N1-Benzyl Tryptamine Pan-SHIP1/2 Inhibitors: Synthesis and Preliminary Biological Evaluation as Anti-Tumor Agents

Inhibition of phosphatidylinositol 3,4,5-trisphosphate 5-phosphatase (SHIP) with small molecule inhibitors leads to apoptosis in tumor cells. Inhibitors that target both SHIP1 and SHIP2 (pan-SHIP1/2 inhibitors) may have benefits in these areas since paralog compensation is not possible when both SHIP paralogs are being inhibited. A series of tryptamine-based pan-SHIP1/2 inhibitors have been synthesized and evaluated for their ability to inhibit the SHIP paralogs. The most active compounds were also evaluated for their effects on cancer cell lines.


Introduction
The PI3K pathway is a major cell signaling pathway in eukaryotic cells, having been shown to have a strong influence on cell division and cell survival. The pathway is named for the phosphoinositide 3-kinase (PI3K) enzyme (in this case referring specifically to the PIK3CA isoform) [1], which phosphorylates the membrane-anchored inositol phospholipid PI(4,5)P 2 to form PI(3,4,5)P 3 . These membrane-anchored inositol phospholipids function as secondary messengers in the PI3K pathway. The diverse phosphorylation patterns present on the inositol ring of these lipids are recognized by PH domains present in signaling molecules such as serine/threonine kinases (such as Akt) which are typically activated upon binding to the inositol phospholipids and also by subsequent phosphorylation by other protein kinases (e.g., 3-phosphoinositide-dependent protein kinase-1, PDK1). This activation initiates a signaling cascade that transduces signals from the membrane to the cell nucleus. Signals from this pathway influence many cellular events, including cell division and survival. Modulating the concentration of the different inositol phospholipids on the interior of the plasma membrane has become a focus in the treatment of cancer, as controlling the amounts of these molecules may be used to curtail undesired cell division or survival. Inhibition of PI3K has been heavily investigated [2][3][4], with a number of clinically relevant PI3K inhibitors having been developed. Resistance to PI3K inhibitors has now been reported under some circumstances [5][6][7], so investigations into alternative approaches to influence PI3K signaling have been initiated. Modulation of the SH2-containing 5'-inositol phosphatases SHIP1 and/or SHIP2, inositol phosphatases that degrade PI(3,4,5)P 3 , may provide another approach to sway PI3K pathway signaling. While SHIP is often considered as reversing signals initiated from PI3K phosphorylation, such as the well-known tumor promoter phosphatase and tensin homolog protein (PTEN) [8], the situation is more complex, with significant evidence pointing to SHIP1/2 facilitating tumor cell survival [9][10][11][12][13][14][15]. The primary difference between PTEN and the SHIP paralogs can be attributed to PTEN hydrolyzing the 3 phosphate to generate PI(4,5)P 2 from PI(3,4,5)P 3 , while SHIP1 and SHIP2 are 5 phosphatases which convert PI(3,4,5)P 3 to PI(3,4)P 2 . Therefore, SHIP and PTEN exert distinct effects on downstream signaling. Additionally, recent results show that the Akt2 isoform binds with greater affinity to PI(3,4)P 2 [16], the product of the SHIP enzymes, Consistent with the "Two PIP Hypothesis", levels of the SHIP1 product PI (3,4)P2 are increased in several leukemia cells lines [19], while increased amounts of PI (3,4)P2 in inositol polyphosphate 4-phosphatase type II (INPP4A/B) knockout mice are implicated in promoting tumor formation [20,21]. We have also shown that the SHIP1 inhibitor 3αaminocholestane (3AC) reduces Akt activation and promotes apoptosis of human blood cell cancers that express SHIP1, further supporting the hypothesis [9,14]. The key role of PI (3,4)P2 in cancer cell signaling was confirmed in these studies by demonstrating that the addition of PI (3,4)P2 onto leukemia cells is protective from apoptosis by SHIP1 inhibition in a dose-dependent fashion [9]. Related studies have shown that a SHIP1 agonist also slows the growth of multiple myeloma cells in vitro [22]. The observation that both agonists and antagonists of SHIP1 are cytotoxic to multiple myeloma cells highlights the delicate balance of PI (3,4,5)P3 and PI (3,4)P2 a cancer cell must maintain in order to develop and perpetuate the malignant state. Altogether a number of different breakdowns in PI3K signaling are possible in order for a cancer cell to reach a malignant state that satisfies the "Two PIP Hypothesis".
