaP and PC-3 cell lines, it increased accumulation of cells in sub-G1, while in PC-3 cells it also caused cell cycle arrest in the G2/M phase (Shabbeer et al., 2009). An ethanolic extract caused an increase of the sub-G1 fraction (30.1% at 100 µg/mL) as well as a cell cycle arrest at G2/M (Sebastian and Thampan, 2007). Diosgenin, a saponin isolated from fenugreek, caused sub-G1 accumulation after 48 h at 50 µM (51 vs. 9% for untreated cells; Shishodia and Aggarwal, 2006). Similarly, the treatment of squamous cell carcinomas with diosgenin caused an increase of the sub-G1 fraction (Das et al., 2012). The saponin causes cell cycle arrest at G2/M phase as soon as after 6 h of treatment (22.9 vs. 17.6%) due to an increase in the expression of p21, while after 24 h of treatment there was also an accumulation in the sub-G1 phase (Leger et al., 2004). After 36 h treatment diosgenin (20 µM) caused cell cycle arrest at G2/M with concomitant decrease in cyclin B1 and p21, and increase in cdc2 in K562 (Liu et al., 2005). In NB4 cells, diosgenin also caused arrest at the G2/M phase with an increase in p53 expression. Furthermore, after 48 h the arrest changed to sub-G1 phase accumulation simultaneously with DNA fragmentation. Diosgenin treatment up to 24 h, arrested cells in the G1 phase (50 vs. 35%) with an associated decrease of cells in the S phase (21 vs. 46%) in 1547 osteosarcoma cells (Moalic et al., 2001). At 48 h, a sub-G1 peak appeared. Concomitantly, an upregulation of p21 and p53 was observed.

Marrubium vulgare
An ethanolic extract of M. vulgare caused DNA fragmentation in a time-dependent manner (Paunovic et al., 2016) in B16 and U251 cell lines. Furthermore, it caused cell cycle arrest at the S and G2/M phases.

Cell Death Induction
Apoptosis is a genetically programmed cell death with two main pathways, an intrinsic pathway characterized by the release of cytochrome c and posterior activation of caspase-9, and an extrinsic pathway characterized by the ligation of a death ligand (Walsh, 2014). While apoptosis controls whole cells, autophagy is a catabolic process in which intracellular organelles and macromolecules undergo destruction and recycling (Lapierre et al., 2015). In Figure 6 the effects of extracts from North African plants and isolated compounds (thymoquinone, diosgenin and protodioscin) in cell death by apoptosis are shown.

Nigella Sativa and Thymoquinone
A supercritical carbon dioxide extract obtained from the seeds of black cumin increased caspase-3 and-7 levels by 2-and 1.5-fold at 100 and 80 µg/mL, respectively, in MCF-7 breast cancer cells. At 100 µg/mL, it also increased caspase-8 levels by 2-fold. The extract increased the apoptotic index (5.65% untreated cells vs. 52.37% at 60 µg/mL; Baharetha et al., 2013). Thabrew et al. (2005) demonstrated that an aqueous extract inhibited DNA (IC 50 = 3 mg/mL) and protein synthesis in HepG2 cells. A nanoemulsion of essential oil caused cell death by either apoptosis (∼44% after 24 h) and necrosis (∼28% after 24 h; Periasamy et al., 2016). Thymoquinone, the major compound of N. sativa, caused DNA fragmentation, induced apoptosis (36% vs. 2.5% after 48 h), induced the expression of caspase-8 precursor, and downregulated procaspases-3 (Khalife et al., 2016). Furthermore, decreased the levels of procaspase-8, -9 and -7 as well as the protein and mRNA levels of Bcl-2, Bcl-xL and survivin in a time-dependent manner while increasing the Bax/Bcl-2 ratio (Woo et al., 2011). Dastjerdi et al. (2016) suggested that thymoquinone induced apoptosis in MCF-7 cells in a p53-and time-dependent manner. In addition, this compound caused DNA fragmentation and morphological changes associated with apoptosis and disrupted the mitochondrial membrane potential, decreased the Bcl2/Bax ratio and activated caspases-3, -7, and -9 (Arafa et al., 2011). Furthermore, it triggered the release of cytochrome c from mitochondria, increased the expression of pro-apoptotic Bax proteins and downregulated the expression of the anti-apoptotic proteins Bcl2 and survivin (Rajput et al., 2013). Rooney and Ryan (2005) demonstrated that the proapoptotic effects are mediated by glutathione (GSH) depletion, since the pre-treatment with buthioninesulfoximine, a selective inhibitor of GSH, increased the number of apoptotic cells (31.7 vs. 55.1%). On the other hand, this effect was caspase-3-dependent, since the number of apoptotic cells decreased 4.9-fold after treatment with a caspase-3 inhibitor (41.1 vs. 8.4%). Kundu et al. (2014) reported that it induced apoptosis in HCT116 cells by decreasing the levels of Bcl-2 and Bcl-xL, but increasing Bax levels. Furthermore, it cleaved caspases-9, -7, -3 and PARP and activated caspase-3. Treating a neuroblastoma cell line, Neuro-2a, with thymoquinone downregulated the mRNA and protein levels of Bcl-2, upregulated the mRNA and protein levels of Bax, caused a loss of mitochondria membrane potential and triggered the release of cytochrome c to the cytosol. In addition, it activated caspases-3 and -9 with concomitant increase in PARP, while decreasing the protein levels of XIAP, a selective inhibitor for caspases-3, -7, and -9. All of them are signals that lead to cell death by apoptosis (Paramasivam et al., 2012). By contrast, the compound induced cell death by autophagy inhibition in glioblastoma cells lines (Racoma et al., 2013). The treatment increase LC3-II and p62 without changing the levels of Beclin-1. Concomitantly, vacuolization of cells, disruption of lysosomal membrane and modulation of lysosome location within the cell was observed. Furthermore, the compound activated cathepsin and induced caspase-independent cell death induction. In DNA-PKcs-normal and defective glioblastoma cell lines, the treatment triggered the release of cytochrome c to the cytosol, induced apoptotic cell death rather than necrosis, caused DNA damage, FIGURE 6 | Plants traditionally used in cancer therapy in North Africa as well as their compounds are able to induce apoptosis in several cancer types. Extracts from N. sativa (Ns) promote the activity of caspase-8, -3, and -7. Thymoquinone (T) enhances the activity of caspase-8, the conversion of procaspase-3 in caspase-3, the activation of caspase-9, Bax, cleavage of PARP, release of cytochrome c (cytC), loss of mitochondrial membrane potential and causes DNA damage. Furthermore, it inhibits the activity of procaspase-3, Bcl-2 and Bcl-xL. T. foenum-graecum (Tfg) exerts a stimulatory effect on caspase-3/7, p53, and Bax and causes the loss of mitochondrial membrane potential and DNA damage. Diosgenin (D) causes loss of mitochondrial membrane potential and blocks the activity of Bcl-2 and Bcl-xL. Protodioscin (PD) cause DNA damage. Extracts from A. longa (Al) and M. vulgare (Mv) causes the activation of caspase-3/7 and the cleavage of caspase-9.
decreased telomerase activity and caused disruption of telomeres length maintenance. These effects were more predominant in DNA-PKcs-normal glioblastoma cells (Gurung et al., 2010). Li et al. (2010) described caspase-3-dependent apoptosis in U266 multiple myeloma cells. According to Zubair et al. (2013), thymoquinone-induced cell death in prostate cancer cell lines was mediated through ROS-induced DNA damage since the effect was inhibited in the presence of a copper-chelating agent (neocuproine).

