Memory is a complex neurocognitive function, important in humans to establish an understanding of the past, present, and future (Kandel, 2006). In Squire’s widely accepted neuropsychological model of human memory, memory is divided into two subgroups: declarative and non-declarative memory. Declarative memory represents our conscious recall and is further divided into two subtypes, semantic and episodic memory (Squire, 1992). Episodic memory stores autobiographic events and semantic memory stores facts, concepts, and our knowledge (Tulving, 1985; Kandel, 2013). Collectively, these defined memory types refer to the past, and recollection of pieces of information from the past, further generalized under the term retrospective memory.

For remembering our intentions after a delay, we use prospective memory (PM). PM is essential for our understanding and sense of the future and refers to the function that enables a person to carry out a planned act after a delay (Burgess et al., 2001). Remembering to remember is the basic definition of PM and it can be divided into time- and event-based sub-types, and can also be triggered by cues (Baddeley et al., 2009; Burgess et al., 2011; Raskin et al., 2012). Remembering to call a friend at a certain time is an example of a time-based PM. Whereas if a friend asks you to give them a call when dinner is ready, remembering and calling the friend is an example of an event-based PM. Forgetting to call your friend at that required time, but remembering that you need to call them after seeing someone on the phone is an example for how PM can be triggered by a cue. One of the age-related memory impairments is PM failure, or forgetting intentions (Einstein et al., 2000). Dementia, Parkinson’s disease, Alzheimer’s disease, and schizophrenia also show PM impairments (Huppert and Beardsall, 1993; Jones et al., 2006; Woods et al., 2007; Kliegel et al., 2011). To date, little is known of the genetic factors which contribute to PM and its impairment, but as genetic variants have been shown to be associated with performance in a variety of memory subtypes, they are likely to also contribute to PM.

The cyclic adenosine monophosphate (cAMP) pathway has been shown to play an essential role in memory and cognitive abilities (Kandel, 2012). The cAMP responsive element binding protein 1 (CREB1) gene encodes a transcription factor that is a member of the leucine zipper family of DNA binding proteins (Quinn and Granner, 1990; Taylor et al., 1990). Studies have demonstrated the importance of CREB family proteins in learning and memory (Mizuno et al., 2002). CREB1 is a member of a family of proteins that function as transcription factors expressed in the brain (Kandel, 2012) with CREB signaling central to spatial, associative, emotional, and social memories in mammals (Gass et al., 1998; Pittenger et al., 2002; Viosca et al., 2009). Studies in mice and rats have described the role of CREB as a universal modulator of memory formation (Lazary et al., 2011). In humans, cognitive disorders and neurodegeneration disease studies on Huntington’s disease, Alzheimer’s disease, Rubinstein–Taybi syndrome, and Coffin–Lowry syndrome, support a key role for CREB signaling (Sweatt, 2009). Furthermore, CREB has been shown to play an essential role in synaptic plasticity, a feature of the pathways of memory formation (Kanehisa, 2002; Mantamadiotis et al., 2002; Saura and Valero, 2011).

Several studies focusing on depression, schizophrenia, rumination, and negative emotionally related memory impairments have found significant associations with CREB regulation (Kawanishi et al., 1999; Barrot et al., 2002; Lazary et al., 2011; Guo et al., 2014). Genetic polymorphisms can influence the expression of linked loci. Lazary et al. (2011) reported the involvement of rumination in memory and identified significant associations between the GG genotype of rs2253206 in CREB1 and rumination and negative emotionality in two independent Caucasian cohorts (n = 651 and n = 1174). The rs2253206 polymorphism is located in the promoter region of CREB1 gene (Sands and Palmer, 2008; Rankinen et al., 2010). Furthermore, Rankinen et al. (2010) revealed that the CREB1 polymorphism rs2253206 major G allele is associated with significantly lower promoter activity than the minor A allele. Given these results we hypothesized that, as CREB1 plays an important role in cognition and memory, polymorphisms that influence CREB1 expression may impact on memory performances of individuals.

In the present work, we conducted an association study focusing on the rs2253206 polymorphism in CREB1 and PM performance. We investigated six different aspects of PM examined by three different memory tests in a quantitative manner, and report significant associations for some of these test scores with the CREB1 rs2253206 polymorphism in a healthy cohort of individuals.

In this study, PM performance was measured using three memory tests widely used to assess memory impairments (Raskin et al., 2012): PRMQ, CAPM, and MIST. PRMQ and CAPM are self-report questionnaires. PRMQ evaluates RM and PM with equal number of questions, but in this study only the aspect that assesses PM was used, which focuses on common everyday-life memory failures such as forgetting to take a pill (Smith et al., 2000). CAPM assesses two different PM failures; basic daily routine failures such as forgetting to eat a meal or forgetting to put on a piece of clothing (e.g., socks) are measured using BADL, and instrumental activity related failures such as leaving the stove on or having trouble remembering personal dates at the right time (e.g., birthdays) are measured using IADL (Chau et al., 2007). The MIST is a one-on-one task-based test measuring time- and event-based planned acts with and without cues, using eight different PM tasks while solving a word-search puzzle as a distraction (Kamat et al., 2014). Examples of MIST for PM assessment include: “In exactly 15 min please tell me it is time to take a break” for time based; and “When I hand you a red pen, please sign your name on your paper” for event based with a cue. MIST also includes a delay test in which participants were asked to send an email to the examiner at a certain time, answering a specific question asked during the test (Woods et al., 2008; Kamat et al., 2014). We have studied these memory measurements independently due to the distinct elements of each assessment.