There may also be applications for pan-SHIP1/2 inhibitors in the treatment of several types of cancer. As stated previously, a SHIP1 inhibitor showed cytotoxic effects on leukemia cells [9]. Other cancers may be more sensitive to SHIP2 inhibition, as SHIP2 expression is increased in breast cancer and promotes survival signals from epidermal growth factor receptor (EGF-R) in these tumors [11,13]. Further studies from our groups have shown that pan-SHIP1/2 inhibitors are also quite effective at slowing the growth of tumor cells [14,23]. Even though small molecule inhibitors are available for Akt [24], SHIP modulation provides a different molecular mechanism to influence PI3K signaling. This is advantageous for treating tumors that have become resistant to Akt inhibitors. The ability to produce a pan-SHIP1/2 inhibitor could prevent tumors from developing a resistance to inhibition of one SHIP paralog leading them to then utilize the other SHIP paralog to produce PI (3,4)  Consistent with the "Two PIP Hypothesis", levels of the SHIP1 product PI(3,4)P 2 are increased in several leukemia cells lines [19], while increased amounts of PI(3,4)P 2 in inositol polyphosphate 4-phosphatase type II (INPP4A/B) knockout mice are implicated in promoting tumor formation [20,21]. We have also shown that the SHIP1 inhibitor 3αaminocholestane (3AC) reduces Akt activation and promotes apoptosis of human blood cell cancers that express SHIP1, further supporting the hypothesis [9,14]. The key role of PI(3,4)P 2 in cancer cell signaling was confirmed in these studies by demonstrating that the addition of PI(3,4)P 2 onto leukemia cells is protective from apoptosis by SHIP1 inhibition in a dose-dependent fashion [9]. Related studies have shown that a SHIP1 agonist also slows the growth of multiple myeloma cells in vitro [22]. The observation that both agonists and antagonists of SHIP1 are cytotoxic to multiple myeloma cells highlights the delicate balance of PI(3,4,5)P 3 and PI(3,4)P 2 a cancer cell must maintain in order to develop and perpetuate the malignant state. Altogether a number of different breakdowns in PI3K signaling are possible in order for a cancer cell to reach a malignant state that satisfies the "Two PIP Hypothesis".
There may also be applications for pan-SHIP1/2 inhibitors in the treatment of several types of cancer. As stated previously, a SHIP1 inhibitor showed cytotoxic effects on leukemia cells [9]. Other cancers may be more sensitive to SHIP2 inhibition, as SHIP2 expression is increased in breast cancer and promotes survival signals from epidermal growth factor receptor (EGF-R) in these tumors [11,13]. Further studies from our groups have shown that pan-SHIP1/2 inhibitors are also quite effective at slowing the growth of tumor cells [14,23]. Even though small molecule inhibitors are available for Akt [24], SHIP modulation provides a different molecular mechanism to influence PI3K signaling. This is advantageous for treating tumors that have become resistant to Akt inhibitors. The ability to produce a pan-SHIP1/2 inhibitor could prevent tumors from developing a resistance to inhibition of one SHIP paralog leading them to then utilize the other SHIP paralog to produce PI(3,4)P 2 . Such SHIP paralog compensation was in fact observed in vivo in a xenogeneic multiple myeloma model treated with the SHIP1-selective inhibitor 3αaminocholestane (3AC). While 3AC was found to be effective in protecting the majority of mice from multiple myeloma growth; however, in those where there was relapse the tumor had up-regulated endogenous SHIP2 expression [14].
A number of groups have reported the identification of small molecules that modulate SHIP activity [25][26][27]. Our own work in the area began with a high-throughput screen (HTS) used to identify molecules that inhibit recombinant SHIP1, which identified a number of lead compounds [9,14,28]. One inhibitor found from this screen was tryptamine K103 (1) (Figure 2), a close analog of the serotonin antagonist benanserin 2 [29,30]. vivo in a xenogeneic multiple myeloma model treated with the SHIP1-selective inhibitor 3α-aminocholestane (3AC). While 3AC was found to be effective in protecting the majority of mice from multiple myeloma growth; however, in those where there was relapse the tumor had up-regulated endogenous SHIP2 expression [14].