Trigonella Foenum-Graecum and Diosgenin
The treatment of several cancer cell lines (T-cell lymphoma, Bcell lymphoma, thyroid papillary carcinoma, and breast cancer) with an aqueous extract from the seeds of fenugreek induced apoptosis, if compared to the control (Alsemari et al., 2014). Similarly, a crude methanol extract triggered apoptosis by activating caspase-3 and up-regulated the expression of p53 and PCNA in HepG2 cells (Khalil et al., 2015). Furthermore, the expression of Bax and the cleavage of PARP were also increased. In DU-123 cancer cells, fenugreek downregulated the expression of mutant p53 (Shabbeer et al., 2009). On the other hand, the extract inhibited the phosphorylation of EGFR in PC-3 cancer cells. A methanol extract of the whole plant induced apoptosis in MCF-7 cells (46.1 and 58.9% apoptotic cells after 24 and 48 h, respectively; Alshatwi et al., 2013). Furthermore, it increased the expression of apoptotic genes, 0.9-fold increase in caspase-3, 0.25-fold in caspase-8, 0.3-fold in caspase-9, 1.7-fold in p53, 8.8-fold in fas, 0.12-fold in FADD, 0.4-fold in bax and 0.7-fold in bak. These results show that fenugreek induced apoptosis via a fas-dependent pathway, however, independent of FADD, Bax and Bak. A chloroform extract induces apoptosis in MCF-7 cells in a dose-and time-dependent manner (23.2% and 73.8% apoptotic cells after 24 and 48 h, respectively, with 50 µg/mL; Khoja et al., 2011). Treating MCF-7 cells with an ethanolic extract induced apoptosis as observed by the flipping of phosphatidylserine from the inner to the outer phospholipid bilayer, loss of mitochondrial membrane potential, and DNA fragmentation (Sebastian and Thampan, 2007). In addition, the treatment increased the expression of apoptotic genes several folds, e.g., caspase-3 (3.5fold), caspase-8 (5.5-fold), caspase-9 (3.7-fold), p53 (1.4-fold), fas (1.7-fold), FADD (2.6-fold), bax (3.2-fold), and bak (4.4-fold). Treatment of Jurkat leukemia cells with an ethanolic extract induced the formation of vacuoles followed by cell membrane disintegration with a concomitant increase of LC-3 expression, all of which are signals of autophagic cell death (Al-Daghri et al., 2012). The treatment of DMBA-TPA-induced tumors in mice with a methanol extract of fenugreek decreased the number of PCNA-positive nuclei and increased the expression of p53 (Ali et al., 2014).
Diosgenin induced apoptosis in HT-29 colon cancer cells by decreasing Bcl-2 expression, while increasing the expression of caspase-3 (Raju et al., 2004). It induced apoptotic cell death as found by morphological and histological changes as well as chromatin condensation in squamous cell carcinomas. In addition, it decreased procaspase-3 levels, decreased Bcl-2 levels, increased Bax levels and inhibited the phosphorylation of JNK and Akt (Das et al., 2012). In A549 lung cancer cells, diosgenin decreased the mRNA levels of hTERT (Rahmati-Yamchi et al., 2014). This saponin disrupted mitochondrial membrane potential as well as the intracellular calcium concentration (578 vs. 210 nM after 12 h; Leger et al., 2004). Furthermore, it increased the Bax/Bcl-2 ratio, caused PARP cleavage with concomitant DNA fragmentation. The increase in intracellular calcium caused a translocation to the membrane and subsequent cPLA 2 activation. In K562 leukemia cells, it also increased the intracellular calcium concentrations, hyperpolarization of mitochondrial membranes after 24 h followed by depolarization after 48 h (Liu et al., 2005). In addition, it activated caspase-3, but decreased the expression of Bcl-2 and Bcl-xL and increased Bax expression. Diosgenin induced apoptosis with an increase in hsp70 mRNA expression and in Bax/Bcl-2 ratio (Moalic et al., 2001).
Protodioscin induced apoptosis in HL-60 leukemic cells with typical morphological changes associated with apoptosis (increased numbers of apoptotic bodies, number of cells in the hypodiploid phase and DNA fragmentation; Hibasami et al., 2003).

