Some epidemiological studies emphasize the significant role of female reproductive hormones and their receptors in the potential connection between breast cancer (BC) and migraine (Mi). Specific patterns of hormonal variations are associated with both BC and Mi. Migraine with aura (MwA) is associated with the peak of estrogen concentration, while migraine without aura (MwoA) is associated with rapid drops in estrogen levels during the hormonal cycle ^[1][27]. On the other hand, BC is linked to high cumulative exposure to estrogens ^[2][28].

Notable differences have been highlighted between MwA and MwoA, regarding the posibility of experiencing headaches during menstruation, with a higher preference for MwoA ^[3][4][7,29]. It has been reported that a drop in circulating estrogen occurring 2–3 days before the onset of menses partially triggers menstrual-associated Mi ^[5][30]. Specifically, it has been found that 60% of women with Mi are more likely to have Mi attacks during peri-menstrual time periods and that there is a strong correlation between the onset of Mi (occurring between days 22 and 27) and the second dramatic drop of estrogens, which occurs in the luteal phase of the menstrual cycle. Similar findings had already been obtained in 1972 by Somerville and colleagues ^[6][31]. They demonstrated that women with Mi, who received a 10 mg intramuscular injection of estradiol, experienced a delayed Mi attack until the estradiol level returned to pre-treatment levels. In line with these data, epidemiologic findings showed that women suffer up to three times more from Mi in the peak of their reproductive years (20–40) compared with men, confirming the role of the hormonal cycle in headache prevalence. In more recent years, a multisite, multiethnic, longitudinal study has shown that female migraineurs are characterized by a faster decline in late luteal phase conjugated urinary estrogens (E1c) than controls ^[7][32]. Specifically, there were significant differences in E1c decline in the late luteal phase, and migraineurs showed a greater rate of E1c decline over the 2 days following the luteal peak than controls (40% vs. 30%; p < 0.001).

In contrast, women with MwA may experience more frequent attacks during estrogen peaks or during pregnancy ^[1][27]. On the other hand, the role of progesterone in Mi attacks appears to be less significant compared to estrogens ^[1][27], except possibly in relation to the intensity of the attack ^[8][9][33,34]. Its action appears to rely on reducing nociceptive inputs at the level of the trigeminal nucleus and dorsal horns directly and through its metabolite, allopregnanolone, which can modulate GABAA (γ-aminobutyric acid type A) receptors ^[10][11][12][35,36,37].

Other hormones, such as vasopressin, prolactin, and orexin, do not have a clear role in Mi ^[1][27]. Therefore, further studies in this field are needed. In addition, sex hormone-binding globulin (SHBG) has been investigated in Mi, demonstrating no association between SHBG and Mi ^[13][38]. However, it has been found that preventive antimigraine treatment with valproate could increase the serum level of SHBG in post-menarchal women ^[14][39].

Over the past 40 years, many studies have demonstrated a predominant role of estrogen exposure, both endogenous and exogenous, in BC development ^[2][28]. Endogenous exposure is closely linked to hormonal cycle patterns. In this context, in two Italian hospital-based case–control studies, the authors found that irregular menstrual patterns have a protective role in BC development ^[15][16][40,41]. Specifically, taking advantage of the large number of women enrolled in these studies (i.e., 1207 and 5606, respectively), the authors demonstrated that irregular menstrual patterns inversely correlate with the risk of both benign breast lesions (RR = 0.6; 95% CI: 0.4–1.0; p = 0.05 ^[15][40]) and BC development (RR = 0.4; 95% CI: 0.3–0.8; p < 0.05; and RR = 0.6; 95% CI: 0.5–0.8; p < 0.05, respectively; ^[15][16][40,41]). Similarly, breastfeeding has been shown to reduce BC risk ^[2][28], possibly by suppressing ovulation and consequently reducing exposure to estrogen. Other endogenous exposure factors that have been demonstrated to impact BC risk include the age at menarche and menopause. As extensively discussed in the 2003 review by Travis and Key ^[2][28], a 1-year delay in the onset of menarche was associated with a 5% reduction in BC risk ^[17][42]. Instead, each 1-year delay in the onset of menopause increased the risk by 3% ^[18][43]. In addition, it has been demonstrated that each birth reduces the BC RR by 7% ^[19][44]. To explain this protective mechanism, it has been suggested that the high serum levels of estrogens and progesterone during pregnancy stimulate the growth of the mammary epithelium and promote the differentiation of epithelial tissue, reducing the number of epithelial structures most vulnerable to epithelial transformations.