A number of groups have reported the identification of small molecules that modulate SHIP activity [25][26][27]. Our own work in the area began with a high-throughput screen (HTS) used to identify molecules that inhibit recombinant SHIP1, which identified a number of lead compounds [9,14,28]. One inhibitor found from this screen was tryptamine K103 (1) (Figure 2), a close analog of the serotonin antagonist benanserin 2 [29,30]. K103 showed inhibition of both SHIP paralogs, marking it as a pan-SHIP1/2 inhibitor, but the molecule showed no effect on OCRL, another 5′ inositol phosphatase [14]. In line with the "Two PIP Hypothesis" the molecule demonstrated significant antitumor effects on several cell lines, especially on breast cancer cells [14,23]. Additional studies on K103 showed that inhibition of SHIP1/2 in multiple myeloma cells resulted in a G2/M cell cycle arrest, followed by caspase cascade activation leading to extensive apoptosis [14]. Tryptamine 1 is compliant with commonly used measures of druglike small molecule properties [31,32], but during the course of this work, it was discovered that 1 induced a psychotropic effect in mice which limited the usefulness of the molecule in vivo. Therefore, some synthetic studies on this tryptamine were initiated to define what features of the molecule were needed to maintain pan-SHIP1/2 inhibition so that an inhibitor with good pharmacodynamic properties and an improved side effect profile could be designed.

Results and Discussion
The synthesis of K103 and related tryptamines was undertaken as shown in Scheme 1. The tryptamine core was accessed from a Fischer indole synthesis [33] with ketone 3 and hydrazine 4 as previously reported [34]. The indole nitrogen was then alkylated with benzyl bromides or benzyl chlorides using cesium carbonate in refluxing acetonitrile using the method of Fink [35]. Removal of the phthalimide protecting group with hydrazine and formation of the HCl salt of the resulting amine resulted in K103 1 as well as the analogs 7b-7e. A 5-thiomethyltryptamine with no benzyl group on the indole nitrogen (7f) was also prepared by deprotection of the phthalimide of 5 with hydrazine followed by the formation of the HCl salt.   K103 showed inhibition of both SHIP paralogs, marking it as a pan-SHIP1/2 inhibitor, but the molecule showed no effect on OCRL, another 5 inositol phosphatase [14]. In line with the "Two PIP Hypothesis" the molecule demonstrated significant antitumor effects on several cell lines, especially on breast cancer cells [14,23]. Additional studies on K103 showed that inhibition of SHIP1/2 in multiple myeloma cells resulted in a G2/M cell cycle arrest, followed by caspase cascade activation leading to extensive apoptosis [14]. Tryptamine 1 is compliant with commonly used measures of druglike small molecule properties [31,32], but during the course of this work, it was discovered that 1 induced a psychotropic effect in mice which limited the usefulness of the molecule in vivo. Therefore, some synthetic studies on this tryptamine were initiated to define what features of the molecule were needed to maintain pan-SHIP1/2 inhibition so that an inhibitor with good pharmacodynamic properties and an improved side effect profile could be designed.