Aristolochia Longa
The treatment of BL41 Burkitt's lymphoma with an aqueous extract from the roots induced apoptosis (20.7 vs. 2.7%; Benarba et al., 2012) with concomitant mitochondrial membrane potential loss (53 vs. 11.6%). Furthermore, it activated caspases-3 and -9 followed by PARP cleavage without affecting the activity of caspase-8. This suggests that A. longa activated the intrinsic pathway of apoptosis.

Marrubium Vulgare
Treating B16 and U251 cells with an ethanolic extract of M. vulgare cause increased fractions of both early and late apoptotic cells (Paunovic et al., 2016). In addition, it activated caspases-9 and -3 with concomitant PARP cleavage. A loss of mitochondrial membrane potential was also observed in both cell lines.

Signal Transduction Pathways
The cellular metabolism is tightly controlled by several pathways. In cancer cells, some of those pathways are dysregulated. NF-κB has been associated with the regulation of several processes such as inflammation, cell growth, and apoptosis (Ghosh and Dass, 2016). This nuclear transcription factor also controls several other biological pathways such as the MAPK or PI3K/Akt pathways. The activation of NF-κB is associated with increased expression of several genes whose products are associated with tumorigenesis, such as antiapoptotic proteins (e.g., survivin, Bcl-2), COX-2, matrix metalloproteinase-9, iNOS, and cell cycle regulation proteins (e.g., cyclin D1; Shishodia and Aggarwal, 2006). Another dysregulated pathway is STAT3. This signal transduction pathway has been associated with cell proliferation, cell survival and angiogenesis (Aggarwal et al., 2009;Becker et al., 2014;Yamamoto et al., 2014). STAT3 is activated by several intrinsic and extrinsic factors such as IL-6, JAKs, ERK, cigarette smoke (Aggarwal et al., 2009). In normal cells, this pathway is only transiently activated and tightly regulated. However, it is constitutively activated in cancer cells (Kortylewski et al., 2005). While NF-κB and STAT3 pathways are activated in cancer cells, PPAR pathway is inactivated, because this pathway negatively controls proliferation and survival (Reka et al., 2011). Furthermore, PPAR inhibits NF-κB and STAT3 pathways. In addition, this pathway is also associated with cell cycle arrest and apoptosis induction (Dicitore et al., 2014). The effects of the extracts and major compounds (thymoquinone and diosgenin) on signal transduction pathways are represented in Figure 7.

