13.3 Alkylating agents

Synthetic drugs that are not cell cycle-specific are mostly alkylating agents . They have diverse reactive groups. Several drugs share the 'N-mustard' structure shown in Figure 13.3-1a. The reaction mechanism (formation of aziridine intermediates that react as electrophiles; Figure 13.3-1b) is actually the same as previously discussed for the irreversible α-blocker phenoxybenzamine; here, however, we don't have a moiety that targets the drug to any particular protein. It should go without saying that most molecules will actually not react with DNA but instead with some other nucleophile hopping about in the cell, in particular glutathione or other sulfhydryls. However, those that do react with DNA are the ones that matter, since the harm done by them has the potential to be permanent. Also note that we have not one but two chloroethyl groups – this creates the possibility of introducing cross-links into the DNA¹. Cross-links between the two strands of the DNA bases are more likely to give rise to permanent mutations than modifications affecting one strand only, since they cannot be removed by excision repair (although there are 'error-prone repair' mechanisms that may remove them).

The most common target of alkylation – quite independently of the alkylating agent used – is the N7 in guanine (Figure 13.3-1b,c). Why is that? The six-membered ring that is part of the guanine is locked up in base pairing within the double helix, i.e. inaccessible; this lets out the other nitrogens. The same applies to the other nucleotides. Why would N7 in guanine be more commonly affected than N7 in adenine? The guanine ring is not as completely aromatic as the adenine ring is. The π electrons of the ring nitrogens are therefore not as completely delocalized, i.e. the nitrogens will be stronger nucleophiles. However, the preference is not absolute, and alkylations of adenine and the pyrimidine bases do occur as well.

An interesting consequence of guanine N7-alkylation is the increased propensity of the guanine ring to adopt the tautomeric form (Figure 13.3-1c). In this form, the arrangement of hydrogen bond donors and acceptors is reversed and now resembles that of adenine, thus enabling guanine to base pair with thymine instead of cytosine. This is believed to contribute to the mutagenic effect of guanine alkylation.

One very commonly used agent containing the di-chloroethyl-amine moiety is cyclophosphamide. This drug may actually be metabolized quite extensively and give rise to several toxic metabolites, the exact contribution of which to the overall therapeutic effect is not very well established (Figure 13.3-2a). Metabolism is initiated by a cytochrome P450 enzyme in the liver (and possibly elsewhere) and continued by several other enzymatic and non-enzymatic steps. One of the final decay products is acrolein², which may form several adducts with guanine, some of which are shown in Figure 13.3-2b. While the metabolites of cyclophosphamide such as acrolein and chloroacetaldehyde seem to be quantitatively more important in the ultimate reactions with DNA than the parent drug itself, the relative significance of individual metabolites is somewhat hard to determine from the literature.

Alkylating agents with other reactive groups do exist – e.g., busulfan and cis-platinum (Figure 13.3-3) – but we will lightly gloss over their respective intricacies, in particular the chemical mechanism of cisplatinum and only note that the preferentially formed lesion observed with the latter drug again involves guanines; two that are located adjacently within the same strand become cross-linked to each other.

What are the biological consequences of the covalent modifications caused by alkylating agents? One that we have already noted is the introduction of mutations opposite a modified base. Another one is the inhibition of DNA synthesis; e.g., some of the acrolein adducts of guanine (Figure 13.3-2b) inhibit the incorporation of any base opposite to them, thus interfering with DNA synthesis and repair. Of note, many of these chemical adducts are removed by DNA repair enzymes only inefficiently or not at all; the enzymes are apparently not 'accustomed' to these peculiar types of modifications³.

1: Or of cross-linking the DNA with some other molecule, e.g. a protein, which might be even more deleterious (I don't know whether this has actually been studied).

2: Acrolein also forms when frying food and may be one of the agents responsible for the rising frequency of colon cancer in western countries.

3: On the other hand, natural radioactivity has been noticeable throughout evolution, and radiation-induced damages are usually efficient triggers of DNA repair.