Results and Discussion
The synthesis of K103 and related tryptamines was undertaken as shown in Scheme 1. The tryptamine core was accessed from a Fischer indole synthesis [33] with ketone 3 and hydrazine 4 as previously reported [34]. The indole nitrogen was then alkylated with benzyl bromides or benzyl chlorides using cesium carbonate in refluxing acetonitrile using the method of Fink [35]. Removal of the phthalimide protecting group with hydrazine and formation of the HCl salt of the resulting amine resulted in K103 1 as well as the analogs 7b-7e. A 5-thiomethyltryptamine with no benzyl group on the indole nitrogen (7f) was also prepared by deprotection of the phthalimide of 5 with hydrazine followed by the formation of the HCl salt. Analogs without the thiomethyl group on the tryptamine were also prepared (Scheme 2). The synthesis of these compounds employs phenylhydrazine hydrochloride 8 and ketone 3 in the Fischer indole reaction. Alkylation of the indole nitrogen in compound 9 with benzyl bromide or 2-chlorobenzylbromide gave the alkylated indoles 10a and 10b, respectively. These compounds were then deprotected leading to 11a and 11b after HCl salt formation. Analogs without the thiomethyl group on the tryptamine were also prepared (Scheme 2). The synthesis of these compounds employs phenylhydrazine hydrochloride 8 and ketone 3 in the Fischer indole reaction. Alkylation of the indole nitrogen in compound 9 with benzyl bromide or 2-chlorobenzylbromide gave the alkylated indoles 10a and 10b, respectively. These compounds were then deprotected leading to 11a and 11b after HCl salt formation. Analogs without the thiomethyl group on the tryptamine were also prepared (Scheme 2). The synthesis of these compounds employs phenylhydrazine hydrochloride 8 and ketone 3 in the Fischer indole reaction. Alkylation of the indole nitrogen in compound 9 with benzyl bromide or 2-chlorobenzylbromide gave the alkylated indoles 10a and 10b, respectively. These compounds were then deprotected leading to 11a and 11b after HCl salt formation.

Scheme 2. Synthesis of sulfur-free K103 analogs.
Substitution on the tryptamine primary amine was also briefly explored (Scheme 3). The primary amine was converted to methylsulfonamide 15 and acetamide 16 to determine if a basic nitrogen was needed for SHIP inhibition. Reduction of the amide 16 led to the ethylamine derivative 17, while treatment of the parent compound with excess paraformaldehyde under reductive amination conditions gave the dimethylamino derivative 14. The methylamine analog was also accessed by the formation of the Bocprotected amine 12, followed by reduction to the methylamine with LiAlH4 and then the formation of the HCl salt 13. This method provided more consistent results than reductive amination methods, which usually led to a mixture of starting material, methylamine and dimethylamine when one equivalent of paraformaldehyde was used.    Substitution on the tryptamine primary amine was also briefly explored (Scheme 3). The primary amine was converted to methylsulfonamide 15 and acetamide 16 to determine if a basic nitrogen was needed for SHIP inhibition. Reduction of the amide 16 led to the ethylamine derivative 17, while treatment of the parent compound with excess paraformaldehyde under reductive amination conditions gave the dimethylamino derivative 14. The methylamine analog was also accessed by the formation of the Boc-protected amine 12, followed by reduction to the methylamine with LiAlH 4 and then the formation of the HCl salt 13. This method provided more consistent results than reductive amination methods, which usually led to a mixture of starting material, methylamine and dimethylamine when one equivalent of paraformaldehyde was used. Analogs without the thiomethyl group on the tryptamine were also prepared (Scheme 2). The synthesis of these compounds employs phenylhydrazine hydrochloride 8 and ketone 3 in the Fischer indole reaction. Alkylation of the indole nitrogen in compound 9 with benzyl bromide or 2-chlorobenzylbromide gave the alkylated indoles 10a and 10b, respectively. These compounds were then deprotected leading to 11a and 11b after HCl salt formation. Scheme 2. Synthesis of sulfur-free K103 analogs.
Substitution on the tryptamine primary amine was also briefly explored (Scheme 3). The primary amine was converted to methylsulfonamide 15 and acetamide 16 to determine if a basic nitrogen was needed for SHIP inhibition. Reduction of the amide 16 led to the ethylamine derivative 17, while treatment of the parent compound with excess paraformaldehyde under reductive amination conditions gave the dimethylamino derivative 14. The methylamine analog was also accessed by the formation of the Bocprotected amine 12, followed by reduction to the methylamine with LiAlH4 and then the formation of the HCl salt 13. This method provided more consistent results than reductive amination methods, which usually led to a mixture of starting material, methylamine and dimethylamine when one equivalent of paraformaldehyde was used.    These compounds were then tested for their activity against the two SHIP paralogs, SHIP1 and SHIP2 in the colorimetric malachite green phosphatase assay [36]. The malachite green assay detects the amount of free phosphate produced from enzymatic reactions with SHIP and an inositol phospholipid substrate [9]. The Malachite assay is useful for ruling out compounds that are fluorescent or have aggregation issues and is inexpensive, so it was utilized as the primary screen even though it is not as sensitive as other assays. Compounds that showed significant inhibition of SHIP in the Malachite had their IC 50 against SHIP1 determined using a Fluorescence Polarization (FP) assay that we have previously described [9,14]. The FP assay is significantly more sensitive than the Malachite assay, and so this assay was used to determine IC 50 of inhibitors that appeared active in the Malachite assay.
