Hunting (Edible) Starfishes for PAK1-blocking Saponins

Recently several anti-cancer saponins were isolated from marine organisms such as “edible” sea cucumbers and starfishes. One of these marine saponins is Frondoside A from sea cucumber (Cucumaria frondosa) which suppresses the growth of pancreatic cancer in vivo (xenograft in mice). Since the growth of pancreatic cancers depends on PAK1, and this saponin induces p21, a CDK-inhibiting protein, whose gene is activated by PAK1-blockers, it is most likely that Frondoside A (FRA) is a PAK1-blocker. Recently we confirmed that FRA indeed inhibits PAK1 directly.

More recently a Korean group found another anti-cancer saponin in methanol extract from an edible starfish grown in Vietnam coast line (1).  The IC₅₀ of this extract against cancer cells is around 1 ppm (1 micro g/ml). In other words, its anti-PAK index is around 100, several times more potent than Bio 30, a CAPE-based propolis. Furthermore, it inactivates the oncogenic kinase ERK, just downstream of PAK1. Thus, we recently started hunting starfishes along Okinawa coast line, screening for a similarly (or even more) potent anti-cancer (PAK1-blocking) saponin(s).

References:

1. Thao NP, Luyen BT, Kim EJ, Kang HK, Kim S, Cuong NX, Nam NH, Kiem PV, Minh CV, Kim YH. 　Asterosaponins from the Starfish Astropecten monacanthus suppress growth and induce apoptosis in HL-60, PC-3, and SNU-C5 human cancer cell lines. Biol Pharm Bull. 2014;　37(2):　315-21.
 

"Anti-PAK Index" : the International standard of PAK1-blocking herbal products such as propolis for quality control

There is a wide variety of edible PAK1-blocking herbal products such as propolis available on the market world-wide which could be very useful for improving our health and even therapy of cancer and many other PAK1-dependent diseases/disorders such as Alzheimer’s disease (1). However, unlike FDA-approved drugs, none of them is associated with any reliable international quality control standard such as IC₅₀ or ED (effective dose). For instance, the quality of propolis depends on the sources of plants where bees harvest from, and the actual content of PAK1-blocking ingredients such CAPE, apigenin, ARC (artepillin C) and propolin G (=nymphaeol C) in each propolis. However, since 1960s till present, the only available quality stardard used for propolis is just “the total flavonoid content” for CAPE-based propolis or the ARC content in Brazillian green propolis. If these “herbal” health-promoting products are regulated by a single reliable pharmacological quality standard, we could compare the quality or effectiveness from one sample to another quite objectively, regardless of their detailed content.

Hence, we would propose here for the fist time to use a universal standard called “Anti-PAK Index” which is the 100 X reciprocal of the IC₅₀ in ppm (micro g/ml). For instance, “Anti-PAK Index” of Bio 30 (CAPE-based propolis from New Zealand) , (propolin G-based) Okinawa propolis (OP) and (ARC-based ) Brazilian green propolis (GP) are 12.5, 8 and 1, respectively, since their IC₅₀ for A549 cancer cells are around 8, 12 and 100 ppm, respectively, whereas the “Anti-PAK Index” of the “pure” compound "Cucurbitacin" (CB) is around 1400, as the IC₅₀ for A549 is around 140 nM (0.07 ppm). The higher the Anti-PAK Index, the more potent a given sample. In other words, 1 mg of CB is equivalent to 112 mg of Bio 30 for therapy of cancers and many other PAK1-dependent diseases/disorders. This rough estimation is not far from the actual  in vivo data where the daily dose of  CB (1 mg/kg),  and that of Bio 30 (50 mg/kg) are their effective dose to suppress the PAK1-dependent growth of pancreatic cancers or NF tumors in mice (1), suggesting that their in vivo bioavailability is quite similar.

