Amyotrophic lateral sclerosis (ALS) is a lethal neurodegenerative disease characterized by the destruction of upper and lower motor neurons in the central nervous system (1). Individuals with ALS initially experience muscle weakness with progressive loss in their ability to speak, eat, move, and, eventually, breathe. Death typically occurs 2–4 years from the time of diagnosis. Despite the rapid and ultimately lethal nature of ALS, only a handful of treatments of limited efficacy exist (2), necessitating the development of new treatment options.

The immune system plays a role in ALS pathology in both humans afflicted with the disease and mouse models (3–5). In humans, many immune cell types and pathways in peripheral blood are associated with ALS survival and progression, including monocytes, neutrophils, CD4+ T cells, and natural killer (NK) cells (6–17). Multiple postmortem studies also report corresponding immune infiltration into human ALS spinal cord tissue (10, 18), findings that have been replicated in mouse models (19–21). More recently, we showed that changes in peripheral immune markers occur upstream of clinical changes, reinforcing the idea that the immune system plays an active role in shaping ALS pathology (6). In mice, the role of the immune system in ALS is even more evident, as manipulating specific immune cells can slow or accelerate disease progression (7, 18, 21–27). Together, these data demonstrate that the immune system is a viable therapeutic target in ALS; however, global immune suppression has proven ineffective or accelerated disease, indicating that target-specific immunotherapies against detrimental immune populations and pathways are needed (28–32).

NK cells are a promising target for focused immunotherapy for ALS. These innate immune cells are responsible for destroying cancerous, infected, and damaged cells in the body (33). Typically excluded from the central nervous system, NK cells accumulate in affected regions of the spinal cord and motor cortex in ALS (7, 18), exacerbating neuroinflammation, polarizing microglia to a pro-inflammatory phenotype, and decreasing regulatory T cell (Treg) numbers (18). Moreover, NK cells are cytotoxic to ALS neurons in vitro (34), and their depletion extends the lifespan of ALS mice (7). Tofacitinib, an FDA-approved janus kinase (Jak) and signal transducer and activator (STAT) pathway inhibitor, preferentially inhibits Jak1/3 primarily utilized by NK cells (35). We recently demonstrated that tofacitinib can inhibit NK cell interferon-gamma production, suppress cytotoxicity, and decrease overall NK cell levels in the peripheral blood of mice in a dose-dependent manner (34). These findings suggest that suppression of NK cell function by tofacitinib treatment may slow ALS progression.

The aims of this study were to test the efficacy of tofacitinib in mutant SOD1 ^G93A mice, the most used rodent model of ALS. Both male and female mice were included, as we previously showed sex differences in the immune mechanisms of ALS, particularly with NK cells (7). Low (5 mg/kg) and high (30 mg/kg) dosages were examined, as tofacitinib’s effects can be dose-dependent (34). For each dose, three treatment regimens were compared: daily prevention beginning before symptom onset, daily intervention beginning near disease onset, and a cycle beginning before symptom onset consisting of one week on tofacitinib and one week without treatment (36). The cycle regimen was designed to target NK cells while sparing protective Tregs; T cell numbers and function recover more rapidly after tofacitinib is discontinued compared to NK cells, and prolonged tofacitinib treatment disproportionally affects Tregs (37–39). Low-dose, but not high-dose, tofacitinib increased survival and delayed weight loss in ALS mice compared to untreated mice. Notably, the treatment regimen had little effect on survival or symptoms. Peripheral NK cell and CD8+ T cells were reduced dose-dependently. Interestingly, low-dose tofacitinib increased NK cell numbers in the spinal cord, though other sources of neuroinflammation were not increased, and RNA-seq analysis found that tofacitinib reversed changes seen in multiple pathways in ALS spinal cord. Together, these results suggest tofacitinib slows ALS progression via NK cell-independent mechanisms and supports the continued study of tofacitinib and other JAK-STAT inhibitors for treating ALS.

