The objective of pharmacological treatment in osteoporosis is to reduce the fracture risk, which is related to an increase in BMD, or at least in a delay in BMD loss [16]. Osteoporosis treatments can be divided in drugs that increase bone formation by stimulating osteoblast activity and drugs that decrease bone resorption by inhibiting osteoclast activity [25]
Alendronate is the main bisphosphonate used as the first-line treatment for patients with osteoporosis, as it has been shown to increase BMD and to reduce the fracture risk [26]. However, the causes responsible for non-response to this drug are unknown. A meta-analysis conducted by Sebba (2008) found that the frequencies of an inadequate response to alendronate, risedronate, and ibandronate in osteoporosis ranged from 8–25%. Francis (2004) showed a frequency of inadequate response to bisphosphonates of 15% in osteoporosis patients. However, this inadequate response may be more frequent in clinical practice. Watts et al. (2009) demonstrated an inadequate response to bisphosphonates ranging between 10% and 50%. However, these authors were unable to identify any risk factor for the inappropriate response upon clinical examination.
Genetic factors account for up to 85% of BMD variability and they are not currently considered when deciding on a pharmacological treatment. There is evidence that SNPs are relevant genetic factors involved in the pathophysiology of osteoporosis [29]. Morris et al. (2018) performed a meta-analysis of genome-wide association studies (GWAS) and identified variations in more than 200 genes involved in bone metabolism and associated with lower BMD and fractures.
Marozik et al. (2019), in a cohort study, found that osteoporosis patients who did not respond to bisphosphonate treatment (40%) had a higher proportion of gene variants compared to patients showing a response to this drug. This result is similar to the present study, where we found an inadequate response to alendronate in 32% of patients.
The generation of profiles from gene variants can contribute to detecting genes involved in the pathophysiology of a disease and to detecting different responses to pharmacological therapies [5]. Hopwood et al. (2009) generated different expression profiles associated with the activation of genes (150) during the bone repair process in patients with osteoporosis who had a femur fracture; the main genes involved were related to the activation and maturation of osteoclasts. Marozik et al. (2019) performed a cohort study in which they looked for an association between response to treatment and combinations of allelic variants of markers associated with the disease, and found that carriers of one combination were predisposed to a negative response to bisphosphonate therapy and a different combination was overexpressed in responders. However, in our study, the analysis of the alleles considered to be risk factors for an inadequate response to alendronate treatment did not show these associations. By generating profiles from a combination of risk alleles, we found a significant difference in patients with profile 2, who all responded to treatment with alendronate (Table 4). In the present study, none of the alleles were considered to be risk factors for osteoporosis and showed an independent association with the therapeutic response to alendronate, supporting the notion of an additive effect in the interaction of these alleles.
In the present study, profile 2 was formed by the SNPs rs700518 of the CYP19 gene, rs1800795 of the IL-6 gene, and rs2073618 and rs3102735 of the OPG gene. The reported frequencies coincide with those published in the literature for the SNPs rs1800795 [31, 32] and rs2073618 [33] in the Mexican population, while rs3102735 [34] was consistent with a Caucasian population. In addition, they were found in HWE (Table 2).
The CYP19 variant generates a change in an amino acid that decreases the activity of the enzyme [35]. The IL-6 variant is associated with an increase in its levels; it is known that this cytokine acts as a pro-resorptive factor by favoring the release of RANKL and increasing osteoclastogenic activity [36, 37]. In addition to this, OPG variants are associated with a lower activity of the protein, which decreases its activity by acting as a decoy and preventing the binding of RANKL to its receptor in osteoclasts. This generates more significant maturation and activation of osteoclasts, so that bone resorption is increased [38, 39]. These effects alter the balance between bone formation and resorption, decreasing BMD so that people with this profile have a greater susceptibility to osteoporosis and a higher risk of fracture. Patients with this profile likely have an imbalance in bone metabolism and more significant osteoclastogenic activity, so receiving antiresorptive drugs such as alendronate would act by inhibiting osteoclasts, allowing for an increase in BMD.
Although the frequencies of these risk alleles are representative of studies reported in the literature for the Mexican population, by stratifying the sample by profile, the number of patients per group decreased considerably, which made it difficult to establish an OR (Table 5).
There are models that help predict responses to treatment, but they only take lifestyle, BMD and anthropometric characteristics into account. By not taking genetic factors into account, the predictive ability is low. A better evaluation should take gene-gene and gene-environment interactions into account [29].
The present study evidence the environmental factors did not significantly modify the response to treatment, so these data support the hypothesis that genetic factors are the primary determinant (Table 1).
Treatment adherence is essential to obtain results, such as an increase in BMD, and to reduce the risk of fracture. Our study clearly shows the importance of considering genetic factors when deciding on the most appropriate treatment for patients and thus reducing the frequency of IR.
The response to treatment is determined after receiving treatment for one year, so if it does not work, this represents lost time for the patient, and the finances and the quality of life of the patient may be affected by not receiving adequate treatment. The results of this study may help guide physicians to personalize treatments more suitably.
The main limitation of our study is the small number of samples analyzed and the minimum follow-up period to observe the response to treatment. A larger sample size could help to establish associations with different profiles. Another limitation is small low number of SNPs analyzed, so it is essential to include more genes that act in different bone metabolic pathways. In addition, the pharmacological management of the disease should be included.
An interesting fact is shown in Table 4, with a high frequency of profile 1 (in 68% of patients), which is the same as profile 2 except for the addition of the ESR1 risk allele. It did not show an association with the response to treatment. This can be explained by the critical contribution of ESR1 in the development of the disease, as it is indicated as one of the most highly involved markers. Therefore, its high frequency reflected in the Hardy-Weinberg disequilibrium is expected (Table 2).