Targeting the plasma membrane of neoplastic cells through alkylation: a novel approach to cancer chemotherapy

Many plasma membrane alkylating agents have been shown to exert significant physiological effects by chemical alteration of surface membrane electron donors, particularly at carbohydrate-specific receptor sites due to a reactive α-halo functional group, as well as their observed amphiphilic nature. In particular, halocarbonyls (haloketones, haloesters, and haloacetamides) have shown significant antitumor activity in murine models after only a single injection of an active agent [16‐18]. However, simple haloacetamides and haloacetates are not active against Ehrlich ascites murine carcinoma in CD2F1 mice [17], suggesting that active halo compounds have particular physicochemical properties atypical of common halo compounds.

Critical analysis of effective halo compounds in vivo has revealed such agents are able to partition from an aqueous to hydrophobic environment (amphiphilic log P value), and have a marked propensity to act as S_N2 alkylators that react with strong electron donors [21]. By contrast, related compounds that are only strong alkylators, or amphiphilic do not potentiate the same antitumor activity [16‐18]. Further, active haloacetamide compounds often have additional electron withdrawing groups which greatly increase alkylation potential. The pK_a of the parent amine from which active haloacetamide compounds are made correlates well with the alkylating activity of the antitumor agents, as is expected; the more acidic the pK_a, the more reactive the derived α-haloacetamide should be. In fact, their reactivity at the CH₂X group is controlled by the electron withdrawing or donating power of the R-group, and the corresponding pK_a of the parent amine (Fig. 1a). This also happens to be the case with nitrogen mustards which have long been used in the clinical setting due to their alkylating potential. However, haloacetamides have apparent hydrogen bond donating and accepting properties due to the presence of the amide nitrogen, indicating that compounds can be designed to have their activity modulated by the dielectric constant of the in vivo environment [21]. That localization could in turn be readily controlled by the lipid: water partition coefficient of the given haloacetamide. Further, nitrogen mustards react by S_N1 mechanisms [22], which are not partial to hydrophobic environments; ideal for alkylating hydrophilic DNA, but not the hydrophobic environment of the plasma membrane. Due to the S_N2 mechanisms observed in haloacetamides, the compounds react with electron donors even in nonpolar environments, allowing alkylation of functional groups on the surface of the plasma membrane to be a feasible prospect.

Experimental data comparing structural activity and chemotherapeutic effectiveness have confirmed that halo carbonyls act as non-charged hydrophobic or amphiphilic electrophiles. Further, the compounds with the most antitumor activity in these series are strong alkylating agents. Since antitumor activity is dependent on the polarity and alkylating activity of halo carbonyls, it has been postulated, and later confirmed that such analogs exert their effects by alkylating cell surface nucleophiles. To further improve the specificity of haloacetamides, the R group has been substituted with carbohydrate derivatives that are found on the plasma membrane. However, inserting a hydrophilic moiety into the base structure significantly reduces antitumor activity, as observed by haloacetamides with free sugars [17, 18]. Therefore, any carbohydrate structure inserted into the haloacetamide base structure needs to be altered so that the resulting compound retains its affinity for nonpolar environments.

One approach to circumvent this issue has been to acetylate the carbohydrate derivatives with acetic anhydride after being attached to the haloacetamide (Fig. 1b and c). From this approach, a comprehensive series of tetra-O-acetate haloacetamido carbohydrate analogs (Tet-OAHCs) have been developed to mimic carbohydrates used at the plasma membrane surface, including 2-deoxy-2-acetamido-β-D-glucose and its α anomer, 2-deoxy-2-acetamido-β-D-galactose, and 2-deoxy-2-acetamido-D-mannose. The mannose derivative is particularly important, as it is a metabolic precursor of sialic acid, an integral component of cell surface biochemistry [23‐25]. In fact, metastatic cancer cells often express a high density of sialic acid-rich glycoproteins, creating a negative charge on cell membranes [26, 27]. The subsequent repulsion between adjacent malignant cells is pivotal for the metastasis of many late stage cancers, providing an optimal target for haloacetamido mannose analogs.