The immune system contributes to ALS progression (3, 4), with NK cells playing a key role (6, 7). We previously demonstrated that tofacitinib, a JAK/STAT inhibitor (45–47), suppresses overall NK cell numbers, cytotoxicity, and cytokine production in vitro, ex vivo, and in vivo (34). Therefore, to determine whether tofacitinib has therapeutic potential for treating ALS, we treated SOD1 ^G93A mice with low-dose (5 mg/kg) or high-dose (30 mg/kg) tofacitinib using multiple dosing regimens. Low-dose tofacitinib treatment significantly prolonged survival and delayed the onset of weight loss, while high-dose treatment had no effect or exacerbated disease. Initiating tofacitinib after symptom onset did not drastically differ from initiating treatment before symptom onset, a particularly promising finding since patients often face years-long delays between ALS disease onset and diagnosis (1). Despite these salutary benefits, tofacitinib did not perceptibly alter motor neuron or NMJ number at disease end-stage, and there was no drug-mediated improvement in grip strength or agility. Tofacitinib decreased peripheral NK cells in a dose-dependent manner, and high-dose tofacitinib also significantly decreased peripheral CD8+ T cells. Interestingly, low-dose tofacitinib increased spinal cord infiltration of NK cells, though no increase in microglial activation or other markers of neuroinflammation were observed. Transcriptomics analysis of spinal cord at end-stage disease found that low-dose tofacitinib reversed several ALS-dysregulated DEGs primarily linked to immune pathways and additionally influenced metabolic pathways unrelated to ALS.

Interestingly, the tofacitinib regimen used did not have a significant effect on the response to treatment. No differences in survival or immune populations and only slight differences in phenotype were observed between mice on prevention or intervention regimens for both low- and high-doses. One potential explanation is that NK cells and CD8+ T cells do not play major protective roles at any point in ALS and, thus, there may be limited advantage with earlier tofacitinib treatment (3, 4, 48). Another possibility is that the use of the intervention regimen beginning at 90 days of age preceded inflammatory changes in ALS, as the immune system shifts from an anti-inflammatory phenotype to a pro-inflammatory phenotype between 90 and 120 days of age in mice (19), after both prevention and intervention regimens were initiated.

Despite an increase in survival and delay in weight-loss onset, the only significant change in immune populations in the spinal cord with tofacitinib treatment was a slight increase in infiltrating NK cells, counter to our hypothesis that infiltrating NK cells are detrimental in ALS. One possible explanation is that NK cell function is more important than NK cell number; we have found that NK cell expression of NKp46, NKp30, and NKG2D, as well as CD62L+ and CD27+ subpopulations, are better predictors of ALS progression than the number of NK cells alone (6). Others have found that tofacitinib has a greater impact on NK cell phenotype, such as cytotoxicity, than on viability (49). This potential decrease in NK cell activity is supported by the lack of subsequent microglial activation, though in vivo NK cell function was not tested in the current study. Indeed, specific NK cell subpopulations may play a role in disease: in our own studies we have found significant dysregulation of NK cells in individuals with ALS Previous studies have shown that CD11b and CD27 expression on human NK cells identify subpopulations with increased ability to differentiate, express cytokines, or lyse target cells depending on their surface expression (50). Moreover, a recent report suggested that specific NK cell subpopulations may contribute to disease progression in individuals with ALS (51). Indeed, at least one study reported that NK cells can drive ALS via the release of IFN-γ inside the CNS which in turn drives activation of resident Immune cells, particularly the microglia (18). These findings suggest that the NK cells accumulating in the CNS of low-dose mice may have altered functionality, particularly reduced release of IFN-γ, though this was not tested in the current study.

The unexpected finding that low-dose tofacitinib extended survival, while high-dose tofacitinib may exacerbate disease, may be due to a number of factors. First, tofacitinib may adversely impact CD4+ T cell function, including Th2 cells and Tregs which have a net protective effect particularly in early stages of disease (12, 26). As ALS progresses, the balance shifts towards Th1 T cells, which drive disease through microglial polarization and Treg inhibition (52, 53). Tofacitinib exerts differing effects on CD4+ T cells based on dose: low-dose tofacitinib impacts effector T cell function, including Th1 polarization, while higher doses may adversely impact the function of protective Tregs (39, 54–56). Thus, a reduction of Treg function early in disease may explain why the high-dose prevention regimen was associated with reduced survival, though T cell function and polarization was not tested in the current study. With regards to NK cells, we found that higher concentrations of tofacitinib were associated with increased cytokine production, though these changes were not statistically significant. Nonetheless, this may explain why mice receiving high-dose regimens saw reduced survival rates. While tofacitinib primarily inhibits Jak1/3 it can affect the activity of other members of the Jak family (57), so it is possible that higher doses adversely impact associated pathways in NK cells, other immune cells, or other cell types such as neurons. Though we observed no off-target effects in NK cells culture in vitro, other cell types were not tested. In addition, mice receiving high doses tofacitinib experienced a trend towards more rapid weight loss. Clearance of the drug, particularly at high doses in female mice, may also have played a role. Our previous pharmacokinetic analyses showed increased maximal concentration in the blood of female mice following drug administration as well reduced drug clearance (34); this may explain why female mice in particular benefited from the use of drug cycles. Other off-target tofacitinib effects, including increased toxicity, likely did not contribute to reduced survival in the high-dose regimens, as tofacitinib is routinely prescribed at higher doses with minimal increase in adverse events (58, 59), and we did not previously observe any signs of toxicity or infection in low-dose or high-dose mice (34). However, it is also unclear whether prolonged use of tofacitinib is toxic to neurons: tofacitinib is somewhat able to cross the blood-brain barrier (60), has shown mixed results in animal models of multiple sclerosis (61, 62), and has caused demyelination in at least one case report (63). Therefore, it is possible that higher doses of tofacitinib may be toxic to motor neurons through heretofore unexplored mechanisms.

