Pharmacogenomic Implications of Variants of Monoaminergic-related Genes in Geriatric Psychiatry

Paulo R Shiroma; Yonas E Geda; David A Mrazek


Pharmacogenomics. 2010;11(9):1305-1330. 

In This Article


A literature search using PubMed and OVID (01/01/1989 to 28/02/2009) was conducted. The terms used were 'pharmacogenetics' or 'pharmacogenomics' and 'elderly' or 'elder' or 'geriatric' or 'late life' or 'aged' and 'psychotropic drugs'. The search was limited to human studies published in English that contained the search terms in the article. Reference lists were gleaned to identify relevant articles that may have been overlooked by the search method.

The terms 'elderly' and 'geriatric' can be used interchangeably to identify adults aged 65 years and older. However, epidemiological, clinical, neuroimaging and genetic studies suggest that distinctive etiological factors in late-life psychiatric disorders may arise after the age of 50 years.[3–5] Therefore, the review contains articles with an a priori defined threshold age of 50 years and older. Articles based on the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study were included as this is the largest sample of elderly subjects who have been pharmacogenomically studied.

Biology of Aging, Altered Compensatory Mechanisms & Pharmacodynamic Changes

Pharmacodynamics is a function of the drug concentration at the receptor site, drug receptor interactions (i.e., number of receptor, affinity, second messengers and cellular response), and homeostatic regulation. Homeostatic reserve, defined as the capacity to maintain an internal equilibrium through physiological adjustments in response to environmental challenges (i.e., medication), is generally reduced in aging.[6]

Mechanisms contributing to functional changes in the aging brain include altered levels of neurotransmitters and receptors, hormonal changes, especially sex and growth hormones, and decreased availability of glucose and oxygen.[7] Age-related changes in monoaminergic systems influence efficacy as well as the occurrence of adverse events by altering pharmacodynamic properties of psychiatric drugs (Box 1). Age-related changes may alter the blood–brain barrier resulting in higher drug levels and prolonged drug exposure of the brain.[8]

Epigenetics & Drug Response in the Elderly

Epigenetics refers to reversible molecular changes to DNA, RNA or proteins (i.e., histones) that may be inheritable but do not involve DNA base changes, and that regulate gene expression, RNA or protein function. The two most extensively studied epigenetic mechanisms are DNA methylation and histone modifications. The DNA methylation of the genome occurs primarily on cytosines located in CpG (cytosine–phosphate–guanine) dinucleotides. The posttranslational modifications of histones proteins (i.e., acetylation) are more complex and affect multiple residues (Arg, Lys, Pro and Ser) at different positions in the N-terminal tails of histones.[9,10]

Epigenetic abnormalities in psychiatric disorders have been recently reported to schizophrenia, depression and drug addiction.[11] During normal aging, global DNA methylation declines,[12] although a number of specific loci have also been reported to be hypermethylated.[13] It appears that age-related hypermethylation preferentially affect loci at CpG islands, while loci devoid of CpG islands lose methylation with age.[14] The loss of DNA methylation was speculated to lead to an age-related decline in learning and memory performance by reducing adult neurogenesis in the hippocampus.[15] Hypomethylation of the promoter of the amyloid precursor protein gene occurs in the cerebral cortex of an aging human brain,[16] and is also observed in brain tissues from Alzheimer's disease patients.[17] How these epigenetic changes may affect cellular and molecular function in the aging process, and therefore, the development of drugs that turn on epigenetic mechanisms in the aging brain function remains unknown. Some psychiatric therapeutics has already been tested in animals and nonelderly subjects by targeting the epigenome. Imipramine, a tricyclic antidepressant, acts as a inhibitor of histones deacetylases and was shown to reverse the downregulation of brain-derived neurotrophic factor transcript in the hippocampus of animals subjected to long-term social defeat stress.[18] Valproate, widely used in bipolar disorder, is a histone deacetylase inhibitor and considered as an 'epigenetic softener'.[19] Haloperidol, clozapine, fluoxetine, sulpiride and sodium butyrate have also been demonstrated to alter gene-expression profiles in neurons and potential to benefit several neurological and psychiatric disorders.[20,21]

Serotonergic System & Aging

Limited data suggest a gradual loss of serotonin production and receptor function in selected regions of the aging brain.[22] Imipramine binding, which provides an estimate of serotonin reuptake transporter protein, is reduced with aging. The activity of tryptophan hydroxylase, which is necessary for the production of serotonin, is also reduced with advanced age.[23]

Serotonin Transporter Gene The serotonin transporter gene (SLC6A4 or SERT or 5-HTT) codes for the serotonin reuptake protein and removes serotonin from the synaptic cleft, returning it to the terminal neuron. This is believed to be the principal site of action of the selective serotonin reuptake inhibitors (SSRIs). Genetic mutations of SLC6A4 have been associated with response to antidepressants.[24–27]