The importance of the hexose tetra-O-acetate carbohydrate moiety for haloacetamides has become apparent in the treatment of tumor-bearing mice. Several compounds in the glucosamine, galactosamine, and mannosamine series have produced considerable rates of long term survival with almost all mice surviving an intraperitoneal (i.p.) injection of 2.5 × 10⁷ Ehrlich ascites carcinoma cells, a concentration that is ten times the minimum dose lethal to 95 % of B6D2F1 mice [21]. By contrast, simple haloacetamide and haloacetyl compounds were completely inactive in the tumor-bearing mammalian system [21]. Further, non-hexose bromoacetamido polyol acetates with alkylating activities and log P values spanning those of the active bromoacetamido hexose tetra-O-acetates did not exhibit antitumor activity [21]. Therefore, the chemotherapeutic potential of Tet-OAHCs requires more than the physicochemical properties of alkylating activity and the octanol: water partition coefficient, as it is associated with the specific interaction these agents can confer with functional groups on the surface of the plasma membrane.

It is has been well-established that carbohydrates are pivotal for leukocyte function. O-linked glycans of glycosylation-dependent cell adhesion molecule-1 (GLYCAM-1) are used by leukocytes to home in on sites of inflammation, as well as direct lymphocytes to lymph nodes [28]. Further, carbohydrates are critical for diapedesis, an innate form of extravasation in which carbohydrate ligands on the circulating leukocytes bind selectin molecules on the inner wall of the vessel, enabling the blood cells to slow down and begin rolling along the inner surface of the vessel wall until they have passed through the vascularized tissue [29, 30]. Therefore, it seems plausible that carbohydrate-based chemotherapeutic agents could have a profound influence on leukocyte influx at the site of administration.

When N-bromoacetyl-β-D-glucosamine tetra-O-acetate (NBrAcGlc-TA) was shown to produce consistent long term remission in Ehrlich tumor-bearing mice after a single injection (0.11 mmol/kg), it was observed that the sites of tumor challenge and drug administration had a rapid granulocyte influx [21]. Further, N-bromoacetyl-β-D-galactosamine tetra-O-acetate (NBrAcGal-TA) injected i.p. at the same dose one day prior to a challenge of 1 × 10⁶ Ehrlich ascites carcinoma cells conferred substantial resistance to lethal tumor outgrowth when compared to vehicle-treated control animals. Although plasma membrane alkylating chlorohydroxyacetone-benzoate derivatives were shown to produce long term survival in the same tumor system [31], they were not associated with an increased granulocyte influx, nor did pretreatment of such compounds one day prior to tumor challenge confer any noticeable resistance.

To further examine whether Tet-OAHCs potentiate a substantial in vivo immune response, NBrAcGlc-TA and chlorohydroxyacetone dinitrobenzoate (ClHA-DNB) were injected at their respective effective doses in B6D2F1 mice in the absence of tumor challenge [31]. Results from the study indicated that only the carbohydrate analog produced a marked early increase in granulocyte influx, as well as a pronounced increase in later lymphocyte response. A quantitative comparison of host leukocyte influx between carbohydrate analogs and another hydrophobic monofunctional non-carbohydrate alkylating agent, bromohydroxyacetone benzoate (BrHAB), later confirmed the in vivo observations [21]. It is worth noting that a difunctional alkylating nitrogen mustard was not able to confer any protection to tumor challenge when administered prior to the injection, and significantly reduced host resistance against Ehrlich ascites carcinoma due to its pronounced immunosuppressive effects [21].

Although most studies investigating the chemotherapeutic efficacy of plasma membrane alkylating agents have been done with Ehrlich ascites carcinoma, the compounds have shown promise in other cancer models as well. In particular, chlorohydroxyacetone esters were shown to produce long term survival in mice challenged with murine sarcoma 180 and P815 murine mastocytoma [31]. However, plasma membrane alkylating agents have demonstrated substantial efficacy in disseminated hematological malignancies as well. Tet-OAHCs were first examined in vitro against Friend murine erythroleukemia cells [17]. Results from the study indicated that the malignant cells in log phase suspension culture were substantially inhibited by low concentrations of bromoacetamido and chloroacetamido analogs of the glucose and galactose series.

