Neuroprotective Benefits of Antidepressants in Multiple Sclerosis: Are We Missing the Mark?

Multiple sclerosis (MS) is the most common neurodegenerative disease to affect young adults (1, 2), with onset typically occurring between the ages of 20 and 40 years (3). Depression and anxiety disorders are frequently present in people with MS (PwMS), and depression is considered a core symptom of the disease (4, 5). The lifetime prevalence of depression is approximately 50% following a diagnosis of MS (6), in contrast to 11.6% for the general Australian population (7). The rate of depression in MS is higher than most other chronic or neurologic diseases (8) and is thought to be driven by a combination of psychosocial factors consequential to the disease and biological changes that result from neurological damage (5). Depression in PwMS has been shown to be more enduring than in people without MS, who usually recover within 6 to 12 months (60% and 80%, respectively) (9). Studies have shown that PwMS and depression at baseline continued to experience depression at 4 and 10 years follow-up (9, 10). There has been less research into anxiety disorders; however, prevalence of clinically significant anxiety has been reported as between 25% and 51% (11–14).

Consistent with the higher rate of psychological disorder experienced by PwMS, compared with another chronic health condition and the general population, there is a greater prevalence of antidepressant medication use, which is reported to be between 21% and 35% in PwMS (15–17), whereas population-wide usage is around 8%−10% (18, 19). However, a study that reported 51% of PwMS had clinically significant depressive symptoms (32% were in the moderate to severe range) found only 41% of PwMS identified as having depressive symptoms were taking antidepressant medication, suggesting underdetection and treatment of depression in PwMS (16).

Stress, Depression, and Progression In MS

Stress is an important biological function that in small amounts is adaptive and has the ability to help motivate us to achieve goals. Sustained or uncontrollable stress, however, may lead to depression, especially in individuals who are predisposed (20). Unsurprisingly, a review of studies that have assessed the relationship between stress and depression in PwMS has found the link to be robust, with a proposed moderate to large effect (5).

The relationship between stress and disease progression has gained considerable research attention because of the long-considered view among many PwMS that significant stressful life events are involved in the onset of disease and occurrence of relapses, as well as the transient worsening of existing symptoms. Animal studies have shown a relationship between induced stress and disease onset and progression in mice infected with experimental autoimmune encephalomyelitis, the mouse model of MS (21). However, the relationship is complex and dependent on factors including gender, timing, duration, and type of induced stress. Additionally, severe stress has been shown to improve disease, whereas milder stress progressed it (22). An association dependent on the severity of a stressful relationship has also been shown in a human study. Brown et al. (23) recruited 101 consecutive PwMS through two MS clinics and followed them over a 2-year period. They reported acute stress, measured as stressful events of less than 6 months, to be related to exacerbation, whereas chronic stress, measured as stressful events of more than 6 months duration, was not. A review of research investigating the relationship between stress and disease onset or progression deduced that there is evidence of stress influencing disease onset and clinical course, although “no secure conclusions could be drawn” (24). There have been contrasting results from studies about the relationship between stress and exacerbation; however, the majority have found a relationship (for a review, see Mitsonis et al. [25]), which was reported to be moderate in a meta-analysis that included results of 14 studies (effect size, Cohen’s d=0.53) (26). Clearly, the relationship between stress and disease onset, progression, or exacerbation is not straightforward and is likely mediated or moderated by a range of psychosocial factors, including depression and anxiety (25, 27).

