Tocilizumab is safe and tolerable and reduces C-reactive protein concentrations in the plasma and cerebrospinal fluid of ALS patients

Abstract

Introduction/Aims:

We tested safety, tolerability, and target engagement of tocilizumab in amyotrophic lateral sclerosis (ALS) patients.

Methods:

Twenty-two participants, whose peripheral blood mononuclear cell (PBMC) gene expression profile reflected high messenger ribonucleic acid (mRNA) expression of inflammatory markers, were randomized 2:1 to 3 tocilizumab or placebo treatments (weeks 0, 4, and 8; 8 mg/kg intravenous). Participants were followed every 4 weeks in a double-blind fashion for 16 weeks and assessed for safety, tolerability, plasma inflammatory markers, and clinical measures. Cerebrospinal fluid (CSF) was collected at baseline and after the third treatment. Participants were genotyped for Asp³⁵⁸Ala polymorphism of the IL6R gene.

Results:

Baseline characteristics, safety, and tolerability were similar between treatment groups. One serious adverse event was reported in the placebo group; no deaths occurred. Mean plasma C-reactive protein (CRP) level decreased by 88% in the tocilizumab group and increased by 4% in the placebo group (−3.0-fold relative change, P<0.001). CSF CRP reduction (−1.8-fold relative change, P=0.01) was associated with IL6R C allele count. No differences in PBMC gene expression or clinical measures were observed between groups.

Discussion:

Tocilizumab treatment was safe and well tolerated. PBMC gene expression profile was inadequate as a predictive or pharmacodynamic biomarker. Treatment reduced CRP levels in plasma and CSF, with CSF effects potentially dependent on IL6R Asp³⁵⁸Ala genotype. IL-6 trans-signaling may mediate a distinct central nervous system response in individuals inheriting the IL6R C allele. These results warrant further study in ALS patients where IL6R genotype and CRP levels may be useful enrichment biomarkers.

INTRODUCTION

Despite the aggressive nature of amyotrophic lateral sclerosis (ALS) and decades of research, no clear nongenetic cause or highly effective treatment has been identified since Charcot’s initial description in 1869.¹ It remains unclear if inflammatory responses have a role in initiating ALS; however, they play a clear role in disease progression.^2,3 Thus, therapeutic approaches targeting inflammatory pathways may prove effective in slowing ALS progression. Although previous clinical trials of immunomodulatory treatments in ALS patients have failed,^4–8 clear target engagement and resulting inflammatory modulation were not demonstrated. Thus, the immune system remains an exciting but unrealized target for ALS therapeutics.

Although ALS patients are reported to have increased expression of several inflammatory cytokines, the interleukin (IL)-6 cytokine pathway is an increasingly relevant therapeutic target.^9,10 Serum and cerebrospinal fluid (CSF) levels of IL-6 are elevated in ALS patients, and serum IL-6 levels in ALS patients negatively correlate with Revised ALS Functional Rating Scale (ALSFRS-R) scores.^11,12 In the central nervous system (CNS), IL-6 is associated with microglia and astrocyte activation.^13–15 A shift in glial phenotype from protective to pro-inflammatory correlates with increased expression of pro-inflammatory cytokines such as IL-6, tumor necrosis factor-alpha (TNF-α), and IL-12, leading to further glial activation, cytotoxicity, and motor neuron cell death.¹⁶ Together, these reports suggest IL-6 signaling in ALS may worsen disease progression, and IL-6 receptor (IL-6R)–blocking therapies could prove beneficial.

In this study, we investigated the safety and tolerability of tocilizumab, a humanized monoclonal antibody antagonist of the IL-6R, in ALS patients with evidence of baseline systemic inflammation. We also assessed the ability of tocilizumab to reduce in vivo indicators of inflammation in such participants and the influence of the IL6R Asp³⁵⁸Ala genotype.

METHODS

Standard Protocol Approvals, Registrations, and Patient Consents

The study was conducted under the Good Clinical Practice Guidance of the International Council for Harmonisation. It was approved by the study sites’ institutional review boards and was registered at ClinicalTrials.gov (NCT02469896). Written informed consent from the participants was obtained before any study activities were conducted.

Trial Design

Participants aged 18 to 75 years, who met the Revised El Escorial criteria for possible ALS, laboratory-supported probable ALS, probable ALS, or definite ALS; who had a slow vital capacity of at least 60% of the predicted value for age, sex, and height; and who had ≤36 months of ALS-related weakness were recruited from the ALS clinics at 5 study sites between 2015 and 2019. During the study, in an effort to aid recruitment, the eligibility criteria for disease duration and slow vital capacity were relaxed to include patients with ALS of any duration and patients with a slow vital capacity of at least 40% of the predicted value for age, sex, and height. Participants were screened for a high-inflammatory peripheral blood mononuclear cell (PBMC) gene expression profile. Participants who showed a 3-fold upregulation of IL6 expression compared to pooled age-matched, healthy controls or a 2-fold upregulation of IL6 expression with a 2-fold upregulation of IL8 or matrix metalloproteinase 1 (MMP1) expression were considered “high inflammatory” and randomized if all other study criteria were met.