Nigella sativa and Thymoquinone
Seed oil obtained from non-heated seeds of N. sativa inhibited NF-κB activity by around 50% while heating the seeds to 50 • C caused a total inhibition of NF-κB in MC38 colon carcinoma cell line (Agbaria et al., 2015). The major compound of black cumin, thymoquinone, acts on several signal transduction pathways, such as PPAR-γ and STAT3. In fact, thymoquinone induced the expression of PPAR-β/δ (unspecifically) and PPAR-γ (specifically) without significant effect on PPAR-α in MCF7 cells. The increase of PPAR-γ activity was dose-and time-dependent (Woo et al., 2011). Furthermore, thymoquinone inhibited the activation of PI3K/Akt signaling pathway, decreased the phosphorylation of the PTEN and PDK1, the expression of phospho-GSK-3β and Bad, increased cleavage of caspase-9, increased the activity of GSK-3β, Bad and caspase-9 and downregulated the expression of Akt downstream of the mTOR-dependent translational machinery (Rajput et al., 2013). In addition, in a doxorubicin-resistant MCF-7/Dox cell line, thymoquinone caused PARP cleavage with concomitant PTEN upregulation and increase of downstream proteins (Arafa et al., 2011). In p53-defective Jurkat lymphoblastic leukemia cells, the treatment increased both p73 α and β expression, while decreasing the expression of UHRF-1, DNMT1 and HDAC1 in a dose-dependent manner. In HCT116 cells, it decreased the phosphorylation of STAT3, the nuclear localization of p-STAT3 and the expression of STAT3 target genes related to survival (e.g., survivin and cyclin-D1 and -D2), while increasing the expression of cell cycle regulatory proteins (e.g., p27). Furthermore, thymoquinone inactivated kinases responsible for the phosphorylation of STAT3, EGFR tyrosine kinase, JAK2 and Src (Kundu et al., 2014). The compound also modulated the STAT3 pathway in U266 multiple myeloma cells. Thymoquinone inhibited the phosphorylation of constitutively expressed STAT3 and depleted the nuclear translocation of STAT3 without neither FIGURE 7 | Plants from the North Africa and major compounds are able to modulate signal transduction pathways associated with cancer survival. N. sativa extracts (Nv) are able to decrease NF-κB activity. The major compound, thymoquinone (T) decreases STAT3 phosphorylation, nuclear translocation and STAT3-induced gene expression. Furthermore, it inhibits or activates the activity of PTEN, inhibits the phosphorylation of PI3K and promotes the activity of PPAR-γ. Extracts from T. foenum graecum (Tfg) exert an inhibitory effect on the phosphorylation of EGFR induced by TGFβ as well as Akt phosphorylation. In addition, it decreases the activity of p38 MAPK. Diosgenin (D) inhibits NF-κB activity induced by TNF, inhibits IKK (IkBα kinase) activity and decreases phosphorylation and degradation of IkBα. Furthermore, it decreases the phosphorylation and nuclear translocation of p65 (RelA). In addition, it inhibits the phosphorylation of PI3K, Akt, ERK1/2, and JNK1/2. Extracts from M. vulgare (Mv) exert a stimulatory effect on p38 MAPK activity. CR, Co-repressor; CA, Co-activator.
affecting protein levels of STAT3 nor the phosphorylation and protein levels of STAT5 (Li et al., 2010). Furthermore, thymoquinone suppressed IL-6-induced phosphorylation of STAT3, activation of Akt and the expression of STAT3dependent genes. In addition, it inhibited the phosphorylation of constitutively expressed JAK2 and of activated protein tyrosine phosphatases and it induced the expression of SH-PTP2, which negatively regulates STAT3 activation. Finally, thymoquinone downregulated the expression of STAT3dependent genes associated with cell survival, proliferation and angiogenesis.

Trigonella Foenum-Graecum and Diosgenin
A fenugreek extract decreased the phosphorylation of EGFR induced by TGFβ in PC-3 cancer cells (Shabbeer et al., 2009). Furthermore, it inhibited the TGFβ-induced Akt phosphorylation and induced p21 expression. In an animal model of skin carcinogenesis induced by DMBA-TPA, a methanol extract decreased p38MAPK expression (Ali et al., 2014). Diosgenin inhibited TNF-induced expression of NF-κB on KBM-5 cells in a dose-and time-dependent manner (Shishodia and Aggarwal, 2006). Furthermore, it decreased the degradation and phosphorylation of IκBα and the activation of IκBα kinase. In addition, it inhibited the phosphorylation and subsequent p65 nuclear translocation and activation of the Akt pathway. Concomitantly, the saponin inhibited the expression of NF-κB-dependent genes, e.g., antiapoptotic genes, MMP-9 and cyclin-D1. These effects were found both in ER-positive and ER-negative cells, thus suggesting that NF-κB inhibition was ERindependent. Diosgenin inhibited the phosphorylation of PI3K, Akt, ERK1/2, JNK1/2 in a dose-and time-dependent manner (Chen et al., 2011). Furthermore, NF-κB nuclear translocation was inhibited in the presence of diosgenin. Diosgenin treatment activated NF-κB in 1547 osteosarcoma cells (Moalic et al., 2001). Paunovic et al. (2016) described the activation of p38 MAPK with concomitant activation of proapoptotic genes and cell cycle inhibitor and inhibition of antiapoptotic genes after treatment with an ethanolic extract. Furthermore, it conversed LC3-I to the LC3-II in both B16 and U251 cell lines with associated upregulation of autophagy-associated genes.

Invasiveness, Migration, and Metastasis
Cancer cell invasion is one of the major challenges to achieve relapse-free and sustainable patient survival (Alexander and Friedl, 2012). Tumor invasion is characterized by cell motility, migration and degradation of the extracellular matrix (Domoto et al., 2016) with concomitant intravasation into blood and/or lymphatic vessels and metastasis of distant organs (Alexander and Friedl, 2012). Although, only very few migrant cells initiate de novo tumor growth in a new organ (Oh et al., 2015), metastasis is responsible for most of the cancer-related deaths. Invasion is regulated by several factors (Alexander and Friedl, 2012), such as the PTEN/PI3K and MAPK/ERK pathways as well as increased MMP expression and activity. The effect of the extracts and major compounds on the major metalloproteinases associated with cancer invasiveness can be found in Figure 8.