Research suggests there is also a relationship between depression and relapse/progression in MS. A study of 132 PwMS reported a correlation between higher depressive symptoms and exacerbation, with depressive symptoms remaining 6 months after the MS relapse (28). Furthermore, a retrospective assessment of 2,312 PwMS found that women (but not men) who had experienced a psychiatric illness (depression, anxiety, or bipolar) within a 10-year period had significantly higher scores on the Expanded Disability Status Scale than those who had not (β=0.31, p=0.0004) (29), which is consistent with an earlier animal study that found a greater stress-progression response in female mice infected with experimental autoimmune encephalomyelitis (30). However, depression was not related to greater disability progression at 10-year follow-up in another smaller study (N=96), despite depression remaining stable in people assessed at baseline (10). It is possible this study was underpowered, although the difference in psychiatric measurements may also be responsible for the contrasting results. Koch et al. (10) recruited through an MS clinic and used the Center for Epidemiological Studies Depression scale (CES-D), with a score of 16 or greater determining depressive symptoms in PwMS. The CES-D ranges between 0 and 60. By comparison, McKay et al. (29) identified psychiatric disorder from physician and hospital claim files (≥ five physician claims or ≥ one hospital claim), suggesting PwMS in this sample had a greater severity of psychiatric difficulties, given that they were seeking health care management in relation to them.

Similarities Of Dysregulation Occur In MS, Stress, and Depression

The experience of stress activates the hypothalamo-pituitary-adrenal (HPA) axis through the secretion of corticotrophin releasing hormone from the hypothalamus, which activates the pituitary gland and secretes adrenal corticotrophin, stimulating cortisol production from the adrenal gland; among other things, it helps regulate stress and immune response. Cortisol receptors provide a negative feedback loop to regulate cortisol production and the cortisol-to-cortisone ratio (20, 31). Production of cortisol is regulated by circadian rhythms, with highest levels in the morning and a reduction of cortisol throughout the day. Regulation of cortisol is important for both physical and psychological health. However, the natural rhythm of cortisol production is altered with HPA-axis disruption (31). HPA-axis dysregulation and inflammation have been related to psychiatric pathology (32). Increased cortisol levels, modulated by the HPA axis, have been associated with acute and chronic stress and depression in people with and without MS (33, 34). Dysregulation of the HPA axis is well established in PwMS; hyperactivity correlates with the clinical course of MS, such that mild hyperactivity is seen in relapsing remitting MS (RRMS), whereas primary-progressive MS (PPMS) shows the most pronounced changes. It is proposed that HPA-axis dysregulation may be driven by inflammation in RRMS but by neurodegeneration in secondary progressive MS (SPMS) (35–37). A study that assessed cortisol awakening response (CAR) in people with RRMS and SPMS in comparison to healthy controls found that CAR was significantly higher in people with RRMS than the control group, but not people with SPMS. Additionally, in people with RRMS, increased CAR at baseline was predictive of greater disability 9 months later (38). Prior studies have also found cortisol hyperresponsiveness and HPA-axis dysregulation predicted clinically measured disease progression in PwMS (39). It is possible that glucocorticoid-insensitive immune cells may be responsible for excessive inflammatory response seen in RRMS, whereas glucocorticoid-insensitivity is absent or reduced in people with chronic disease types SPMS and PPMS (39). Taken together, the reregulation of the HPA-axis may be time critical for people with RRMS, before irreversible neuronal damage that causes dysregulation in the absence of inflammation. Importantly, treatment with antidepressant medication has shown attenuation of HPA-axis function in depression (40), and HPA-axis hyperactivity in MS is attenuated by immunomodulatory medications (35).

There is a well-established association between immune system inflammation and severity of depression (41, 42). The MS disease-modifying medication fingolimod, which has anti-inflammatory effects, has recently displayed antidepressant effects, including hippocampal neurogenesis in mice exposed to either chronic unpredictable stress or depressive symptoms induced by chronically treated corticosterone (43). Fingolimod increases brain-derived neurotropic factors (BDNF). BDNF support neurons and glial cells within the central nervous system and are important to neuronal survival and plasticity. BDNF are expressed in immune system cells and have been linked to inflammation. Other immunomodulatory therapy used in MS, including glatiramer acetate and hematopoietic stem cell therapy, have been shown to modulate BDNF. BDNF are similarly reduced in people experiencing depression, and fluoxetine has been found to increase BDNF in cultured mouse astrocytes (44–46).