Testing of these tryptamines indicated that the halogen on the N-benzyl group seemed to improve inhibition of SHIP, as compounds 1, 7b-7d and 11b showed significant activity while 7f and 11a showed virtually no activity against SHIP1 or SHIP2 ( Table 1). Deletion of the thiomethyl group did allow for an active compound (11b), but the determination of the IC 50 showed that there was some loss of activity when compared to 1. Interestingly, returning the thiomethyl group to the 5-position of the indole along with the addition of an unsubstituted benzyl group did provide an active analog (7e), appearing to indicate that the thiomethyl is more responsible for activity than the halogen on the N-benzyl group. Modification of the N-benzyl group by moving the chloride to the 4-positon (7b) or using a 2,3-dichlorbenzyl (7d) provided analogs with significant activity, while the addition of a 4-bromobenzyl group gave less active analog. These compounds were then tested for their activity against the two SHIP paralogs, SHIP1 and SHIP2 in the colorimetric malachite green phosphatase assay [36]. The malachite green assay detects the amount of free phosphate produced from enzymatic reactions with SHIP and an inositol phospholipid substrate [9]. The Malachite assay is useful for ruling out compounds that are fluorescent or have aggregation issues and is inexpensive, so it was utilized as the primary screen even though it is not as sensitive as other assays. Compounds that showed significant inhibition of SHIP in the Malachite had their IC50 against SHIP1 determined using a Fluorescence Polarization (FP) assay that we have previously described [9,14]. The FP assay is significantly more sensitive than the Malachite assay, and so this assay was used to determine IC50 of inhibitors that appeared active in the Malachite assay.
Testing of these tryptamines indicated that the halogen on the N-benzyl group seemed to improve inhibition of SHIP, as compounds 1, 7b-7d and 11b showed significant activity while 7f and 11a showed virtually no activity against SHIP1 or SHIP2 ( Table 1). Deletion of the thiomethyl group did allow for an active compound (11b), but the determination of the IC50 showed that there was some loss of activity when compared to 1. Interestingly, returning the thiomethyl group to the 5-position of the indole along with the addition of an unsubstituted benzyl group did provide an active analog (7e), appearing to indicate that the thiomethyl is more responsible for activity than the halogen on the N-benzyl group. Modification of the N-benzyl group by moving the chloride to the 4-positon (7b) or using a 2,3-dichlorbenzyl (7d) provided analogs with significant activity, while the addition of a 4-bromobenzyl group gave less active analog. Compounds with modifications of the primary amine of the tryptamine were also tested. Acylation was first explored, but acetamide 16 proved to be difficult to test, as it tended to precipitate when mixed with buffer. The methylsulfonamide 15 oddly proved to be a weak agonist of SHIP1. While SHIP1 agonists may have applications as antiinflammatories [39,40], the agonist activity of 15 was weak and therefore was not further pursued. The addition of two methyl groups to the amine was also explored, with the Compounds with modifications of the primary amine of the tryptamine were also tested. Acylation was first explored, but acetamide 16 proved to be difficult to test, as it tended to precipitate when mixed with buffer. The methylsulfonamide 15 oddly proved to be a weak agonist of SHIP1. While SHIP1 agonists may have applications as antiinflammatories [39,40], the agonist activity of 15 was weak and therefore was not further pursued. The addition of two methyl groups to the amine was also explored, with the tertiary amine salt 14 being inactive. Monoalkylation with either a methyl (13) or ethyl group (17) did lead to analogs with activity against SHIP2; however, no inhibition of SHIP1 was observed. While these analogs may present a path towards selective inhibitors of SHIP2, these compounds will not be useful in determining the effects of pan-SHIP inhibition in tumor models.