On the other hand, exogenous exposure to estrogens comprises the use of oral contraception and hormonal replacement therapy. Data from 54 published studies showed that the current or recent use (in the past 10 years) of oral contraceptives posed a slight increase in the BC RR ^[20][45]. However, further research is needed, as the formulations of oral contraceptives are changing. The use of hormonal replacement therapy showed a 2.3% risk of being diagnosed with BC for each year of use in the span of 1–4 years. Instead, the excess risk of being diagnosed with BC for patients who were on hormonal replacement therapy for 5 years or more was 35% ^[18][43]. Considering the type of hormonal replacement therapy, recent studies showed that using treatments containing estrogens plus progestins for 5 years induced a 26–30% increase in BC risk ^[21][46].

The role of estrogens in BC development was also confirmed by the demonstration that increased concentrations of circulating estrogens were linked to an increased risk of BC development in postmenopausal women. In fact, in a pooled analysis of nine prospective studies, including 663 postmenopausal BC patients, it was shown that the increase in circulating levels of estradiol, free estradiol, and oestrone significantly increased the risk of BC ^[22][47]. Circulating estrogen levels depended in part on the concentrations of SHBG. With high SHBG concentrations, the level of free circulating estrogen was lower; however, in obese postmenopausal women, the levels of SHBG were lower, thus increasing the exposure to free circulating estrogens ^[23][48].

Researchers were not able to exclude the contribution of other hormones, such as progesterone, prolactin, and testosterone, in BC development, since current epidemiological and experimental data suggest a role for these hormones in the etiology of BC. Indeed, breast cell proliferation has been found to be greater during the luteal phase of the menstrual cycle, concomitant with high levels of progesterone ^[24][49]. In addition, high levels of prolactin have also been shown to increase BC risk because of the effect of prolactin in stimulating the proliferation, survival, and motility of mammary epithelial cells ^[25][50]. The conversion of testosterone into estrogen in the breast has been investigated in a pooled analysis of prospective studies, showing an increased BC risk in postmenopausal women with high testosterone levels ^[22][47].

Besides the hormonal cycle, changes in estrogen levels and their consequent effects on susceptibility to Mi and BC could also be due to mutations in estrogen pathways. The cytochrome p450 family 19 gene (CYP19A1), located on chromosome 15, encodes for the enzyme aromatase. Aromatase catalyzes the final step in estrogen biosynthesis and metabolism, converting androgens (androstenedione and testosterone) to estrogens (estrone and estradiol, respectively) ^[26][51]. Therefore, genetic alterations in this gene may affect estrogen synthesis and, in turn, influence the risk of Mi, BC, or both.

In this context, several polymorphisms have been evaluated in association with Mi or BC ^{[27][28][29][30][31]}[52,53,54,55,56]. However, only the CYP19A1 rs4646 (NC_000015.10:51210646:A:C, NC_000015.10:51210646:A:G in the gene CYP19A1) variant has been analyzed in both diseases. Specifically, in 2012 Ghosh and colleagues investigated the role of polymorphisms in CYP19A1 and ER genes in the susceptibility to Mi in the North Indian population by conducting a case–control study and comparing it with other studies in a pooled meta-analysis ^[27][52]. Through genotyping experiments, the authors found that the CYP19A1 rs4646 variant alone or in combination with ESR1 mutations at locus rs9340799 (NC_000006.12:151842245:A:G, in the same gene) conferred a significant protective effect. In the context of BC risk, in 2015, a population-based case–control study conducted by Alanazi and colleagues explored the possible influence of CYP19A1 polymorphisms on BC incidence in Saudi Arabian patients ^[28][53]. The genotyping analyses revealed no statistically relevant correlation between single nucleotide polymorphisms (SNPs) rs4646 and BC risk. Similarly, a previous investigation conducted in 2014 by Boone and colleagues showed an inverse correlation between CYP19A1 rs4646 polymorphism and BC risk; however, after multiple comparison adjustments, no statistical significance was reached ^[29][54].