A number of neurotransmitters have been shown to be altered in both MS and depression. For example, norepinephrine, known for its anti-inflammatory and neuroprotective effects, has been found to be reduced in PwMS (47, 48), as have serotonin transporters and platelet content (49, 50). Excessive glutamine, known to have neurotoxic effects, has been shown in PwMS (51). Serotonin or norepinephrine depletion has long been considered a factor in depression, and this is supported by antidepressant medications that act to increase the synaptic availability of these neurotransmitters to treat depression (52, 53). Glutamate has similarly been implicated in depression, and there is evidence that chronic antidepressant administration reduces the presynaptic release of glutamate (54). These commonalities between neurotransmitter disturbance in MS and depression may point to a potential overlap in disease mechanisms.

What We Know About Antidepressants and Neuroprotection In MS

Animal Studies

The importance of the serotonergic system in immune modulation has been established in animal models with autoimmune encephalomyelitis, where knocking out the serotonin transporters resulted in increased susceptibility to inflammatory response (55). Serotonergic transmission has been shown to be altered in PwMS, with fewer seretonergic transporters (SERT) in the limbic and paralimbic systems and more SERT in the frontal cortex (50). Studies have shown benefits of the antidepressant medications sertraline and fluoxetine in suppressing inflammatory responses and reducing disease onset and progression in mice with autoimmune encephalomyelitis (56–59). In addition, the MAO inhibitor phenelzine reduced progression of autoimmune encephalomyelitis and showed increased levels of serotonin, norepinephrine, dopamine, and GABA (60), whereas the tricyclic clomipramine was found to reduce T-cell proliferation, oxidative stress, and B-lymphocytic activity in acute and chronic phases of autoimmune encephalomyelitis (61). Although it would be ideal to assess whether neuroprotective effects of antidepressants in mouse model autoimmune encephalomyelitis show a difference in effect size dependent on level of stress or depression-like symptoms, this has not been possible, because the motor impairment produced by autoimmune encephalomyelitis precludes assessment of animal depressive-like behavior (43).

Human Studies

A summary of human studies is provided in Table 1. Most human studies have been small. A double-blind placebo-controlled study (N=38, n=19 selective serotonin reuptake inhibitor [SSRI]/placebo) that assessed the effect of fluoxetine on new MRI lesion formation in people who had RRMS or SPMS with relapses found a reduced frequency of new lesions in the group administered fluoxetine compared with the control group (62). Additionally, a study assessing the effect of stress as a risk for relapse in people with RRMS taking an SSRI, escitalopram, compared with a control group, found that the control group had a 2.9 times greater chance of relapse; this was predicted by long-term but not short-term stress-related events (63). Another very small study (N=11) showed a neuroprotective effect from short-term use of fluoxetine through an increased N-acetylaspartate/creatine ratio of cerebral white matter, proposing a reversible effect of axonal damage (64). In contrast, a study undertaken by Mostert et al. (65) that assessed fluoxetine in people with SPMS (N=42, n=20 SSRI, n=22 placebo) found no significant difference in treatment groups on disability progression. However, the authors noted that disability progression in the placebo group was slower in their study than other studies assessing progressive MS using similar measures, and they proposed that the study was underpowered (65). An important difference between the studies that have shown an effect and the Mostert et al. (65) study is that the latter included only people with SPMS. The Mostert et al. (62) study required PwMS to have evidence of relapse or gadolinium-enhancing lesion on MRI within the prior 1–2 years, whereas the Mostert et al. (64) study included people with RRMS (N=7) and SPMS (N=4). Results from these small studies point to the possibility of inflammation as the important aspect in the disease-modulating ability of antidepressant medications in MS.