Participants were randomized in a 2:1 ratio to either 3 monthly intravenous treatments of tocilizumab (8 mg/kg) or placebo, beginning at the baseline visit and followed in a double-blind fashion every 4 weeks for 16 weeks. The primary outcomes were safety and tolerability. Safety was assessed as the proportion of participants experiencing severe or nonsevere treatment-emergent adverse events, as classified by the Medical Dictionary for Regulatory Activities (MedDRA, version 17.1). Tolerability was defined as completing all infusions of the study intervention and completing all study visits. Secondary measures included PBMC gene expression profiles (IL6, IL8, MMP1); plasma and CSF inflammatory cytokines (IL-6, IL-8, IL-1β, IL-17, TNF-⍺) and C-reactive protein (CRP); CSF-soluble IL-6R (sIL-6R) and soluble glycoprotein 130 (sGP130) concentrations; IL6R Asp³⁵⁸Ala genotype; and standard clinical measures (ALSFRS-R score, slow vital capacity, handheld dynamometry, and grip strength). Analysis of plasma cytokine levels focused on changes from baseline to the mean of postbaseline assessments at weeks 4, 8, 12, and 16. For CSF levels, the analysis focused on the change from baseline to week 8 because CSF was not collected at other visits. The principal measure of target engagement was the proportion of participants who experienced a 2-fold or greater decline between baseline and the geometric mean of all follow-up assessments in at least 2 of the 3 pro-inflammatory genes. Site evaluators were trained to perform the clinical measures via a standardized testing protocol. CSF was collected by an experienced clinician via lumbar puncture before the first dose of the study drug and 2 hours after the third dose.

PBMC Isolation and Gene Expression

Blood was collected into two 8-mL BD Vacutainer Cell Preparation Tubes (sodium citrate). Tubes sat for 30 to 60 minutes and were then centrifuged at 1800 g for 20 minutes in a swinging bucket rotor without braking. Tubes were shipped overnight on cold packs to a central laboratory at Barrow Neurological Institute (Phoenix, AZ). The plasma containing the PBMCs was decanted from the tubes. The PBMCs were collected and processed for ribonucleic acid (RNA) isolation and IL6, IL8, and MMP1 gene expression profiles, as described in the Supplementary Materials.

Concentration of Pro-inflammatory Cytokines, CRP, and sIL-6R in Plasma and CSF

Plasma and CSF samples were collected from participants at all sites using the same standardized protocols, processed immediately, and stored at −80 °C before shipment to the central laboratory. Briefly, CSF samples were ensured as being clear and absent of turbidity or visible blood after collection, were immediately centrifuged at 1750 g for 10 minutes, were aliquoted into cryovials, and frozen at −80 °C. For plasma preparation, blood was collected using K2EDTA tubes and was centrifuged at 1750 g for 10 minutes. Supernatants were aliquoted into cryovials and frozen at −80 °C.

Measurements of CRP were performed in duplicate by standard ELISA using the R&D Systems Human C-Reactive Protein DuoSet ELISA kit (R&D Systems, Minneapolis, MN, catalog #DY1707) according to the manufacturer’s instructions. Measurements for IL-6, IL-8, IL-1β, and TNF-α were performed using a Meso Scale Discovery V-PLEX Proinflammatory kit (catalog #15049). Detailed methods are provided in the Supplementary Materials.

IL-6, sIL-6R, and sGP130 concentrations were measured with commercial ELISA kits (R&D Systems, USA, catalog # Q6000B, lot P186912; catalog # DR600, lots P185493 and P191955; catalog # DGP00, lot P161984, respectively) at Wake Forest School of Medicine. Details of methods are provided in the Supplementary Methods. Of note, IL-6 was measured by two different approaches. While absolute values varied between the approaches, inferences of within-treatment and between-treatment differences were consistent across both methods (see Supplemental Table S1). Results from the V-PLEX assay for IL-6 are presented for consistency with results for other cytokines measured by this method.

Genotyping of IL6R Asp³⁵⁸Ala Variant (rs2228145)

DNA was purified from whole blood using the Qiagen Autopure LS with standard Puregene chemistry. DNA was genotyped using a validated TaqMan assay for the single nucleotide polymorphism rs2228145 in the IL6R gene according to the manufacturer’s instructions (Assay ID: C__16170664_10; IL6R Ala³⁵⁸ variant; Applied Biosystems, Waltham, MA) and read using an allelic discrimination protocol on a Real-Time PCR System (Model 7500; Applied Biosystems, USA). Positive controls for each genotype and 3 no-template controls (blanks) were run for each assay set.