The only difference between these two is that Bio 30 has been available on the market world-wide for clinical uses for almost a decade, but CB is not as yet.  The only way for us to take CB is to eat the edible bitter melon (Goya) grown in Okinawa (which contains around 1 g of CB per kg) or drink Goya teas. Since roughly 90% of Goya is water, the “Anti-PAK Index”of Goya extract/tea could be around 14, pretty close to that of Bio 30.

We recently managed to synthesize a highly cell-permeable and water-soluble compound whose IC50 is around 24 nM (0.01 ppm) against the growth of A549 cancer cells (2). Its "Anti-PAK Index" is 10,000, several times more potent than CB, and we have filed a US patent on this new PAK1- blocker called "15K" for its clinical application (The related Japanese patent was granted). It is derived from an old FDA-approved synthetic anti-inflammatory pain killer (Ketorolac).  It passes the BBB (blood brain barrier), and could be useful for treating a variety of PAK1-dependent neuronal diseases/disorders such as brain tumors, AD (Alzheimer's disease), PD (Parkinson's disease), epilepsy, depression,  schizophrenia and even autism. Furthermore, it extends the healthy lifespan of C. elegans by 30% at 50 nM, clearly indicating that "15K"  causes no side effect!

I bet that leading pharmaceutical giants such as Pfeizer, Roche and Novartis, who are currently racing for developing potent PAK1-blockers which would be useful for clinical application, would be most likely keen to buy an exclusive license (worth more than 200 million USD) from this patent of ours. 

References:

１．Maruta H. Herbal therapeutics that block the oncogenic kinase PAK1: A practical approach towards PAK1-dependent diseases and longevity. Phytother Res. 2014; 28:656-672..

2.    Nguyen BC, Takahashi H, Uto Y, Shahinozzaman MD, Tawata S, Maruta H. Chemistry"-based highly potent PAK1-blocking cancer-killer. Eur J Med Chem. 2016 ; 126: 270-6.

 

Lesson 5: Triazolylation of "Acidic" Herbal PAK1-blockers for Improving Their Cell-permeability

There are a variety of  "acidic" herbal compounds that block the oncogenic/ageing kinase PAK1.
They include CA (caffeic acid) and ARC (artepillin C) from propolis, UA (ursolic acid) and RA (rosmarinic acid) from Rosemary leaves, and so on.  The major problem for their clinical application is their poor cell-permeability. Their COOH keeps these molecules "acidic", and this acidicity (negative charge) hampers their efficient penetration through the "acidic" (phospholipid-based) cell membranes.

In the case of CA,  phenethyl esterization of its COOH (converting to CAPE) significantly improves its cell-permeability.  That is why CAPE is 10 times more potent than CA as a PAK1-blocker in cell culture. However, this type of esterization significantly reduces its water-solubility, and therefore its bioavailability (absortion through intestine) in vivo.

Recently, owing to the potentially Nobel-winning "Click Chemistry" (invented by the 2001 Nobel laureate, Barry Sharpless) , which allows us to couple the "water-soluble" triazole ring to OH of the COOH moiety of these acidic herbal PAK1 blockers in a very high yield with copper salt as a catalyzer, an Indian group managed to improve the cell-permeability of a triterpene called UA by 200 folds (lowering the IC50 below 100 nM in cell culture) (1). 

Furthermore, a Chinese group managed to replace OH at position 3 of UA by amine, adding a "positive" charge (for improving its cell-permeability), and enhaunced its anti-cancer activity by 20 folds (2).  In theory, if we could combine these two "specifics", UA's IC50 in cell culture could be lowered to around 5 nM.  I believe a similar modification could be applied to many other acidic herbal PAK1-blockers for a robust improvement of both their cell-permeability and bioavailability.

In fact we recently managed to boost the cell-permeability (=anti-cancer activity) of an acidic PAK1-blocker (Ketorolac) over 500 times by the "Click Chemistry" (3, and US patent).

References:

1.      Rashid S, Dar BA, Majeed R, Hamid A, Bhat BA. See comment in PubMed Commons belowSynthesis and biological evaluation of ursolic acid-triazolyl derivatives as potential anti-cancer agents. Eur J Med Chem. 2013 ; 66:238-45. 