This is further supported by studies that have found an effect of fluoxetine on recovery following ischemic stroke (66). Studies have also shown untreated depression following stroke is related to greater impairment at follow-up, including cognitive impairment and greater difficulty performing activities of daily living, as well as increased mortality. Similar to MS, there is a high prevalence of depression following stroke; this is thought to be a combination of reactive and acute neurological inflammation factors (67).

A recent evaluation of potential neuroprotective oral antidepressant medications suggested that fluoxetine should be assessed for possible repurposing in SPMS, highlighting the relevant mechanisms to SPMS (68), including stimulation of glycogenosis and enhancement of brain-derived neurotrophic factors found to be reduced in SPMS. A four-arm trial that assessed three potential medications, including fluoxetine, against placebo (N=445) found no support for a neuroprotective effect in people with SPMS (69). A previous study similarly assessing the neuroprotective potential of fluoxetine against placebo in people with SPMS or PPMS (N=108) reported preliminary results at ECTRIMS 2016 also reported no significant results between groups for clinical disease progression, although there was a trend toward significance, p=0.07 (70). Both of these studies excluded people experiencing symptoms of depression (70, 71). MRI results from these trials are yet to be published.

What Are We Missing?

An important consideration for interpreting human studies that assess the relationship between antidepressant administration and exacerbation or progression in PwMS is that they excluded people with moderate to severe depressive symptoms. Excluding people with depressive symptoms is problematic, given that depression is a symptom of MS with a biological basis beyond that of reactive depression (5). The inclusion of people with depressive symptoms in a double-blind randomized-controlled study provides a layer of complexity because, once identified, the randomization of people experiencing moderate to severe depressive symptoms to the placebo arm raises ethical concerns. However, there is precedent for this practice when assessing antidepressant efficacy (72). Alternatively, a single-blind randomized control trial would provide participants in the placebo arm with autonomy to discontinue the trial if they chose to seek treatment for depressive symptoms, while still blinding study outcome evaluation.

Furthermore, studies that are limited to PPMS or SPMS, both characterized with ongoing progression of disability and brain atrophy in the absence of, or with minimal, inflammation, may limit the ability to detect an effect or require a larger sample to detect a smaller effect. Inflammation is a common mechanism associated with both MS and depression (41, 42) and is therefore a potential common mechanism by which antidepressant medication may exert an effect to reduce progression. Similarly, a shared disruption to neurotransmitters, including serotonin, norepinephrine, and glutamate, as well as to BDNF and HPA-axis activity in PwMS and people with depression supports the inclusion of PwMS and comorbid depression (35, 40, 50, 53, 73), especially given the ability of antidepressant medication to reregulate BDNF levels and HPA-axis activity in people experiencing depression (40, 74). Thus, if there is a link between depression and disease progression in MS, exclusion of participants with comorbid depression may be considered a systematic sampling error, because it risks not finding a neuroprotective relationship from SSRI use (or any antidepressant in general) when there is one, if results are generalized to the broader MS population.

Although there is a high priority to find neuroprotective agents for SPMS and PPMS given the dearth of therapeutics currently available for these subtypes, it is remiss to exclude people with RRMS, who often still experience neurological damage with current therapeutics. These people may be granted a greater level of neuroprotection, and therefore quality of life, if the addition of antidepressant medication offers neuroprotective benefits.

Conclusions

Given the high prevalence of depression and other emotional disorders in PwMS, and the similarities of neurotransmitter, BDNF, and HPA-axis dysregulation in PwMS and people experiencing stress and depression, it is vital that studies assessing the neuroprotective benefits of antidepressant medication in PwMS include people who are also experiencing depressive symptoms, and people with RRMS subtype, where inflammation is a key disease component. Should current research assessing the neuroprotective benefits of antidepressants in MS miss a signal because the study design excludes the very PwMS who would have shown a neuroprotective benefit, it risks the MS research community losing enthusiasm for this line of research. Thus, a potential opportunity to improve the quality of life for people living with MS and to reduce health care expenditure resulting from greater physical and psychological disability progression will also be lost.