Statistical Analysis

As a prespecified analysis, the proportion of participants who experienced a 2-fold or greater decline between baseline and the geometric mean of all follow-up assessments in at least 2 of the 3 pro-inflammatory genes was compared between tocilizumab and placebo participants using the Fisher exact test.

Additional target engagement measures were analyzed after log-transformation in a shared-baseline, repeated-measures mixed model with fixed terms for visits (5 levels: baseline and weeks 4, 8, 12, and 16), treatment group (2 levels: placebo and 4 mg/kg tocilizumab) × postbaseline visit (4 levels), and random participant-specific visits with unstructured covariance. Linear contrasts were used to estimate treatment-specific mean change from baseline and treatment-dependent differences. Estimates and their 95% confidence bounds are back-transformed for reporting as a treatment-dependent ratio of postbaseline concentration relative to baseline. Treatment differences in relative change from baseline (i.e., differences of differences on the log₂ scale) are reported as relative-fold changes. Measures of clinical progression were analyzed without transformation in shared-baseline, random-slope linear mixed models with fixed effects of time and treatment × time interaction and random participant-specific intercepts and slopes with unstructured covariance.

Differences in the effect of tocilizumab on target engagement measures associated with IL6R genotype were estimated under an additive genetic model by adding the IL6R C allele count, C allele count × visit, and C allele count × treatment × postbaseline visit interaction terms to the repeated-measures target engagement model.

Participants receiving at least 1 infusion of the study drug were included in safety, tolerability, target engagement, and clinical progression analyses. Analyses were performed using SAS (version 9.4; SAS Institute, Cary, NC). Two-tailed P-values <0.05 were declared significant. The primary inference for secondary outcomes is based on a step-down Bonferroni adjustment for multiple comparisons. Some unadjusted, comparison-wise P-values are also reported given the study’s low power and are so noted.

A sample size of 24 was planned to provide 80% power to detect a significant treatment-dependent difference in the proportion of participants who experience a 2-fold or greater decline between baseline and the mean of all follow-up assessments in at least 2 of the 3 pro-inflammatory genes if no more than 5% of placebo participants experienced such a change and at least 30% of tocilizumab participants did so.

Data Availability

Results for the prespecified primary and secondary outcome measures are posted on ClinicalTrials.gov. Other data required to replicate the procedures and results, which are not presented here because of space limitations, will be shared with other investigators upon request.

RESULTS

Baseline Characteristics

Of 51 patients screened with PBMC gene expression profiles, 22 had a high-inflammatory profile and met all other eligibility criteria. Of these 22 study participants, 8 were randomized to placebo, and 14 were randomized to tocilizumab treatment (Figure 1). Treatment groups were well balanced for demographic characteristics and were not significantly different among IL6R genotypes (Table 1 and Supplemental Table S2).

Figure 1.

CONSORT diagram.

Table 1.