2.   Ma CM, Cai SQ, Cui JR, Wang RQ, Tu PF, Hattori M, Daneshtalab M. The cytotoxic activity of ursolic acid derivatives. Eur J Med Chem. 2005; 40 (6): 582-9.

3.   Nguyen BC¹, Takahashi H², Uto Y², Shahinozzaman MD¹, Tawata S³, Maruta H⁴. 

      "Click Chemistry"-based highly potent PAK1-blocking cancer-killer. Eur J Med Chem. 2016 ;126:270-276.

  

Lesson 4: PAK1-blockers=AMPK-activators=HSP-inducers

Among Tyr-kinases essential for the activator of PAK1, FYN plays a unique role. It inactivates the anti-oncogenic kinase LKB1 which in turn inactivates PAK1, and activates the anti-oncogenic kinase AMPK, simultaneously.  Mainly due to this kinase cascade (and a few other signal cascades) , a unique "formula for longevity" has been established at least among herbal compounds, without any exception:  PAK1-blockers=AMPK-activators (1, 2).

The longevity transcription factor "FOXO" is inactivated by PAK1, and activated by AMPK. The FOXO is essential for the expression/induction of a variety of HSP (heat shock protein) genes which protect a variety of intracellular proteins from heat-denaturation or many other stresses. Thus, the higher the HSP level, the longer the healthy lifespan.

In other words, PAK1-blockers=AMPK-activators=HSP-inducers.  Thus, using a unique strain called CL2070 of C. elegans which carries a transgenic gene called HSP16-GFP, in a few days we can screen in vivo for a variety of HSP-inducers, which are PAK1-blockers and AMPK-activators as well (3).

References:

1．Maruta H.（2011）Effective neurofibromatosis therapeutics blocking the oncogenic kinase PAK1. Drug Discov. Ther. 5, 266-78.

2．Maruta H. （2014） Herbal Therapeutics that Block the Oncogenic Kinase PAK1: A Practical Approach towards PAK1-dependent Diseases and Longevity. Phytother Res. 28、656-72．

3．Yanase S, Luo Y, Maruta H.（2013）PAK1-deficiency/down-regulation Reduces Brood Size, Activates HSP16.2 Gene and Extends Lifespan in C. elegans. Drug Dev Ther. 7: 29-35.

Lesson 3: Tackle Tyr-kinases (ETK, JAK2, FYN), only upstream of PAK1

I have pointed out that three Tyr-kinases (ETK, JAK2 and FYN) are essential for the full activation of PAK1 in cells. Interestingly, none of these Tyr-kinases is involved in the activation of other members of PAK family (PAK 2-6).  In other words, inhibiting ETK, JAK2 or FYN would block selectively the activation of PAK1 only, and would not affect the remaining five members of PAK family. Furthermore, an ETK inhibitor called AG879 or GL-2003, a JAK2 inhibitor called cucurbitacin I, and a FYN-inhibitor called PP1/PP2,  block the activation of PAK1 in cells, mostly with the IC50 ranging 5-10 nM, clearly indicating that these Tyr-kinase inhibitor are highly potent (cell-permeable), and therefore would be very useful for clinical purposes such as cancer therapy. 

Thus, instead of developing (poorly cell-permeable) "PAK1-specific" inhibitors, we shall encourage "bright" (open-minded) scientists to tackle "PAK1-specific" activators (ETK, JAK2 and FYN) by these (or new) "highly cell-permeable" compounds (synthetic or natural).  You should push the boundary (Grenzgebiet) further for a great leap.  Why not?

We have recently patented a series of water-soluble PP1/PP2 derivatives which are potentially useful for clinical application. One of them called PP12 inactivates PAK1 by inhibiting FYN directly, and inhibits the growth of human colon and lung cancer cells, which carry the oncogenic K-RAS mutant, with the IC50 around 50 nM.  We believe PP12 would be potentially useful for therapy of pancreatic, colon, and lung cancers as well as brain tumors such as glioma　and NF tumors. 