To be eliminated from the body, estrogens must be converted to inactive metabolites and then excreted in urine/feces. Estrogen metabolism is primarily dependent on cytochromes P450 (CYP450) enzymes (including CYP1A1, CYP1A2, CYP1B1, and CYP3A4), which are responsible for oxidation; UDP-glucuronosyltransferase, which manages glucuronidation; sulfotransferase, which induces sulfation; and catechol O-methyltransferase (COMT), which is responsible for O-methylation ^[32][57]. Hence, polymorphisms in genes related to the estrogen catabolism pathway may theoretically affect estrogen levels and, in turn, impact Mi and/or BC development.

However, despite preliminary in vitro findings suggesting that the mutant COMT-Met isoforms may increase BC risk due to variations in COMT catalytic activity associated with significant differences in the level of catechol estrogens ^[33][58], the role of COMT polymorphisms in BC risk has not yet been established, as results of several studies have been inconclusive or controversial ^[34][35][59,60]. Their investigations hypothesized that carrying these variants would increase BC susceptibility, as the CYP1B1 variants displayed higher catalytic activity than the wild type ^[36][61]. In comparison, the COMT variants showed lower thermal stability and thus lower enzymatic activity, overall increasing the amount of “carcinogenic catechol estrogens”. However, neither the case–control study nor the meta-analysis showed any statistically significant association between BC risk and the COMT Val158Met (i.e., a non-synonymous G→A SNP (rs4680, NC_000022.11:19963747:G:A) in exon four leads to a valine (Val) to methionine (Met) peptide change in the mature protein) ^[35][60]. Finally, the lack of correlation between COMT Val158Met variants and BC development was confirmed in a large meta-analysis involving a total of 56 case–control observational studies conducted in 2012 by Qin and colleagues ^[37][62].

Similarly, no statistically significant association was found between COMT polymorphisms and Mi susceptibility. In fact, in a 2012 systematic review and meta-analysis regarding COMT polymorphisms and chronic pain ^[38][63], the authors did not find any correlation between the analyzed COMT SNPs and Mi. Specifically, the Val158Met polymorphism (rs4680, NC_000022.11:19963747:G:A), which has previously been shown to produce an enzyme with lower thermostability, thus causing a decrease in enzyme activity, was investigated in correlation with Mi (with and without aura). None of the three studies included in the analysis found any correlation between the aforementioned SNP and Mi. This lack of correlation was also observed in a case–control study exploring the possible association between four functional polymorphisms involved in estrogen metabolism and menstrual Mi in a UK population ^[39][64].

Besides COMT SNPs, CYP1B1 polymorphisms have also been extensively investigated in association with BC ^{[34][35][40][41]}[59,60,65,66]. The hypothesis behind these investigations is that CYP1B1 may play a key role in breast and endometrial carcinogenesis ^[32][57]. The main activity of CYP1B1 is represented by the 4-hydroxylation of estradiol, which in turn produces free radicals that may induce DNA damage. In addition, abundant levels of CYP1B1 have been found in the so-called “estrogen target tissues”, which are mammary, ovarian, and uterine tissues, as well as in tumor tissues, thus suggesting that the specific and local 4-hydroxylation of estradiol could enhance carcinogenesis in these tissues ^[32][57]. However, the results of two meta-analyses highlighted the possible race-specific effects of CYP1B1 polymorphisms ^[40][41][65,66]. Specifically, in the meta-analysis by Paracchini and colleagues, a lack of association between the CYP1B1 Val432Leu (i.e., C→G transversion at position 1666 in exon 3, resulting in an amino acid substitution of leucine (Leu) with valine (Val) at codon 432; rs1056836, NC_000002.12:38071059:G:C) polymorphism and BC was observed in Asian individuals, whereas in populations of mixed/African origin, a negative correlation emerged (OR = 0.8; 95% CI: 0.7–0.9; p < 0.05). In addition, the pooled analysis showed a possible association in Caucasians (OR = 1.5; 95% CI: 1.1–2.1; p = 0.05), but this was age-dependent, being higher for the middle age classes (45–59 years) and lower among older and younger women. Unfortunately, no data on CYP1B1 polymorphisms and Mi susceptibility have emerged so far.