Baseline participant characteristics

		Treatment assignment		IL6R genotype*
Variable	Overall No. (%)	Placebo No. (%)	Tocilizumab No. (%)	AA	AC	CC	P-value
No. participants	22	8	14	6 2P, 4T	10 4P, 6T	6 2P, 4T
Sex							0.33‡
Female	6 (27)	3 (38)	3 (21)	2	4	2
Male	16 (73)	5 (62)	11 (79%)	4	6	4
Race							0.55‡
Black	1 (4)	1 (12)	0 (0)	1	0	0
White	21 (96)	7 (88)	14 (100)	5	10	6
Ethnicity							>0.99‡
Hispanic or Latino	1 (4)	1 (12)	0 (0)	0	1	0
Non-Hispanic or Latino	21 (96)	7 (88)	14 (100)	6	9	6
Revised El Escorial							0.36‡
Definite	6 (27)	2 (25)	4 (29)	1	3	2
Probable	6 (27)	3 (38)	3 (21)	3	2	1
Probable, laboratory-supported	7 (32)	2 (25)	5 (36)	0	4	3
Possible	3 (14)	1 (12)	2 (14)	2	1	0
Onset site							>0.99‡
Bulbar	6 (27)	2 (25)	4 (29)	2	3	1
Limb	16 (73)	6 (75)	10 (71)	4	7	5
Riluzole use							0.84‡
No	7 (32)	2 (25)	5 (36)	1	4	2
Yes	15 (68)	6 (75)	9 (64)	5	6	4
Family history of ALS							>0.99‡
No	20 (91)	7 (88)	13 (93)	6	9	2
Yes	2 (9)	1 (12)	1 (7)	0	1	4
Age at baseline, mean±SD (y)	60.9±8.6 (45.2–75.2)	59.2±8.6 (45.2–73.0)	61.9±8.7 (48.7–75.2)	65.1±9.8 (48.7–75.2)	60.7±9.0 (45.2– 74.1)	57.1±5.2 (50.5– 62.6)	0.28§
Years since symptom onset, mean±SD	2.0±1.3 (0.4–4.8)	1.6±0.9 (0.4–2.7)	2.2±1.4 (0.6–4.8)	2.5±1.9 (0.6–4.8)	1.8±1.1 (0.4–4.0)	1.8±0.9 (0.8–2.7)	0.53§
Years since diagnosis, mean±SD	1.2±1.1 (0.2–4.0)	0.9±0.6 (0.4–1.8)	1.4±1.3 (0.2–4.0)	1.6±1.7 (0.3–4.0)	1.2±1.0 (0.2–3.5)	1.0±0.6 (0.2–1.8)	0.60§
ALSFRS-R total score, mean±SD	33.0±6.9 (21.0–43.0)	32.9±7.4 (21.0–41.0)	33.1±6.9 (22.0–43.0)	30.5±5.8 (22–37)	31.7±7.3 (21–41)	37.8±5.6 (27–43)	0.13§
SVC, mean±SD (max %-predicted)	82.2±21.6 (43.1–119)	86.7±25.1 (43.1–113)	79.7±19.9 (43.4–119)	73.3±18.3 (52.6–96.6)	80.7±23.2 (43.1–108)	93.6±20.0 (72.2–119)	0.27§
Grip strength, stronger hand, mean±SD (lb)	45.0±24.2 (5.00–91.5)	42.9±29.7 (5.00–91.5)	46.1±21.5 (12.0–86.7)	32.9±16.5 (18.5–57.0)	41.4±21.4 (5.0–76.5)	63.0±27.7 (20.0–91.5)	0.07§
BMI, mean±SD (kg/m2)	27.7±5.3 (21.7–42.0)	26.7±3.5 (22.5–31.7)	28.3±6.1 (21.7–42.0)	30.9±7.3 (21.9–42.0)	26.4±4.8 (21.7–35.6)	26.8±2.2 (24.1–29.7)	0.22§

Results are expressed as number (%) unless otherwise noted.

^*A = IL6R major allele; C = IL6R coding variant; Asp³⁵⁸Ala, rs2228145, C allele.

^‡Statistical comparison across groups was performed using the Fisher exact test.

^§Statistical comparison performed using one-way analysis of variance.

Abbreviations: ALS, amyotrophic lateral sclerosis; ALSFRS-R, Revised ALS Functional Rating Scale; BMI, body mass index; lb, pound; P, placebo; SVC, slow vital capacity; T, tocilizumab.

Safety and Tolerability

Treatment with tocilizumab was safe. Both adverse events and serious adverse events were rare. One participant in the placebo group experienced a severe adverse event of aspiration pneumonia. Adverse events of special interest included injection-site reactions and other infections. Two such events were seen in the placebo group (1 injection-site infection, 1 aspiration pneumonia) and 1 in the tocilizumab group (Clostridium difficile colitis). Other adverse events were similar between groups (Supplemental Table S3 and ClinicalTrials.gov NCT02469896). No deaths occurred in either group. Tolerability was 75% in the placebo group versus 86% in the tocilizumab group (P=0.60). Two participants in the placebo group withdrew consent (1 began a prohibited medication, 1 because of disease progression). Two participants in the tocilizumab group completed the trial visits but discontinued infusions after adverse events (1 after completing 2 infusions because of C. difficile colitis, 1 after completing 2 infusions because of asymptomatic neutropenia). Both adverse events resolved without sequelae.

Profiles of PBMC Inflammatory Gene Expression and Plasma and CSF Cytokines and Proteins

No participant experienced a 2-fold or greater decline between baseline and the geometric mean of postbaseline assessments in gene expression of at least 2 of the 3 pro-inflammatory genes. There were no significant changes in PBMC IL6, IL8, or MMP1 gene expression between screening and baseline and no treatment-dependent differences in the ratio of postbaseline geometric mean gene expression versus baseline (P≥0.12) (Table 2 and data not shown).

Table 2.