Diosgenin and Trigonelline
To the best of our knowledge, no studies assessing the antiinvasiveness and anti-migration capacities of whole extracts from T. foenum-graecum have been carried out as of yet. Therefore, we only focus on results related to diosgenin. Treatment with diosgenin inhibited TNF-induced invasion in H1299 cells by 50% and ∼40%, respectively, prior and after invasion induction (Shishodia and Aggarwal, 2006). Non-cytotoxic doses decreased the migration of PC-3 cells in a time-dependent manner as well as the cell invasion in a dose-dependent manner (22 and 40% invasion inhibition at 10 and 20 µM, respectively; Chen et al., 2011). Furthermore, the saponin decreased MMP-2 and MMP-9 activity with concomitant suppression of the mRNA and protein expression. In addition, it also decreased the mRNA levels of MMP-7 and EMMPRIN (extracellular inducer of matrix metalloproteinase), while increasing the levels of TIMP-2 (tissue inhibitor of metalloproteinase-2). Tube formation in HUVECs induced by conditioned medium by PC-3 cells was inhibited suggesting that diosgenin might inhibit angiogenesis.
Trigonelline (Figure 4, 4), the major alkaloid of fenugreek has been described to act in an anticarcinogenic manner (El Bairi et al., 2017). In fact, Hirakawa et al. (2005) demonstrated that the ROS-induced migration of AH109A hepatoma cells was inhibited in a dose-dependent manner by this alkaloid. Furthermore, this compound decreased the motility of Hep3B (Liao et al., 2015) as well as the expression of MMP-7, the expression of phospho-Nrf2, PKCα, ERK1/2, and p38, while increasing the phosphorylation of Raf on serine-259. In addition, trigonelline decreased the expression and activity of several anti-oxidative enzymes.

TOXICOLOGICAL EVIDENCE
Toxicity testing represents an important step in the drug development process evaluating the potential of a medicinal plant, before it can be further considered for clinical trials.

Nigella Sativa and Thymoquinone
Black cumin is widely used in traditional medicine. However, it is commonly used in low doses due to the lack of scientific evidence regarding its toxicity (Al- Ali et al., 2008;Dollah et al., 2013). Nevertheless, some authors described toxic effects of N. sativa both in vitro and in vivo. In fact, Khader et al. (2007) described that 0.3 mg/mL of a seed extract from N. sativa exerted genotoxicity since it promoted the formation of MNNG-induced chromosomal aberrations and the appearance of micronuclei. The same authors described that thymoquinone (20 µM) decreased cell proliferation, while necrosis was augmented at concentrations as low as 2.5 µM. Higher concentrations exerted acute cytotoxicity (Khader et al., 2009). Regarding the genotoxicity of thymoquinone, the number of chromosomal aberrations and micronuclei increased in a dose-dependent manner up to 10 µM. In vitro assays are of limited predictive power regarding potential toxicity in human subjects due to the fact that cell cultures do not reflect the biotransformation that occurs in vivo (Reichling et al., 2009). The LD 50 of a fixed oil from the seeds of N. sativa was determined for its acute and chronic toxicity (Zaoui et al., 2002). For acute toxicity, the LD 50 was 28.8 and 2.06 mL/kg body weight of mice for oral and intraperitoneal administration, respectively. The LD 50 of an alcoholic extract administered intraperitoneally was 561 mg/kg (Paarakh, 2010). The acute toxicity of thymoquinone, the major constituent of N. sativa, was determined by Badary et al. (1998). The LD 50 was 2.4 g/kg. High doses of thymoquinone induced hypoactivity and breathing difficulties before death. After 24 h of treatment, high doses of thymoquinone decreased liver, plasma and heart GSH levels. Furthermore, the activities of ALP, LDH and CPK, and plasma concentration of urea and creatinine were elevated. The chronic toxicity was determined in rats by giving 2 mL/body weigh/day for 12 weeks (Zaoui et al., 2002). At the end of the treatment, no differences in key enzymes were detected, serum cholesterol, triglycerides, and glucose significantly decreased. Furthermore, the counts for leukocytes and platelets decreased while hematocrit and hemoglobin increased. The treated groups presented a lower body weight at 6 weeks and onwards, whereas the organ weight remained unchanged. The subchronic administration of 30-90 mg/kg/day thymoquinone for 90 days showed no effect on key enzymes, body weight, hematological parameters nor did it cause histopathological changes (Badary et al., 1998). The oral administration of aqueous, methanol and chloroform extracts exerted no mortality 7 days post-treatment. However, at 21 g/kg of aqueous and chloroform extracts lowered the body weight (Vahdati-Mashhadian et al., 2005). The subacute effects of these extracts were determined by giving orally 2 g/kg/day for 14 days. At the end of the treatment, the aqueous extract decreased alkaline phosphatase (ALP), if compared to the control. Other enzymes were not affected. The chloroform extract decreased the activity of all tested enzymes. The oral intake of either fixed oil or essential oil was considered as safe, since no changes on serological indexes and white and red cell count appeared even 56 days after treatment (Tauseef Sultan et al., 2009). Al- Ali et al. (2008) demonstrated that the toxicity of N. sativa depends on the type of administration. Intraperitoneal injections to mice and rats demonstrated low LD 50 values (104.7 and 57.5 mg/kg for mice and rats, respectively), whereas upon oral administration the lethal doses were 870.9 and 794.3 mg/kg, respectively. Furthermore, the intraperitoneal administration led to abdominal muscle contraction and ataxia, which persisted for a couple of hours. After 6 h, the animals were drowsy and less responsive. For the oral administration, the drowsiness and poor responsiveness was more gradual and lasted until the animal's death or vanished after 24 h. The oral administration of grounded seeds to rats in doses of 0.01, 0.10, and 1.00 g/kg body weight for 5 weeks caused no changes in urea and creatinine levels, neither did it cause histopathological changes in the kidney (Dollah et al., 2013). In addition to both in vitro and in vivo studies, several clinical trials have been conducted using either N. sativa and thymoquinone without demonstrating any severe side effects (Qidwai et al., 2009;Akhondian et al., 2011;Amin et al., 2015;Mohtashami et al., 2015). In addition to the rather safe toxicological profile, i.p. pre-treatment with 12.5 mg/kg. of N. sativa volatile oil decreased the hepatotoxicity induced by CCl 4 by decreasing the level of serum enzymes and MDA content (Mansour et al., 2001). On the other hand, thymoquinone decreased cisplatin-induced nephrotoxicity by lowering urea and creatinine levels and improvement of polyuria, kidney weight, and creatinine clearance (Badary et al., 1997).