Transcriptomics analysis of the spinal cord revealed that tofacitinib reversed pathways dysregulated during ALS. These pathways were primarily related to inflammation, but metabolic pathways were impacted as well. Tofacitinib reversed the most upregulated pathways in ALS, including pathways associated with hematopoietic cell lineage, neutrophil extracellular trap formation, and cytokine-cytokine receptor interactions. Interestingly, tofacitinib treatment also resulted in the downregulation of multiple metabolic pathways that were not dysregulated in untreated ALS mice. These included pathways associated with lipid metabolism for steroid hormone biosynthesis, linoleic acid metabolism, arachidonic acid metabolism, and fatty acid degradation. Tofacitinib’s association with altered lipid metabolism is well-established, primarily observed as increases in LDL and HDL cholesterol in multiple studies of patients on tofacitinib (64, 65). Additionally, tofacitinib has been shown to reduce cachexia in rodent arthritis models (66, 67), and is associated with weight gain in humans (68, 69). This is particularly promising given that ALS has long been associated with metabolism alterations observed in plasma (70, 71), and dysregulation of lipid metabolism is a top predictor of disease progression (72).

There are limitations to this study. Tofacitinib blood levels were not assessed, and differences in the ease and ability for ALS mice to reach the chow may have affected dosing at later stages of disease. We have also previously found sex differences in tofacitinib clearance rates (34). The impact of tofacitinib was not assessed in other mouse models of ALS. SOD1 ^G93A mice, as well as ALS patients harboring SOD1 mutations, do not develop the TDP-43 proteinopathy seen in 95% of ALS patients (73, 74). Thus, additional experiments using alternate mouse models, such as those based on C9orf72 expansions, may be warranted. Mouse models also present issues in both the lack of upper motor neuron disease and the limited amount of tissue. For this reason, motor neurons were counted in the lumbar spinal cord, and the thoracic spinal cord was used for flow cytometry and transcriptomics. Moreover, while specific NK cell subpopulations may contribute to ALS progression (51), NK cell subpopulations were not characterized in this study. In addition, clinical trials could benefit from a better understanding of the role of NK cells and NK cell subpopulations in ALS. Preclinical studies including adoptive transfer of specific subpopulations of NK cells or NK cells pre-treated them with tofacitinib could further elucidate the mechanisms by which low-dose tofacitinib slow disease progression. Finally, it must be acknowledged that few ALS treatments have translated from mouse to man (73, 74). However, it is likely that tofacitinib will be effective in human ALS as well, as similar dysregulation of Jak3 is seen in both mouse and man (75). Tofacitinib preferentially inhibits Jak1/3; however, other jakinibs, such as baricitinib, may also warrant study as they differ slightly in Jak-specificity and blood-brain barrier penetrability (47, 76–78).

There is great potential for tofacitinib as an ALS therapeutic. Further preclinical studies, particularly those in ALS mouse models that develop TDP-43 proteinopathy or studies designed for increased sensitivity for symptom onset and progression, may be helpful in guiding clinical translation. However, these studies face diminishing returns due to the poor predictive power of mouse studies for human translation. Additionally, there is already a significant body of research on tofacitinib in humans, including studies on dose-effect, long-term use, and the impact of tofacitinib on other immune cell types including monocytes, T cells, and B cells, facilitating dosing in individuals with ALS and translation to the clinic (37, 58, 79). Lastly, tofacitinib is not predicted to interact with either riluzole or edaravone, the two most commonly prescribed ALS medications (80, 81). Based on those factors and the findings presented here, tofacitinib is a promising therapeutic candidate for ALS and is ready for translation into clinical trials.