Fold differences for tocilizumab versus placebo in PBMC inflammatory gene expression and plasma and CSF cytokine levels^*

	Total participants (n=22)		AA (n=6)		AC (n=10)		CC (n=6)
	Diff (95% CI)	P-value	Diff (95% CI)	P-value	Diff (95% CI)	P-palue	Diff (95% CI)	P-value
PBMC gene expression
IL6	0.4 (−0.4 to 1.2)	0.28	0.5 (−1.1 to 2.1)	0.53	0.3 (−1.1 to 1.7)	0.63	−0.1 (−1.7 to 1.5)	0.91
IL8	0.4 (−0.9 to 1.8)	0.53	1.5 (−0.5 to 3.4)	0.13	−1.5 (−3.4 to 0.4)	0.12	1.1 (−1.3 to 3.4)	0.36
MMP1	−0.0 (−0.6 to 0.5)	0.88	0.1 (−1.1 to 1.3)	0.85	0.2 (−0.9 to 1.3)	0.67	−0.3 (−1.5 to 0.9)	0.65
Plasma cytokine and protein
CRP	−3.1 (−4.6 to −1.5)	0.003	−3.4 (−6.0 to −0.5)	0.02	−3.5 (−5.9 to −1.0)	0.006	−2.2 (−5.0 to 0.6)	0.12
IL-1β	0.3 (−0.2 to 0.7)	0.25	−0.001 (−0.9 to 0.9)	>0.99	0.2 (−0.6 to 1.0)	0.58	0.5 (−0.4 to 1.4)	0.24
IL-6	3.0 (2.4 to 3.5)	<0.001	2.9 (1.9 to 4.0)	<0.001	3.3 (2.3 to 4.3)	<0.001	2.8 (1.7 to 3.8)	<0.001
sIL-6R	3.4 (2.9 to 3.8)	<0.001	3.6 (2.8 to 4.4)	<0.001	3.3 (2.6 to 4.1)	<0.001	2.8 (1.9 to 3.6)	<0.001
IL-8	0.2 (−0.1 to 0.6)	0.19	0.1 (−0.6 to 0.9)	0.74	0.5 (−0.2 to 1.2)	0.13	−0.1 (−0.9 to 0.8)	0.84
IL-17	0.3 (−0.5 to 1.2)	0.43	0.3 (−1.2 to 1.7)	0.70	−0.1 (−1.4 to 1.2)	0.85	1.1 (−0.3 to 2.5)	0.12
TNF-α	−0.1 (−0.3 to 0.1)	0.29	−0.1 (−0.6 to 0.3)	0.55	−0.1 (−0.4 to 0.3)	0.78	−0.2 (−0.6 to 0.3)	0.47
CSF cytokine and protein†
CRP	−1.8 (−3.1 to −0.4)	0.01	−0.9 (−2.0 to 0.3)	0.15	−1.7 (−2.4 to −1.1)	<0.001	−2.5 (−3.3 to −1.7	<0.001
IL-1β	−0.3 (−1.5 to 0.9)	0.60	−3.9‡ (−5.3 to −2.6)	<0.001	0.7 (0.0 to 1.4)	0.049	−0.6 (−1.5 to 0.3)	0.21
IL-6	1.1 (0.6 to 1.6)	<0.001	0.9 (−0.4 to 2.1)	0.15	1.6 (0.9 to 2.2)	<0.001	0.5 (−0.3 to 1.3)	0.18
sIL-6R	0.6 (0.4 to 0.8)	<0.001	0.7 (0.04 to 1.3)	0.04	0.6 (0.3 to 1.0)	0.002	0.5 (0.2 to 0.8)	0.002
IL-8	0.1 (−0.2 to 0.4)	0.59	−0.1 (−0.7 to 0.6)	0.84	0.3 (−0.1 to 0.7)	0.10	−0.1 (−0.6 to 0.4)	0.65
IL-17	ND		ND		ND		ND
TNF-α	0.1 (−0.31 to 0.5)	0.70	0.1 (−1.0 to −1.2)	0.91	0.3 (−0.3 to 0.9)	0.28	−0.1 (−0.8 to 0.7)	0.82

Data shown are fold difference (tocilizumab vs placebo from baseline to mean of weeks 4–16) and P-value (95% CI). Statistical differences are determined by shared baseline, repeated-measures analysis of variance. Bold indicates statistical significance.

^*A = IL6R major allele; C = IL6R coding variant; Asp³⁵⁸Ala, rs2228145, C allele

^†All CSF cytokine and protein data through week 8.

^‡Significance reflects increased expression in AA placebo-treated patients from baseline to 8 weeks with no change in tocilizumab-treated AA patients.

Abbreviations: CRP, C-reactive protein; CSF, cerebrospinal fluid; Diff, fold difference; IL, interleukin; MMP1, matrix metalloproteinase 1; ND, no difference; PBMC, peripheral blood mononuclear cell; sIL-6R, soluble IL-6 receptor; TNF, tumor necrosis factor.