Trigonella Foenum-Graecum and Diosgenin
Fenugreek has been widely used in traditional medicine for treating several ailments and diseases, although several sideeffects have been reported (Al-Ashban et al., 2010), including anti-fertility and abortifacient effects (Ouzir et al., 2016). The lethal dose of an alcoholic extract of fenugreek seed was above 1 g/kg i.p. (Sur et al., 2001). The intraperitoneal administration of an aqueous extract yielded an LD 50 close to 4,000 mg/kg (Javan et al., 1997). The LD 50 upon i.p. administration of an aqueous extract from the leaves of T. foenum-graecum was 1.9 g/kg, while the LD 50 increased to 10 g/kg upon oral administration (Abdel-Barry et al., 1997). The i.p. administration of a glycoside extract from fenugreek revealed an LD 50 of 0.65 g/kg, while the LD 50 value increased to 7 g/kg upon oral administration (Abdel-Barry and Al-Hakiem, 2000). A fenugreek extract enriched in furostanolic saponins (>60% w/w) did neither show acute toxicity (LD 50 > 5,000 mg/kg) nor sub-chronic toxicity (Swaroop et al., 2014). Supplementation with 100 mg/kg/day in drinking water for 3 months did not show alarming signs of chronic or acute toxicity (Al-Ashban et al., 2010). Similarly, intragastrical administration of fenugreek extract did not exert acute toxicity upon oral supplementation (Muralidhara et al., 1999). Although, several studies have demonstrated low toxicity of fenugreek extracts, some authors have established that these extracts are toxic to the fetus development (Khalki et al., 2010;Mozaffari et al., 2010). Furthermore, Khader et al. (2007) also described the mutagenic potential of T. foenum-graecum. Several clinical trials have been conducted using fenugreek, which described mostly non-serious side effects. In fact, the supplementation of diabetic individuals with 25 g/day for 24 weeks revealed a few side effects (diarrhea and excess flatulence) that subdued after a few days (Sharma et al., 1996). In a single-blind trial, 20 healthy men treated with an aqueous extract minimal side effects were observed (feeling of hunger, dizziness and frequency of micturition; Abdel- Barry et al., 1997). A double-blind, randomized and placebo-controlled 6-week trial in 60 healthy adult males showed no adverse side effects using a standardized fenugreek extract, except for three individuals that complained about slight stomach discomfort after taking the extract in the absence of food (Steels et al., 2011). In another double-blind, randomized, placebo-controlled three-period cross-over trial, 25 healthy male individuals were supplemented with fenugreek for three 14-day treatment period with a 14-day washout period, only few side effects were described (abdominal pain and urine smell; Chevassus et al., 2009). In a double-blind and placebocontrolled, Gupta et al. (2001) described that treatment with 1 mg/day of a hydroalcoholic seed extract for 2 months caused dyspepsia and mild abdominal distension without affecting the kidney and liver. A double-blind, randomized and placebocontrolled trial on 120 healthy adult male using a standardized seed extract showed no adverse effects during 12 weeks of treatment (Rao et al., 2016). Some minor side effects were observed that were dispersed between both treatment and control groups. Thus, they cannot be associated with fenugreek consumption. Swaroop et al. (2015) conducted an open label, one-arm, non-randomized, and post-marketing surveillance in 50 premenopausal women and did not find hepatotoxicity, nephrotoxicity and cardiotoxicity during the 90 days observation period. Furthermore, changes in the hemogram and total leukocyte count were not observed, although the hemoglobin levels slightly increased. A short-term, single site, double-blind, randomized and placebo-controlled study was carried out in 80 women, who consumed 600 mg/day of a standardized seed extract (Rao et al., 2015). No major adverse effects were observed with only two patients manifesting minor side effects in the treatment group.

Aristolochia Longa
Although beneficial effects of A. longa were reported not only against cancer, but also against several other ailments, this plant is rich in aristolochic acids (Cherif et al., 2014). These acids are responsible for the toxicity associated with many species of the Aristolochia genus (Yamani et al., 2015) and cause a syndrome named aristolochic acid nephropathy (Benzakour et al., 2012;Cherif et al., 2014). By contrast, several studies have assessed that extracts from the roots of A. longa to be safe in a single dose treatment (Benzakour et al., 2011;Benarba et al., 2012Benarba et al., , 2016Cherif et al., 2014). On the other hand, Cherif et al. (2014) described 1/6 mortality upon application of 4 g/kg and 2/6 mortality upon 5 g/kg. Low doses failed to exhibit sings of sub-chronic toxicity (Benzakour et al., 2011;Cherif et al., 2014). By contrast, high doses demonstrated serious toxicity, such as atypical locomotion, anorexia, asthenia, ataxia, diarrhea, and urination (Benzakour et al., 2011;Cherif et al., 2014).