The ESR1 gene is located on chromosome 6 and encodes for ER and ligand-activated transcription factor. The receptor plays a key role in BC, endometrial cancer, and osteoporosis. At the same time, the encoded protein has been found to play a role in growth, metabolism, sexual development, gestation, and other reproductive functions, via the regulating the transcription of many estrogen-inducible genes. Besides its well-known role in BC cancer management, the discovery that ESR1 was localized in brain regions considered to be involved in Mi pathogenesis ^[42][67] suggested the role of ER in Mi development as well. The estrogen receptor α (ERα), related to the ESR1 gene, and the estrogen receptor β (ERβ), related to the ESR2 gene, activate a mitogen-activated protein kinase pathway ^[43][44][68,69]. In contrast, a third ER (G protein-coupled estrogen receptor 1, GPER) is coupled with a G-protein pathway ^[45][70]. ERα and GPER are expressed in the brainstem pons, and ERα in the periaqueductal grey ^[46][71]. Moreover, ERα and ERβ are expressed in the cerebral cortex, possibly explaining why estrogens are able to stimulate MwA ^[47][72]. Additionally, all three ERs are expressed in the hypothalamus, which has been found to be a key structure in regulating the recurrence of Mi attacks ^[48][73]. In particular, ERβ is mostly concentrated in the supraoptic area and the paraventricular nucleus ^[49][74]. ERβ-knockout murine models showed an increased level of the calcitonin gene-related peptide (CGRP); therefore, Krause et al. recently suggested that a direct action of estrogens on their receptors was able to alter the CGRP signaling and the recurrence pattern of Mi attacks ^[1][27].

Based on this hypothesis, polymorphisms in ESR1 could represent a key point for understanding the relationship between BC and Mi susceptibility. Among the ESR1 polymorphisms analyzed in correlation with both BC and Mi risk, one of the most investigated was an intronic polymorphism in the ESR1 gene (rs2234693, NG_008493.2:g.190510T>C), also called ESR1 PvuII. In the 2012 case–control study by Ghosh and colleagues, this intronic polymorphism was related to patients who had MwA ^[27][52]. On the contrary, no statistically significant correlations were observed between the ESR1 PvuII polymorphism and the risk of any type of Mi in two different meta-analyses, each including eight studies ^[50][51][75,76].

The ESR1 PvuII polymorphism (rs2234693, NG_008493.2:g.190510T>C) has been extensively evaluated in correlation with BC risk, with controversial results. In a cohort study published in 2008 by González-Zuloeta Ladd and colleagues ^[52][77], the authors found no correlation between the PvuII and BC risk in their Caucasian population. Similarly, a very recent retrospective case–control study showed that this variant was not associated with BC risk in Saudi women ^[53][78]. In contrast, in a previous large population-based cohort study, the authors showed that the PvuII pp genotype, compared with the PP one (CC-genotype), was a 1.5-fold increased BC risk. However, the correlation did not reach the threshold for statistical significance ^[54][79]. Interestingly, the authors also found a statistically significant interaction between PvuII polymorphism and E2 level on BC risk, and the effect was stronger among women with the Pp or pp genotypes. Consistently with these findings, in a large meta-analysis of 25 case–control studies, the authors showed that people with PvuII T > T + T > C or T > T genotypes were at a greater risk of BC than those with the C > C variant; however, regarding T > T polymorphism, the higher risk occurred only in an Asian population ^[55][80]. By contrast, in BC patients in Moskow, a significant association was found between the CC variant of the rs2234693 (NC_000006.12:151842199:T:C, NC_000006.12:151842199:T:G) polymorphism and the risk of BC development ^[56][81].

In the meta-analyses mentioned above by Li and Schürks ^[50][51][75,76], besides PvuII variants, the authors also analyzed ESR1 594 G > A and ESR1 325 C > G polymorphisms associated with Mi. In both studies, ESR1 594 G > A (rs2228480, NC_000006.12:152098959:G:A, NC_000006.12:152098959:G:T) and ESR1 325 C > G (rs2295190, NC_000006.12:152122608:G:T) polymorphisms were associated with Mi susceptibility, in a race-specific manner. In addition, in Schürks’s meta-analysis ^[51][76], both polymorphisms were associated with Mi, regardless of the presence of aura.

Given the BC risk, in 2012, an extensive review regarding ER gene mutations and polymorphisms in disease susceptibility attempted to bring together all the studies on this topic. It was found that the ESR1 rs2228480 polymorphism was the most investigated among ESR1 variants in correlation with BC risk ^[57][82]. Specifically, this variant was associated with BC risk in a considerable number of populations (e.g., Caucasian, Turkish, cohorts with Western European ancestries, Tunisian, and Korean).