The success of plasma membrane alkylating agents against Friend erythroleukemia in vitro prompted an in vivo study examining the effects of such compounds on the immunogenicity of the malignant cells. When DBA/2 J mice treated with chlorohydroxyacetone benzoate (ClHAB) survived a primary challenge of a single injection of 1 × 10⁶ tumorigenic cells, the survivors were completely resistant to rechallenge. By contrast, a single injection of sonically disrupted Friend erythroleukemia cells or of virus harvested from cultured cells had no effect on the survival time of animals subsequently rechallenged with intact cells, indicating that ClHAB potentiates an immunogenic response. Further, mice given injections of cells killed by exposure to 50 μM ClHAB in vitro exhibited increased survival times upon rechallenge. It was later demonstrated that Friend erythroleukemia cells treated for 3 h with 30 μM ClHAB were non-proliferative in subculture, and completely non-tumorigenic when implanted either i.p. or subcutaneously (s.c.) into DBA/2 J hosts. In fact, mice receiving a single injection of 1 × 10⁶ cells altered under these conditions were fully protected against rechallenge. These results are indicative of substantial reductions in tumorigenicity, as it is known that as few as 10 cells of Friend erythroleukemia are needed to kill murine models [32, 33]. Therefore, controlled modification of Friend erythroleukemia cells by exposure to relatively low concentrations of ClHAB abrogated the tumorigenicity of treated cells, without altering their ability to confer protection to subsequent rechallenge with unaltered cells.

Although carbohydrate analogs offer a potentially specified approach to alkylate functional groups vital for neoplastic cell stability, it is likely that some cells within various cancer types will have inherent or acquired resistance to the series of compounds. Therefore, finding additional agents that can supplement the mechanisms of plasma membrane alkylating agents is clinically relevant, as it will give clinicians a greater diversity of options in which to treat problematic tumorigenic growths.

NAD-linked glycerol-3-phosphate dehydrogenase (GPDH) is a key enzyme linking carbohydrate metabolism with the synthesis and degradation of glycerolipids, yet it is present in low amounts or entirely absent in most neoplastic tissues [34‐36]. Since phospholipids are required for the synthesis of new membranes in proliferating cells, absence of this enzyme in malignant cells is profound, and indicates that alternative pathways to glycerolipids are pivotal for the potentiation of neoplastic tissue. Such an alternative pathway exists in which dihydroxyacetone phosphate (DHAP) is acylated and reduced to lysophosphatidic acid by a microsomal NADP-linked enzyme entirely independent of cytosolic GPDH [37‐40]. The acyl-DHAP pathway results in phospholipid synthesis, as well as production of ether lipids that are often elevated in various cancer types, indicating that such compounds influence the abnormal growth patterns of these malignancies [35]. Therefore, DHAP analogs designed to exploit the aberrant activity of GPDH could profoundly interfere with lipid metabolism and membrane glycerol-lipid synthesis, making affected cells much more sensitive to chemotherapeutic agents that directly attack the plasma membrane.

Attempts have been made to exploit these potential differences and to alter membrane-mediated phenomena in neoplastic tissue directly at the level of membrane lipids, as well as indirectly by affecting lipid metabolic pathways. Success has been achieved with halo analogs of GPDH and DHAP [41‐43] which are designed to exploit the differences in GPDH production between malignant and normal cells. The glycerol analogs directly exploit the difference in GPDH, while the DHAP analogs influence the GPDH and acyl-DHAP pathways. These analogs have shown some in vivo antitumor activity, but they experience difficulties traversing the plasma membrane, as they are charged, phosphorylated compounds [44]. Therefore, non-phosphorylated derivatives of dihydroxyacetone may be pivotal for increasing chemotherapeutic activity. This reasoning is in alignment with previous observations that monobenzoate esters of dihydroxyacetone inhibit the acyl-DHAP pathway. However, the inhibitory effect of monobenzoate esters is substantially limited in vivo by endogenous kinase activity which phosphorylates at the 1′ position of the compounds, resulting in the formation of 1,3-DHA-monobenzoate, a non-inhibitory derivative [45]. Since non-phosphorylated halo derivatives of dihydroxyacetone are not susceptible to kinase inactivation, they could be used as potent inhibitors of the acyl-DHAP pathway.