Lesson 2: Start with a "highly cell-permeable" ATP antagonist to develop "potent" PAK1-blockers

Back in 2001, we identified the first potent PAK1-inhibitor among "marine" derivatives of staurosporine (ST) called "ST-2001" (or 3-OH ST) in which the position 3 of ST is hydroxylated. Its IC50 for both PAK1 in test tube and anti-cancer activity in cell culture is around 1 nM, indicating that its "cell-permeability" is very high. Unfortunately, however, it turned out to be a "non-specific" kinase inhibitor and "toxic" for normal cells as well.

The best (and probably simplest) way to make this marine compound "PAK1-specific" would be to link "bulky" and "positively charged" side chains (such as hexylamine) to the position 9 by an enzymatic synthesis of ST-3009. This modification would unable many kinases such as PKC (essential for the growth of normal cells) to accomodate this bulky compound in their relatively small ATP-binding pocket, except for PAK1 (non-essential for the growth of normal cells). PAK1 forms an inactive homodimer whose ATP-binding pocket is larger than any other kinases.

The major set-back of this project is that the marine source of this compound suddenly vanished from the coast line of Guam Island, perhaps due to a global warming (a rise in sea temperature) . On the other hand, chemical hydroxylation of ST only at position 3 (or 9) is possible, but "economically unfeasable" (the yield is very low!).  Thus, as an alternative, we search for an enzyme (called ST hydroxylase, STOHase) which is able to hydroxylate ST at either position 3 or 9 (but not both positions!).

Currently, we are screening for extracts from marine organisms such as sea cucumber which contain PAK1-blockers and would be able to convert ST to ST-2001 in vitro. Once ST-2001 is produced in test tube, its chemical modification at position 9 would be a piece of cake. In the past we have learned lots of lesson from CEP-1347 (1).  Recently we learned another key lesson from the enzymatic synthesis of ARC (artepillin C) using a gio-specific yeast lipase B (Novozym 435) which could distinguish the ester bond at position 3 from that at 9 of ST (2).

References:

1. Nheu, T., He, H., Hirokawa, Y., Tamaki, K. et al. The K252a derivatives, inhibitors for the PAK/MLK kinase family selectively block the growth of RAS transformants. Cancer J. 2002, 8, 328-36.

２．Yashiro K, Hanaya K, Shoji M, Sugai T. New synthesis of artepillin C, a prenylated phenol, utilizing lipase-catalyzed regioselective deacetylation as the key step. Biosci Biotechnol Biochem. 2015、 18:　1-5.

Lesson 1 (Golden Rule): Don’t screen for direct PAK1-inhibitors in test tube!

This year a "Genentech" team developed a new series of PAK1-3 (so-called "group 1") inhibitors:

J Med Chem. 2015 Jun 1. [Epub ahead of print]

Structure-Guided Design of Group I Selective p21-Activated Kinase (PAK) Inhibitors.

Crawford JJ, Lee W, Aliagas I, Mathieu S, Hoeflich K, Zhou W, Wang W, Rouge L, Murray L, La H, Liu N, Fan PW, Cheong J, Heise C, Ramaswamy S, Mintzer R, Liu Y, Chao Q, Rudolph J.

Abstract

Following a high-throughput screen, we identified an aminopyrazole scaffold-based series that was optimized to yield group I selective PAK inhibitors. A structure-based design effort aimed at targeting the ribose pocket for both potency and selectivity led to much-improved group I vs. II selectivity. Early lead compounds contained a basic primary amine, which was found to be a major metabolic soft spot with in vivo clearance proceeding predominantly via N-acetylation. We succeeded in identifying replacements with improved metabolic stability, leading to compounds with lower in vivo rodent clearance and excellent group I PAK selectivity.

The IC50 (cell permeability) of a new PAK1-3 inhibitor (called compound 23) in cell cuture is around 120 nM, similar to a herbal PAK1-blocker called "cucurbitacin I" from an edible bitter melon grown in Okinawa. We shall see its "bioavailability" and "safety" in vivo (animal test).