In this study, we used a non-pathological population-based sample of individuals to investigate the effect of the rs2253206 CREB1 gene polymorphism on PM. After multiple testing corrections, we found that the minor A-allele at rs2253206 was significantly associated with better PM performance in an additive model of inheritance and the homozygous AA genotype was nominally associated in a recessive model of inheritance. In contrast, the homozygous major G allele was shown to be correlated with poorer PM and IQ scores (Tables 2, 3). The difference between the memory test scores for the three genotypes (AA, AG, and GG) are small and this is likely to due to the fact that the cohort is comprised of healthy, and mainly young, individuals. Nevertheless, small increments in the scores obtained in the PRMQ, CAPM, and MIST can reflect a large impact on memory performance and daily living. While these tests are commonly used to assess elderly individuals or those that might be affected by Alzheimer’s disease, Parkinson’s disease, schizophrenia, etc., our results suggest that genetic factors may contribute to performance of PM tasks in healthy individuals.

Two of the memory measures used in this study, PRMQ-PM and CAPM-BADL, are very similar in the questions asked, yet they are not highly correlated due to the different construction of the tests. Neither test was found to be significantly associated with rs2253206 in the study cohort. MIST-IMD, which evaluates PM behavioral performance with respect to immediate memory, was nominally associated with rs2253206, while MIST-Delay which measures another aspect of PM encompassing long-term memory, did not show any association. We also found that MIST-IMD and MIST-Delay test scores were not correlated with each other. All tests assess measures of PM, but lack of correlation between test scores suggest that it is a heterogeneous construct.

Several studies have demonstrated that CREB is essential for long-term memory and synaptic plasticity (Impey et al., 1996; Lamprecht et al., 1997; Josselyn and Nguyen, 2005; Alberini, 2009; Wallace et al., 2009). Additionally, Williams et al. (2008) have shown changing expression levels of CREB influences working memory in a rat model. SNPs in the CREB1 gene, including rs2253206, have previously been significantly associated with traits linked to cognitive vulnerability. Guo et al. (2014) investigated the effect of CREB1 gene polymorphisms on cognitive dysfunction in Chinese patients with major depression and their results suggest the CREB1 gene is a promising marker for cognitive function in major depression patients. Juhasz et al. (2011) investigated risk factors for depression, and while studying the neuroplastic pathway, they demonstrated a significant association between the CREB1 rs2253206 polymorphism with rumination and severe depression, reporting that minor allele carriers are less likely to ruminate. These results suggest that CREB1 rs2253206 is, or is linked with, a functional SNP. However, we were unable to find evidence that CREB1 is an eQTL of rs2253206 using the GTEx database, although it suggested that the nearby genes METTL21A and CCNYL1 were in some non-neuronal tissues. Little is known of the function of these genes. Furthermore, we could not detect any other polymorphisms in high LD with rs2253206 that suggested a role in CREB1 expression or function.

Functional analysis of the rs2253206 SNP by Rankinen et al. (2010) demonstrated that the major G-allele has lower promoter activity, likely due to loss of a CCAAT enhancer-binding protein alpha (C/EBPα) transcription factor binding site. Therefore, the effect of rs2253206 on cognitive function observed in our cohort may occur via altered expression levels of CREB1. While analysis of CREB1 expression using GTEx did not show high levels in brain-related tissues, CREB1 transcripts were present in all neural tissues and highest in cerebellar tissues. Although that part of the brain has not been reported to be particularly associated with PM, Tomlinson et al. (2014) found contribution of the cerebellum in working memory.

A limitation for our study is the small sample size which impacts on the power to find significant associations. Nevertheless, our findings suggest further investigation of the CREB pathway in our memory cohort may be justified, and also that the rs2253206 polymorphism should be tested in a similarly characterized independent cohort. However, with respect to PM few studies have investigated the genetic factors involved and thus have both PM phenotypes and genotyping data available. We have previously reported a significant effect of the APOE 𝜀4 polymorphism in PM using the CAPM test in a smaller subset of the cohort investigated here (Donges et al., 2012). Evans et al. (2014) also reported an association of APOE 𝜀4 genotype with PM using different PM tasks to our study; interestingly they found that genotype affects changed with age of the participants (Evans et al., 2014). Kennedy et al. (2015) investigating the brain-derived neurotrophic factor (BDNF) Val66Met polymorphism (rs6265) also found an adverse effect of age on PM performance across the lifespan which was much stronger in the BDNF Met carriers than for the Val homozygotes. Therefore, it would be of interest to also study the CREB1 rs2253206 polymorphism in an older cohort tested for PM performance, when impairments influenced by genetic factors may become more pronounced.

In summary, we have identified associations between CREB1 rs2253206 and PM in a healthy cohort; one of the memory phenotypes (CAPM-IADL) was identified as being significantly associated and another (MIST-IMD) identified as nominally associated with rs2253206. When we consider its location and evidence for an effect of the rs2253206 polymorphism on CREB1 promoter activity, it is conceivable that genotypic variation may affect PM performance via a regulatory effect on the CREB1 gene, a gene known to be involved in cognition and re-modeling of synaptic plasticity. Our results suggest that further studies of polymorphisms in CREB1 and other genes in the cAMP pathway would be of interest to explore its role in the genetic basis of PM.

NA performed the laboratory work, data analysis, and interpretation. HS and RL supervised the study and interpretation of the results. DS designed the memory performance evaluations. LS participated in phenotype data collection. LH and LG supervised the study. All authors contributed to writing and editing of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.