Although a considerably diverse set of plasma membrane alkylating agents have been examined for chemotherapeutic potential, they are only preliminary investigations in what could turn out to be a large class of similarly designed chemotherapeutic agents. Plasma membrane alkylating agents are only in their initial stages of development, and many more compounds can be designed and studied for anticancer activity. For example, it has been observed that haloacetamide compounds that exhibit alkylating potential have their reaction rate dictated by hydrogen bonding [17]. While hydrogen bond accepting by C = O, C-O, and C-N structures accelerate the rate of alkylation, hydrogen bond donation by N-H decreases observed activity. It would therefore seem apparent that agents without a hydrogen on the amide nitrogen could be particularly effective alkylators. Further, since hydrogen bonding can be exploited as a means to enhance alkylation in a hydrophobic environment, compounds should be designed to examine potential therapeutic benefits conferred from such activity. Pragmatic requirements to observe hydrogen bond controlled enhancement of alkylation by haloacetamides in low polarity medium are as follows: 1) Hydrogen bond donation by N-H is an effective gain of an electron by N, indicating that the alkylation rate should decrease; 2) Hydrogen bond acceptance by N-H or C = O is an effective withdrawal of an electron, inherent for acceleration of the alkylation rate; 3) Intramolecular hydrogen bonds forming 5 and 6-membered rings (including the shared hydrogen) would be highly favored due to the stabilization of ring formation; 4) Formation of ring forming intramolecular hydrogen bonding would be favored in a medium of a low dielectric constant (such as the plasma membrane) compared to a polar medium; 5) Multiple intramolecular hydrogen bonds would be more effective than solvent donated hydrogen bonding in enhancing alkylation rate; 6) Any intramolecular hydrogen bonding should significantly enhance the rate of alkylation. Due the potential importance intramolecular hydrogen bonding could have in increasing plasma membrane alkylating activity, several novel compounds have been posited for chemotherapeutic investigation (Fig. 2a and b).

One of the most profound characteristics of plasma membrane alkylating haloacetamides is their ability to potentiate a substantial granulocyte influx at the drug administration and primary tumor site when attached to a tetra-O-acetylated carbohydrate. Not only does administration of such compounds 24 h after tumor challenge result in particularly high cure rates and resistance to rechallenges, but pretreating mice with the congeners effectively prevents tumor formation. Although consistent alkylation of the plasma membrane is enough to warrant further investigation due to the importance of cell surface biochemistry in neoplastic growth, the ability to potentiate an immune response at the primary tumor site has notable clinical implications, and further substantiates the utility of acetylated carbohydrate haloacetamide analogs. It is apparent that these compounds could have immediate therapeutic potential, and novel congeners have been proposed to further examine this intriguing in vivo response (Fig. 2c and d). The compounds utilize analogs of carbohydrates typically observed on the cell surface to promote alkylating specificity. In addition to monofunctional carbohydrate derivatives, difunctional haloacetamide compounds are also posited as a means to further increase alkylation rates. It is also well known that difunctional alkylating agents have the propensity to crosslink adjacent nucleophiles [1], and in theory should give haloacetamido analogs another mechanism by which to damage neoplastic cells.

Although this review has commented on novel plasma membrane-directed agents, there is a class of clinically approved antineoplastic agents that has a marked influence at the cell surface. Anthracyclines are most commonly known as nucleic acid-directed agents, as the compounds intercalate base pairs, produce free radicals, and are potent inhibitors of DNA topoisomerase II [4]. However, it has also been well established that anthracyclines (particularly doxorubicin) alter the fluidity of neoplastic cell plasma membranes [46, 47], and bind phospholipids with considerably affinity [48, 49]. Some studies have also indicated that extracellular doxorubicin is important for anticancer activity and that the compound demonstrates marked cytotoxicity without entering the cell [4, 50].

While the combination of anthracyclines and plasma membrane alkylating agents has yet to be assessed for drug synergy, this concomitant chemotherapeutic approach is an intriguing possibility that would give these novel agents more versatility if they ever reached the clinical setting. Indeed, anthracyclines and DNA-directed alkylating agents are often used in combination, with synergy being attributed to the production of free radicals and alkylated adducts. Interestingly, it has been demonstrated that doxorubicin produces free radicals, including highly reactive hydroxyl species, at the cell surface. Therefore, it is possible that the same synergy between anthracyclines and DNA-directed alkylating agents may be reciprocated with plasma membrane alkylating agents, due to the combined effect of unstable free radicals and alkylated adducts. However the potential cardiotoxicity of plasma membrane directed agents has also yet to be assessed, and would likely decide whether these agents could be used concomitantly with anthracyclines, unless dexrazoxane or other cardiotoxic reducing agents mitigate any substantial aberrant drug interaction.