Plasma CRP was reduced 88% (95% confidence interval [CI], −94% to −76%) among tocilizumab participants from baseline to the mean of all follow-up assessments versus a 4% increase (95% CI, −59% to +259%) among placebo participants (−3.1-fold relative change, 95% CI, −4.6 to −1.5-fold, adjusted P=0.003) (Figure 2A, Table 2). Plasma IL-6 and sIL-6R were dramatically increased following tocilizumab treatment, but not after placebo (Table 2). Plasma IL-6 increased 2.9-fold (95% CI, 2.6 to 3.2-fold) in the tocilizumab group versus a 0.09-fold decline (95% CI, −0.6 to +0.4-fold) in the placebo group (3.0-fold relative change, 95% CI, 2.4 to 3.5-fold, adjusted P<0.001) (Figure 2B). Plasma sIL-6R increased 3.4-fold (95% CI, 3.1 to 3.6-fold) in the tocilizumab group versus a 0.02-fold decline (95% CI, −0.44 to +0.40-fold) in the placebo group (3.4-fold relative change, 95% CI, 2.9 to 3.8-fold, adjusted P<0.001) (Figure 2C). No treatment-dependent difference was seen in any other plasma cytokine concentration.

Figure 2.

Plasma (A–C, logarithmic scale) and cerebrospinal fluid (CSF; D–F, linear scale) trajectories of select cytokines and proteins. Adjusted mean estimates and their 95% confidence intervals are plotted by treatment group and visit. Concentrations of the inflammatory indicator C-reactive protein (CRP; A, D) are reduced, while interleukin 6 (IL-6; B, E) and its soluble receptor (sIL-6R; C, F) levels are significantly increased in plasma and CSF in participants treated with tocilizumab. Results are expressed as mean concentration ± 95% CI. Statistical analysis to determine significant differences was performed as described in the methods. Adjusted P-values for differences between participants treated with tocilizumab versus placebo at 8 weeks (a time reflective of treatment effect) and 16 weeks (a time when treatment effects are reduced), and differences between the two groups from baseline to the average of weeks 4–16, were all P<0.001 for plasma CRP, IL-6, and sIL-6R. The differences between tocilizumab versus placebo at 8 weeks were statistically significant for CSF measurements ([adjusted] CRP, P=0.01; IL-6, P<0.001; sIL-6R, P<0.001). For all measures there were no differences from baseline to weeks 4–16 in placebo-treated individuals (plasma unadjusted p–values : CRP, P=0.94; IL-6, P=0.72; IL-6R, P=0.93; CSF: CRP, P=0.41; IL-6, P=0.26; sIL-6R, P=0.70). Changes were consistently observed in tocilizumab-treated individuals (plasma unadjusted p–values: CRP, P<0.001; IL-6, P<0.001; sIL-6R, P<0.001; CSF unadjusted p–values: CRP, P<0.001; IL-6, P<0.001; sIL-6R, P<0.001).

A similar profile was observed for concentrations of CRP, IL-6, and sIL-6R in CSF (Figures 2D–F, respectively, Table 2). Relative to placebo, CRP declined 1.8-fold from baseline to week 8 (95% CI, −3.1 to −0.4-fold, unadjusted P=0.01, adjusted P=0.22) in the tocilizumab group, IL-6 increased 1.1-fold (95% CI, 0.6 to 1.6-fold, adjusted P=0.002), and sIL-6R increased 0.6-fold (95% CI, 0.4 to 0.8-fold, adjusted P<0.001). No other cytokine changes were detected between treatment groups (Table 2).

IL6R Genotype Dependent Effects

Six participants had the IL6R Asp³⁵⁸Ala AA genotype, 10 had the AC genotype, and 6 had the CC genotype, a minor allele frequency of 0.50 (95% CI, 0.35 to 0.65) for the IL6R C allele. Baseline plasma sIL-6R concentrations were higher among participants with more copies of the C allele (P=0.02). Although concentrations in the CSF were parallel, the association was not significant (Supplemental Table S4).

Tocilizumab treatment dramatically reduced plasma CRP concentrations and increased plasma IL-6 and sIL-6R concentrations independent of IL6R Asp³⁵⁸Ala genotype (Table 2). Conversely, the effects of tocilizumab on change in CSF concentrations of CRP from baseline to week 8 were nominally sensitive to IL6R Asp³⁵⁸Ala genotype (Table 2, Figure 3). For each additional C allele, the relative fold-change from baseline to week 8 for CSF CRP between tocilizumab versus placebo participants was a decline of 0.8 fold (95% CI, −1.5 to −0.09 fold, unadjusted P=0.03, adjusted P=0.74). CSF concentrations of IL-6 and sIL-6R increased in all tocilizumab-treated individuals regardless of genotype (Table 2, Figure 3), although the IL-6 increases did not reach significance in all genotypes (Table 2).

Figure 3.