The direct PAK1-inhibitors called FRAX486 and FRAX597 developed by AFRAXIS (founded by Susumu Tonegawa of MIT, the 1987 Nobel laureate) have the IC50 around 10 nM for inhibiting PAK1 in test tube. However, their IC50 for inhibiting the PAK1-dependent growth of tumor cells in cell culture is above 1 micro M (1000 nM), clearly indicating that they are "very poorly cell-permeable", and therefore not useful for clinical application. In fact,  a recent in vivo study confirmed that FRAX597 (3 mg/kg, i.p.) alone has no effect on the growth of human pancreatic cancer xenograft in mice, while a herbal PAK1-blocker called glaucarubinone (1-2 mg/kg, i.p.) inhibits the growth of the same cancer xenograft by 70%.   

Again, a Novartis team recently developed a new PAK1-specific compound 3, which inhibits PAK1 with the IC50 around 10 nM in test tube, but in cell culture its IC50 is around 2 micro M (2000 nM), clearly indicating that compound 3 is no good for clinical application. So have you got any lesson? 

Combination of immunoprecipitation (IP)-ATP_Glo kinase assay and melanogenesis for the assessment of potent and safe PAK1-blockers in cell culture.

Drug Discov Ther. 2015; 9(4): 289-95.

Nguyen BC, Be Tu PT, Tawata S, Maruta H  (PAK Research Center, Okinawa, Japan)

Abstract

Cucurbitacin I (CBI) is a triterpene from a bitter melon called Goya grown in Okinawa, Japan, and directly inhibits both the Tyr-kinase JAK2 and the G protein RAC, leading to the inactivation of PAK1 (RAC/CDC42-activated kinase 1). Bio 30, a propolis produced in New Zealand, contains CAPE (caffeic acid phenethyl ester) as the major anti-cancer ingredient which directly down-regulates RAC, leading to the inactivation of PAK1. Since PAK1 is essential for the growth of RAS cancer cells such as A549 cell line which carry an oncogenic K-RAS mutant, and the melanogenesis in skin cells, here using these PAK1-blockers as model compounds, we introduce a new approach to the quick assessment of PAK1-blockers in cell culture. 

First, combining the immuno-precipitation (IP) of PAK1 from cell lysate and the in vitro ATP_Glo kinase assay kit (called "Macaroni-Western" assay), we confirmed that both CBI and Bio 30 inactivate PAK1 in A549 lung cancer cells in 24 h, and inhibit their PAK1-dependent growth in 72 h. Furthermore, we verified that CBI inhibits the PAK1/PAK4-dependent melanogenesis in melanoma cells by far more than 50%, while Bio 30 inhibits the melanogenesis only by 50%, with only a merginal effect on their growth per se. Since the "Macaroni-Western" kinase assay and melanogenesis are both rather simple and quick, the combination of these two cell culture assays would be highly useful for selecting both "potent" (highly cell-permeable) and "safe" (non-toxic) natural or synthetic PAK1-blockers.

Oncogenic/Ageing Pathways Leading to the Activation of PAK1

Within cells, there are several distinct proteins essential for the full activation of the oncogenic/ ageing kinase PAK1. First of all, two distinct G proteins called RAC and CDC42 bind to the GBD domain of this kinase. This interaction leads to opening-up of this molecule.  Cucurbitacin, a triterpene from a bitter melon (Goya) in Okinawa, and CAPE (caffeic acid phenethyl ester) from propolis inactivates RAC, leading to the inactivation of PAK1.

Secondly, two SH3 adaptor proteins called PIX and NCR bind to the Pro-rich domains called "PAK18" and "PAK13" in the N-terminal half of this kinase. The PIX-PAK1 interaction requires phosphorylation of PAK1 at Tyr 285 by an oncogenic Tyr-kinase called JAK2. Cucurbitacin and Ursolic Acid (UA) from rosemary leaves directly inhibits JAK2 as well, thereby blocking the PIX-PAK1 interaction, leading to the inactivation of PAK1. In addition, the cell-permeable peptide "WR-PAK18" could block the PIX-PAK1 interaction, leading to the inactivation of PAK1.