The most active pan-SHIP1/2 inhibitors (1, 7b, 7d, and 7e) were then evaluated for cytotoxicity on several cancer cell lines using the MTT colorimetric assay for cell metabolic activity (Figure 3) [41]. The assay monitors the ability of NADPH-dependent cellular oxidoreductases to reduce the dye MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide), which when reduced produces a purple color that can be easily quantified. This assay is often used to determine cell viability [42]. Initially, the three new tryptamines and the parent compound 1 were evaluated on a number of leukemia cell lines (NB4, HSB2, K562 and 697-Pre-B) as well as the multiple myeloma cell line OPM-2. All of these cell lines express SHIP1 due to their hematopoietic origin, as SHIP1 is selectively expressed in blood and bone marrow cells [43]. Cytotoxicity was also found on the breast cancer cell lines MBA-MD-231 and MCF-7, which is likely due to the inhibition of SHIP2. SHIP2 is known to be overexpressed in breast cancer cells [11,13], and inhibition of SHIP2 has been shown to be cytotoxic to these cells previously [14,23]. Unexpectedly, the compound that exerted the greatest effect on these cell lines was 7d, which in nearly every cell line exerted the greatest cytotoxicity. While not quite as active as 1 in the FP assay, 7d is more hydrophobic with a ClogP of 4.6 vs. the ClogP of 4.11 for 1. Being more hydrophobic, tryptamine 7d may spend more time associated with the plasma membrane where SHIP activity is most important in signal transduction, and this may explain the greater influence on cells in the MTT assay.
Molecules 2022, 27, x FOR PEER REVIEW 6 of 16 tertiary amine salt 14 being inactive. Monoalkylation with either a methyl (13) or ethyl group (17) did lead to analogs with activity against SHIP2; however, no inhibition of SHIP1 was observed. While these analogs may present a path towards selective inhibitors of SHIP2, these compounds will not be useful in determining the effects of pan-SHIP inhibition in tumor models. The most active pan-SHIP1/2 inhibitors (1, 7b, 7d, and 7e) were then evaluated for cytotoxicity on several cancer cell lines using the MTT colorimetric assay for cell metabolic activity (Figure 3) [41]. The assay monitors the ability of NADPH-dependent cellular oxidoreductases to reduce the dye MTT (3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide), which when reduced produces a purple color that can be easily quantified. This assay is often used to determine cell viability [42]. Initially, the three new tryptamines and the parent compound 1 were evaluated on a number of leukemia cell lines (NB4, HSB2, K562 and 697-Pre-B) as well as the multiple myeloma cell line OPM-2. All of these cell lines express SHIP1 due to their hematopoietic origin, as SHIP1 is selectively expressed in blood and bone marrow cells [43]. Cytotoxicity was also found on the breast cancer cell lines MBA-MD-231 and MCF-7, which is likely due to the inhibition of SHIP2. SHIP2 is known to be overexpressed in breast cancer cells [11,13], and inhibition of SHIP2 has been shown to be cytotoxic to these cells previously [14,23]. Unexpectedly, the compound that exerted the greatest effect on these cell lines was 7d, which in nearly every cell line exerted the greatest cytotoxicity. While not quite as active as 1 in the FP assay, 7d is more hydrophobic with a ClogP of 4.6 vs. the ClogP of 4.11 for 1. Being more hydrophobic, tryptamine 7d may spend more time associated with the plasma membrane where SHIP activity is most important in signal transduction, and this may explain the greater influence on cells in the MTT assay.

Conclusions
In summary, an initial evaluation of some tryptamine pan-SHIP1/2 inhibitors has been conducted. The parent tryptamine and some analogous tryptamines have been synthesized and evaluated for activity against both SHIP paralogs in the Malachite Green assay. Several of these analogs showed significant activity against these inositol phosphatase enzymes. The most active analogs all contained chlorinated benzyl groups on the indole N1 nitrogen and a thiomethyl group at the 5-position on the indole. Some alkyl substitution was tolerated on the primary amine of the tryptamine, but disubstitution with alkyl groups or conversion to an amide or sulfonamide led significantly to reduced SHIP inhibition. The most active analogs showed good activity against a number of cancer cell lines. The analog that was best at killing cancer cells was not the most potent SHIP inhibitor, but this may be due to the lipophilicity of the compounds, as the most lipophilic compound was the most potent on cells and SHIP is known to be recruited the cell membrane. Future work will include efforts to further differentiate the structure from the tryptamine scaffold to avoid any psychotropic side effects.