Cerebrospinal fluid (CSF) trajectories of select cytokines and proteins. Adjusted mean estimates and their 95% confidence intervals are plotted by treatment group, IL6R Asp³⁵⁸Ala genotype, and visit. (A) Individuals who inherit the IL6R Ala³⁵⁸ allele exhibited reductions in CSF C-reactive protein (CRP). (B) Increases in interleukin 6 (IL-6) or (C) soluble IL-6 receptor did not appear to be influenced by the allele’s presence. Results are expressed as mean concentration ± 95% CI. Adjusted P-values for differences from baseline to week 8 between participants treated with tocilizumab versus placebo are shown. There were no differences from baseline to week 8 in any placebo-treated individuals for all measures and genotypes. Reductions in CSF CRP were observed in tocilizumab-treated AC and CC, but not AA individuals, suggesting influence by the allele (tocilizumab vs placebo X #C 8-wk change, unadjusted P=0.03). However, its significance was not maintained after correction for multiple comparisons (adjusted P=0.74). Presence of the Ala³⁵⁸ allele had no influence on tocilizumab-induced increase in IL-6 or sIL-6R (unadjusted IL-6 #C, P=0.64; sIL-6R #C, P=0.66).

Effects on ALSFRS-R and Other Clinical Indicators of Disease Progression

Rates of change in ALSFRS-R total score, ALSFRS-R domain scores, slow vital capacity, handheld dynamometry, and grip strength did not differ significantly between treatment groups or genotypes. However, given the small sample size, confidence bounds were too wide to exclude potential toxicity or benefit (data not shown).

DISCUSSION

ALS and IL-6

Our lack of understanding of disease mechanisms impedes the identification of appropriate ALS therapeutic approaches and clinical trial design optimization. The absence of accepted pathophysiologically relevant biomarkers risks enrollment of participants in trials whose predominant disease mechanisms are unlikely to respond to a given investigational drug. Response by participants with an underlying disease mechanism appropriate for the investigational drug would likely be diluted by the larger pool of participants who lack this disease mechanism, potentially masking the demonstration of efficacious treatment. For this reason, we sought to enrich our study population by including only patients with evidence of increased inflammatory gene expression.

IL-6 regulates inflammatory responses through two signaling paradigms: classical IL-6 signaling¹⁷ and IL-6 trans-signaling.¹⁸ Although classical signaling is mediated through a membrane-bound receptor on limited cell populations, trans-signaling is modulated by the extracellular sIL-6R.¹⁹ Tocilizumab has been shown to block both classical and trans-signaling.¹⁷ A common IL6R coding variant (Asp³⁵⁸Ala, rs2228145) accounts for >50% of the variability in sIL-6R levels in human subjects.^19,20 The IL6R Asp³⁵⁸Ala (C allele) occurs at different frequencies in different populations and is present in more than half of ALS patients.²¹ In a longitudinal study, serum IL-6 levels were negatively correlated with patients’ ALSFRS-R.¹² Additionally, the specific correlation with subscore and respiratory function measured by the percent predicted FVC was only present in patients who carried the IL6R Ala³⁵⁸ variant.¹¹

Thus, in addition to selecting patients with evidence of increased inflammatory gene expression, we sought to identify patients with ALS who carry the IL6R C allele to provide insight for predicting disease progression and identifying those who might benefit most from IL6R-blocking therapies, using tocilizumab as a potential treatment.

Safety and Tolerability of Tocilizumab in ALS Patients

Tocilizumab was safe and well tolerated in our study. It is also well established in the rheumatoid arthritis literature.²² Nonetheless, we felt it was essential to demonstrate this safety and tolerability profile in ALS patients, given their general debility and susceptibility to infection.²² Only 2 tocilizumab-treated participants experienced adverse events that might raise a safety concern. One patient had neutropenia, but a review of baseline laboratory results revealed this patient had borderline neutropenia, suggesting that neutropenia may have preceded treatment with tocilizumab. The patient’s levels returned to normal by week 16. Another patient developed C. difficile colitis and stopped study infusions but continued to complete all study visits. The tocilizumab package insert includes a black box warning regarding opportunistic infections, such as tuberculosis or invasive fungal infections. Additionally, gastrointestinal perforations and injection-site or hypersensitivity reactions are listed as potential serious adverse effects. None of these was seen in our cohort.

PBMC Inflammatory Gene Expression

Despite evidence that tocilizumab blocked IL-6 signaling in our participants, PBMC IL6, IL8, and MMP1gene expression did not appear to change during the study. It may be that PBMC gene expression is not an accurate pharmacodynamic biomarker for tocilizumab’s impact on ALS. Additionally, it may be that IL-6 signaling blockade does not influence PBMC gene expression or protein production and only counteracts circulating IL-6 protein levels and downstream signaling responses. We conclude that our PBMC biomarker does not predict the anti-inflammatory effect of tocilizumab in ALS patients.