Thirdly, its interaction with another Tyr-kinase called ETK is also required for the activation of PAK1. Thus, a synthetic compound called AG 879 that blocks the ETK-PAK1 interaction also inactivates PAK1. A third Tyr-kinase called FYN is also involved in the activation of PAK1. The anti-oncogenic kinase LKB1 is known to directly phosphorylate PAK1 at Thr 109 to inactivate PAK1, while phosphorylate another anti-oncogenic kinase called AMPK at Thr 172 to activate AMPK. FYN phosphorylates LKB1 for inactivation, leading to the activation of PAK1, and inactivation of AMPK. Thus, a synthetic FYN inhibitor called PP1/PP2 inactivates PAK1, while activating AMPK.

However, none of these PAK1-blocking compounds is available on the market for clinical application as yet. So far only a bee product(s) called "propolis" which blocks PAK1 through its anti-oncogenic ingredients including CAPE and ARC (artepillin C) is widely available as a health food/supplement for the treatment of cancers and other PAK1-dependent diseases/disorders since the ancient Egyptian era.

Continued

A brief history of our PAK research "for improving our QOL and longevity"

The dawn of our PAK research could be traced back to 1977 when our team at NIH found the very first member of PAK (RAC/CDC42-activated kinase) family called "myosin I heavy chain kinase" (MHCK)  in a soil amoeba (1). The F-actin-induced activation of ATPase activity of this myosin requires this kinase that phosphorylates the heavy chain. In other words this kinase is essential for the acto-myosin-based cell motility/migration (so-called  "amoeboid movement").

Around 1994, a similar kinase was cloned by Ed Manser in mammals, and coined PAK1, for p21 (RAC/CDC42)-activated kinase 1 (2).  A few years later, PAK1 was found by us and others to be essential for the RAS-induced malignant transformation of normal cells. Since then (around the turn of century), PAK1 has been the major target for developing anti-cancer drugs. Interestingly, PAK1 is not essential for the growth of normal cells. Thus, unlike the conventional anti-cancer drugs (DNA/RNA/microtubule poisons), PAK1-blockers would not cause any side effect.

More recently, PAK1 was found to be essential for many other diseases such as Alzheimer's diseases (AD), diabetes (type 2), obesity, hypertension, a variety of inflammatory and infectious diseases, as well as neuronal diseases/disorders such as epilepsy, depression, schizophrenia, autism and even headache. Furthermore, a few years ago, we revealed that PAK1-deficient mutant (RB689) of C. elegans lives significantly (by more than 50%) longer than the wild-type (3).

Thus, there is no doubt that PAK1-blockers (synthetic or natural) would have a potentially "huge"  market value in pharmaceuticals. Furthermore, this year our team at "PAK Research Center" in Okinawa found that PAK1 is essential for melanogenesis in skin cells as well. Among the best known natural PAK1-blockers is a bee product called "propolis", and it was recently found to promote hair growth in vivo. Thus, these PAK1-blockers would gain an additional attraction from cosmetics industry for skin-whitening and therapy of hair loss （alopecia).

References:

１．Maruta, H. and Korn, E.D. Acanthamoeba cofactor protein is a heavy chain kinase required for actin activation of the Mg2+-ATPase activity of Acanthamoeba myosin I.  J  Biol  Chem. 1977, 252, 8329-32.

２．Manser, E., Leung, T., Salihuddin, H., Zhao, ZS. et al. A brain Ser/ Thr protein kinase activated by Cdc42 and Rac1. Nature. 1994, 367, 40-6.
３．Yanase S, Luo Y, Maruta H. PAK1-deficiency/down-regulation Reduces Brood Size, Activates HSP16.2 Gene and Extends Lifespan in C. elegans. Drug Dev Ther. 2013, 7: 29-35.