SCIENTIFIC EVIDENCE ON MEDICINAL PLANTS NOT TRADITIONALLY USED FOR CANCER TREATMENT
Several plants are widely used as by herbalists in North Africa to treat diseases other than cancer. Nevertheless, several studies have been conducted on some of those plants highlighting their potential as anticancer agents. Therefore, we also report on these medicinal plants. Diethyl ether and petroleum ether extracts from Alhagi maurorum demonstrated a comparable cytotoxicity toward several cancer lines, including C32 cells as the most susceptible to the first extract (IC 50 = 2.7 µg/mL) and HeLa cells as the most susceptible to the latter (IC 50 = 40.1 µg/mL; Loizzo et al., 2014). Methanolic extracts from Artemisia herbaalba, Ruta chalepensis, and Peganum harmala all demonstrated cytotoxic effects in a dose-dependent manner (Khlifi et al., 2013). An ethyl acetate extract of Argania spinosa inhibited the growth of MCF-7 cells (IC 50 = 42 µg/mL; El Babili et al., 2010). Extracts of Inula viscosa were cytotoxic toward several cancer cell lines, including multidrug-resistant lines, with the hexane extract demonstrating the most promising activity (IC 50 = 9.56-15.67 µg/mL; Merghoub et al., 2016). In a comparative study with 16 plants, the methanol extracts from I. viscosa and Ononis hirta were the most potent ones against all tested cancer cell lines. For instance, in MCF7 cells the IC 50 values were in a range from 15.78 to 27.96 µg/mL (Talib and Mahasneh, 2010). A crude alkaloid extract from Glaucium flavum inhibited the growth of several human cancer cell lines in a dose dependent manner (IC 50 = 7.9-13.6 µg/mL; Bournine et al., 2013a). Furthermore, this type of extract significantly inhibited the growth of MDA-MB-231 cells in a dose-dependent manner (Bournine et al., 2013b). Treating mice bearing Ehrlich ascites tumor with an methanolic extract of Calligonum comosum decreased the number of viable cells (99.2 × 10 6 vs. 192.8 × 10 6 ; Badria et al., 2007). In another study, a methanolic extract of C. comosum inhibited the viability of HepG2 carcinoma cells in a dose-dependent manner (IC 50 = 9.60 µg/mL; Shalabi et al., 2015). Similarly, Aloe vera also decreased cell viability of HepG2 in a dose-dependent manner (IC 50 = 10.45 µg/mL). The aqueous extract obtained from garlic (Allium sativum) decreased the viability of murine CT26. WT colon carcinoma cells (Lee et al., 2013). Several extracts from mulberry (Morus alba) decreased the viability of HepG2 in a time-dependent manner (Fathy et al., 2013). The essential oil of Artemisia campestris inhibited the growth of HT-29 cells (IC 50 = 46.82 µg/mL; Akrout et al., 2011). Several extracts of Thymelaea hirsuta only weakly inhibited cell growth at high concentrations of 100 µg/mL (58.19-65.54% inhibition). Several roots extracts from Linum usitatissimum significantly inhibited the growth of Jeg3 trophoblast tumor cells (Abarzua et al., 2007). Infusions from Mentha spp., Rosmarinus officinalis and Origanum majorana inhibited the growth of several cancer cell lines, including HeLa, Jurkat and MCF-7 . Seed oil isolated from Ecbalium elaterium inhibited the growth of HT-29 colonic adenocarcinoma (IC 50 = 4.86 µg/mL) and HT-1080 colonic fibrosarcoma cells (IC 50 = 4.16 µg/mL; Touihri et al., 2015). Treating MCF-7 and OVCAR cells with 50 mg/L of hydromethanolic extract from the flower buds of Cistus salviifolius inhibited the growth by 35.16 and 36.85%, respectively (El Euch et al., 2015). The methanolic extract of Achillea odorata (50 µg/mL) inhibited the MCF7 cell growth of MCF-7, Hep2 and WEHI cancer cells in a dosedependent manner with 42.90, 61.54, and 81.13% inhibition, respectively (Boutennoun et al., 2017). The growth of B16F10 was inhibited in a dose-and time-dependent manner (IC 50 ∼736 µg/mL after 24 h and ∼650 µg/mL after 48 h) after treatment with an aqueous extract from the gall of Limoniastrum guyonianum (Krifa et al., 2014). Furthermore, it also decreased both tumor weight and size in B16F10-bearing mice. In a similar study, the same type of extract exerted antiproliferative activity toward HeLa cells in a dose-and time-dependent manner (IC 50 = 170 and 140 µg/mL after 24 and 48 h, respectively; Krifa et al., 2013). A hexane extract from the bark of Ficus drupacea decreased the growth of several cancer lines with T24 cells as the most susceptible ones (IC 50 = 21.32 µg/mL; Yessoufou et al., 2015). The anticancer effects of a methanolic extract of Cleome arabica have been described against several cell lines. SK-N-BE neuroblastoma cells were the most sensitives ones to this extract (Tigrine et al., 2013). Saffron (Crocus sativus) exerted antiproliferative activity against two aggressive prostate cancer cell lines injected into mice by decreasing tumor weights (Festuccia et al., 2014). Furthermore, it decreased the metastatic capabilities of the tumors. In another study, saffron extracts inhibited the growth of both p53 +/+ and p53 −/− HCT116 cell lines in a dose-and time-dependent manner (Bajbouj et al., 2012).
In a MNNG-induced gastric carcinogenesis model, treatment with an aqueous extract from C. sativus inhibited the progression to the adenoma stage in a dose-dependent manner (Bathaie et al., 2013). Saffron also inhibited the growth of A549 lung cancer cells in a dose-and time-dependent manner (IC 50 = 380 and 170 µg/mL after 48 and 72 h; Samarghandian et al., 2013). Dichloromethane and methanol extracts from Limonium densiflorum were cytotoxic against different cancer cell lines being the dichloromethane extract the most promising one (Medini et al., 2015). The ethyl acetate extract from Cyperus rotundus weakly inhibited the proliferation of K562 erythroleukemia cells (IC 50 = 100 µg/mL; Kilani-Jaziri et al., 2009). This type of extract only weakly inhibited the growth of L1210 cell line (IC 50 = 200 µg/mL), while a lyophilized infusion and a methanol extract exerted negligible activity (Kilani et al., 2008). Several extracts from the leaves of Punica granatum were screened for their antiproliferative effects toward MCF-7 cells (Bekir et al., 2013). The methanol extract was the most effective (IC 50 = 31 mg/L). A hydroethanolic extract from the grains of Echinochloa crus-galli demonstrated potent cytotoxicity against several cancer cell lines (El Molla et al., 2016). Extracts from Ceratonia siliqua and Quercus ilex inhibited the proliferation of U87 glioblastoma in a dose-dependent manner (Amessis-Ouchemoukh et al., 2017). The growth of KB cancer cell line was inhibited in a dose-dependent manner by an aqueous extract from Moringa oleifera leaves (Sreelatha et al., 2011). Several extraction techniques from the leaves of various olive cultivars inhibited the growth of the JIMT-1 breast cancer cell line in a dose-dependent manner (Taamalli et al., 2012). An aqueous extract from Jatropha podagrica strongly inhibited the growth of both PC12 and A549 in a dose-dependent manner (Ghali et al., 2013). Henna (Lawsonia inermis) at doses of 180 mg/kg for 15 days significantly reduced the growth of Dalton's lymphoma ascites-bearing mice (Priya et al., 2011). Treating Ehrlich ascites carcinoma-bearing mice with henna extract decreased the number of cancer cells (Ozaslan et al., 2009). Similarly, the chloroform extract of L. inermis was very cytotoxic toward HepG2 (IC 50 = 0.3 µg/mL) and MCF-7 (IC 50 = 24.8 µg/mL; Endrini et al., 2002Endrini et al., , 2007. Ethyl acetate and petroleum ether extracts both decreased viability of MCF7 cells (IC 50 = 27 and 22 mg/mL, respectively; El Babili, 2013). The essential oil of henna also showed potent cytotoxicity toward HepG2 (IC 50 = 24 µg/mL; Rahmat et al., 2006). An aqueous extract of Urginea maritima inhibited the growth of SH-SY5Y in a dose-and time-dependent manner (IC 50 = 10 µg/mL, 1 µg/mL and 100 ng/mL after 24, 48, and 72 h; Elghuol et al., 2016). Both aqueous and ethanol extracts from leaves, seeds and roots of Plectranthus amboinicus similarly inhibited the growth of two cancer cell lines, HepG2 and MCF-7 (El-hawary et al., 2012). Hexane, ethyl acetate and chloroform extracts from Aristolochia baetica significantly inhibited the growth of MCF-7 cells in a doseand time-dependent manner (Chaouki et al., 2010). Similarly, ethyl acetate and methanol extracts of Origanum compactum both inhibited the proliferation in dose-and time-dependent manner.