The serotonin transporter insertion/deletion (indel) promoter variant (5-HTTLPR or rs4795541) consists of a series of 22-bp sequences (Figure 3). A short allele (S) has 14 copies of this sequence and a long allele (L) has 16 copies.[28] Individuals who are homozygous for the L allele produce more mRNA expression and greater serotonin transporter density in platelets than individuals with the S/S genotype.[24,29,30] The allele frequencies of the indel promoter polymorphism have been studied in diverse ethnic and racial groups (Table 1). The 5-HTTLPR gene has a common SNP variation (rs25531) that influences the expression of the L allele.[28,31] The G allele of rs25531 reduces SLC6A4 mRNA expression to a level that is believed to be equivalent to the expression of the S allele. The A allele of rs25531 has a higher expression. Thus, 5-HTTLPR is functionally triallelic (LA vs LG or S) rather than biallelic (L vs S). The allele frequency of the 5-HTTLPR gene in patients of European ancestry is 40% S, 50% LA, and 10% LG. In patients of African ancestry, the allelic frequency is 25% S, 51% LA and 24% LG. In Native Americans, the allelic frequency is 65% S, 34% LA, and 1% LG (1%).[20] The S allele can occur in up to 80% of individuals in some Asian populations.[32,33]

Figure 3.

SLC6A4 location and allelic variation of the indel promoter polymorphism (rs4795541).
MAOA: Monoamine oxidase A.
Adapted with permission from [214].

Serotonin Transporter-linked Polymorphic Region & Antidepressant Response in Older AdultsTable 2 summarizes research findings on the serotonin transporter gene in older subjects. In older subjects of European ancestry, patients with the 5-HTTLPR L/L genotype carriers have usually had a better response to paroxetine as compared with S allele carriers.[25,34,35] However, the preferential clinical response described in L allele, extended to other SSRIs, seems more prominent at early stages of treatment by affecting the speed of response. There was no significant difference in the final antidepressant outcome.[34,36]

Rausch and colleagues proposed that the enhanced responsivity associated with the L allele may primarily occur before saturation dose is reached.[37] The inhibition of serotonin reuptake by SSRIs at the initiation of treatment may only be the initial step in a complex adaptive process.[38] Paroxetine plasma concentration was found to influence the outcome at 2 weeks in S allele carriers but not in L/L subjects. When paroxetine concentration was less than 60 ng/ml and the subject was an S allele carrier, they were less likely to be in remission at 12 weeks with 88% sensitivity and 64% specificity.[35]

Subjects with a paroxetine concentration above 60 ng/ml had an equal likelihood of remission that was not related to their 5-HTTR genotype. If this finding is replicated, it would be possible to identify individuals who are likely to remit at an earlier point following treatment initiation. Early improvement has been associated with later response in older[39,40] and middle-aged adults.[41] One consideration is that rapid improvement results have been hypothesized to be associated with a placebo response, and a higher placebo response has been reported among subjects with the L allele.[37] An adequate run-in period before randomization could clarify the mechanisms of early pharmacological response.

Analyses on the influence of genetic variation in the SLC6A4 gene on citalopram treatment of depression have been conducted in STAR*D sample, which includes a large cohort of elderly individuals. A significant association of remission or response and genotype was not identified for the entire sample while adjusting for significant clinical (years of schooling) and clinical covariates (i.e., baseline depression).[42,43] However, a SLC6A4 gene resequencing demonstrated large allele and variant allele frequency differences between ethnic groups (white non-Hispanics, black and white Hispanic subjects), prompting analysis of ethnic subgroups separately. Thus, an analysis of the white non-Hispanic subgroup did find that the L allele was associated with a better antidepressant response,[44] suggesting that the genomic ancestral background varied in a systematic manner between ancestral groups. The genetic effect remained significantly associated with remission even after adjusting for baseline depression, days in the study, and final dose of antidepressant. In addition, analyses of rs25531 functional expression variants, namely LG + S alleles (lower function) versus LA allele (higher function), did not have different treatment outcomes when treated with citalopram.[44,45] The LA/LA genotype (p = 0.04) and the LA allele (p < 0.001) were associated with a higher frequency of adverse events burden when compared with the less active variants.[45]

Studies involving older Asians found that S allele carriers exposed to SSRIs and other nonserotonergic antidepressants were more likely to respond.[32,46] Studies in middle-aged Asians converged with a more favorable genetic profile of S/S genotype.[47] However, this was not true in all studies.[48,49]

Drug response is a complex human trait controlled by multiple polymorphic genes with both a small and wide range of effects.[50] Increased variability of antidepressant treatment may reduce the signal:noise ratio to establish significant genetic associations. While several SSRIs were described as having similar pharmacological effects on 5-HTTLPR genotypes and alleles,[51] other classes of antidepressants such as mirtazapine (noradrenergic and specific serotonergic antidepressant),[25] milnacipran (serotonin noradrenergic reuptake inhibitor [SNRI],[52] and nortriptyline (tricyclic antidepressant [TCA])[32] rendered different and even opposite clinical outcomes between alleles by tapping combined or nonserotonin monoaminergic systems (Table 2). Clinical decision to initiate treatment could be made on the allelic predicted probability of response or tolerability in combination with consideration of the specific mechanism of action of antidepressants.

Serotonin Transporter-linked Polymorphic Variability & Antidepressant-related Adverse Events Most pharmacogenomic studies have focused on efficacy rather than tolerability. Adverse drug reactions are reliably measured (e.g., either the patient experienced headaches or not), less susceptible to placebo effect than clinical improvement, and the mechanism may be better understood.[53] From a methodological perspective, efficacy should be understood within the context of side effects. If only efficacy is studied, survival bias may occur in retrospective studies. Adverse drug reactions also influence efficacy rating scales. For instance, elevated paroxetine levels may cause agitation and insomnia thus increasing the total score on Hamilton Depression Rating Scale (HAM-D) and paradoxically affecting a measure of response.[54]

With respect to adverse reactions of the antidepressants, the S allele has shown higher predisposition among older[25] and middle-aged individuals[55] treated with SSRIs although these associations are not completely consistent.[32,56] Older adults carrying the L/L genotype had increased adverse drug reactions to mirtazapine. The mechanism of action of mirtazapine is to induce the release of norepinephrine and serotonin by inhibiting a2 adrenergic receptors.[57] This may have interfered with the efficiency of L/L patients to reuptake serotonin in the synaptic cleft, hence increasing the frequency of side effects.

Other Variants in the Serotonin Transporter Gene & Antidepressant Response Previous studies have estimated that up to 7% of antidepressant response variance can be explained by 5-HTTLPR.[58] Rare variants for the L allele (16A, 16D and 16F) and two others for the S allele (14A and 14B) were examined in older patients to improve predictability.[59] Overall, the presence of these rare polymorphisms added 0.6% to the previous 7% with no difference in response rate by the presence of a different S allele. Subjects with the excess of 16F L allele (A of the A/G polymorphism together with a T instead of a C in the sixth nucleotide of the sixth repeat) may have more treatment resistant depression.

A rare gain-of-function missense variant of the SLC6A4 gene, which consists of an A to G transversion at nucleotide 1273 of the coding sequence, converts Ille425 into Val425.[60] These SLC6A4 gene mutations were discovered in two unrelated families with obsessive compulsive disorder (OCD), obsessive compulsive personality disorder and other serotonin-related disorders that included Asperger's syndrome, social phobia, anorexia nervosa, tic disorders and alcohol and other substance abuse/dependence.[33] Affected members showed a severe clinical course and are treatment resistant. The Ille425Val polymorphism leads to increased rates of serotonin transport,[60] which coincides with a previous report of high-functioning 5-HTTLPR L/L genotype associated with OCD.[61]

Serotonin Transporter Intron 2 Variable Number of Tandem Repeat The variable number of tandem repeats (VNTR) in the second intron of the serotonin transporter gene is a polymorphic region with 9, 10, 11 or 12 copies of a 16–17 bp repeat element.[62] The most common alleles are 10 and 12 alleles with the 12 allele designated as L allele and all others as S alleles. The variability in allele frequency of the serotonin transporter intron 2 (STin2) VNTR has been studied in different ethnic populations (Table 3). The STin2 VNTR is not located in the coding region of the serotonin gene but it might have a role in the regulatory element of gene transcription expression through an adjacent activator protein-1 motif.[63] The 9-copy allele has been associated with a higher level of expression than the 10- or 12-copy VNTR alleles.[64]

The STin2 VNTR 12/12 genotype showed better antidepressant response fluoxetine or sertraline in older Koreans.[32] The combination of 12/12 genotype and 5-HTTLPR S/S genotype had the highest response with independent effects from each gene as suggested by these variants being in weak linkage disequilibrium (LD; r2 = 0.04, D' = 0.40). Conversely, white non-Hispanic subjects with 12/12 genotype were found to have a remission rate of 44.1% as compared with 51.9% of remaining genotypes. Further exploratory analysis with multifocus effects indicated that carriers of the S-12 haplotype had the lowest probability of remission when treated with citalopram in both dominant model (p = 0.029) and additive model (p = 0.023). Other investigations in older adults found no association with antidepressant response to fluvoxamine[65] or fluvoxamine-induced nausea.[56]