CONCLUSION
The extensive research on the anticancer activity of plants from the North Africa highlighted their potential for future chemotherapeutic agents, especially plants from Morocco and Algeria, which have the most ethnobotanical references regarding the anticancer effects of plants. Amongst all the botanical families ascribed, Lamiaceae, Apiaceae, Compositae, and Fabaceae are the ones with more plants used in folk medicine. In fact, the most used plants, T. foenum-graecum, A. longa, M. vulgare, and C. absus belongs to these families. On the other hand, N. sativa (Ranunculaceae) is also extensively used as herbal anticancer drug in North African traditional medicine. The anticancer effects of N. sativa and T. foenum-graecum are widely described both in cytotoxic assays and in mechanisms underlying those effects. Although A. longa and M. vulgare are referred by herbalists for the treatment of cancer, very few studies assessing their anticancer properties have been conducted while studies demonstrating the anticancer effects of C. absus are yet to be conducted. The anticancer effect of these plants act on a panoply of cellular pathways being the induction of apoptosis and modulation of signal transduction pathways the most studied, while the ability to suppress invasiveness, migration and metastasis is poorly evaluated. Although, the safety profile of the plants has been described with minor side effects, T. foenumgraecum have been associated with anti-fertility and abortifacient effects and A. longa have been described as inducing severe side effects after chronic consumption of the plant. Despite the myriad of studies regarding the anticancer effect as well as extensive literature on the underlying mechanisms, very few in vivo studies have been conducted and clinical trials are still to be conducted.
In order to establish these plants as potential chemotherapeutic agents several points must be addressed namely the assessment of anticancer potential using in vivo models. Furthermore, clinical trials are still lacking, which limits the therapeutic application of these plants. In addition, a better understanding on the mechanisms of action is still needed in order to establish rational phytotherapeutic approaches.