IL6R Genotyping

The expected minor allele frequency for the IL6R C allele is 0.37 in white individuals and 2 previously assessed ALS populations.^21,23 The IL6R C allele frequency of 0.50 in our study was higher but did not differ significantly. The selection of ALS patients with enhanced levels of PBMC IL6 messenger RNA (mRNA) may have biased the cohort toward individuals inheriting the variant allele.

Plasma and CSF IL-6 and sIL-6R

The increases in circulating plasma and CSF IL-6 and sIL-6R in ALS patients agree with those in other patient populations treated with tocilizumab.¹⁷ Since the literature suggests that circulating IL-6 levels negatively correlate with ALSFRS-R, there is a concern that treatment with tocilizumab may be detrimental.¹² However, tocilizumab-induced increases likely reflect altered catabolism of the sIL-6R/IL-6 complexes instead of compensatory upregulation of the IL-6 pathway.¹⁷ Because the negative effect of IL-6 levels as they relate to ALSFRS-R scores is presumed to be mediated by the activation of the IL-6 pathway via sIL-6R and that negative effect is functionally blocked by tocilizumab, tocilizumab should not exacerbate ALS. Importantly, tocilizumab-induced increases in sIL-6R levels in both plasma and CSF suggest that the peripherally administered agent can have desired physiologic effects on the CNS. Consistent IL-6 and sIL-6R results between patients with or without the IL6R C allele indicate that tocilizumab tissue distribution is not affected by genotype. Although we did not measure tocilizumab levels in the CSF in our study, a prior study in nonhuman primates demonstrated that measurable levels of tocilizumab could be detected in the CSF several hours after intravenous administration.²⁴

Reductions in CRP Concentrations

Treatment with tocilizumab dramatically reduced CRP concentrations in the plasma and CSF. The effect was rapid but quickly waned in the weeks after treatment was stopped. The reduction in plasma CRP levels with tocilizumab treatment across genotypes supports a functional effect of the drug. CRP is produced by the liver and is viewed as an indicator of systemic inflammation. Interestingly, IL-6 itself appears to regulate CRP and other acute-phase protein production in the liver.²⁵ Therefore, the reductions in plasma CRP levels suggest that tocilizumab administration effectively reduces IL-6 activity in the periphery. The association of peripheral CRP levels with ALS disease progression remains in question.^26–28 As such, an ALS-specific benefit remains a topic for investigation.

Prior studies in the rheumatoid arthritis literature have also demonstrated tocilizumab’s ability to reduce CRP. Still, our study is the first to show this effect in CSF. As CSF drug levels were not assessed, this situation serves at least to demonstrate potential on-target effects in the CNS. CRP reduction is particularly important in CNS diseases where the neuroinflammatory insult occurs within the confines of the blood–brain barrier. The precise neuroinflammatory mechanisms associated with CRP expression that contribute to motor neuron injury in ALS are unclear. For example, CRP is produced by activated astrocytes and microglia that contribute to neuronal dysfunction and degeneration.^29–31 It is found within neurofibrillary tangles, suggesting a role in Alzheimer disease and other neurodegenerative diseases and may inhibit acetylcholine activity.^29–31 Many portions of the complement system have also been implicated in neuroinflammation. CRP binds to ligands on damaged cells and activates the complement cascade.³² Additionally, prior studies have shown that CRP levels in ALS patients correlate with survival and the rate of progression.²⁸ Thus, the reduction of CRP in the CNS may reduce complementary or other inflammatory damage to motor neurons and potentially slow progression in ALS patients.

Unfortunately, since the effects of tocilizumab on CRP were seen in a posthoc analysis, CRP was not used as a selection criterion. Had a CRP concentration been used as one of the inclusion criteria, it may have allowed us to better select patients who might benefit from a CRP lowering agent like tocilizumab. In retrospect, replacing the PBMC gene expression profiles with CRP concentration may have been a more useful choice.

The reduction of CSF CRP was not uniform across genotypes, with greater reductions among those who inherited more IL6R C allele copies. This result was unexpected, although the potential influence of the C allele on disease progression has been suggested.¹¹ Because CSF CRP was not prespecified as a biomarker of primary interest, its statistical significance after correction for multiple comparisons was not maintained. However, the greater reduction of CSF CRP by tocilizumab among those who inherited more copies of the IL6R C allele suggests a trans-signaling specific mechanism in a subpopulation of ALS patients that may be targetable. Further studies are needed to confirm this trans-signaling mechanism.

No clinical benefit of tocilizumab was demonstrated in this study. Moreover, this study was not originally powered to detect clinical effects. Additionally, participants had a slower-than-average disease progression because the average decrease in ALSFRS-R was less than 3 points, making it even more difficult to demonstrate a clinical benefit in such a short time frame.³³ CRP may also prove a beneficial biomarker. Other inflammatory biomarkers, such as monocyte chemoattractant protein-1, chitinases, and macrophage migratory inhibitory factor, have been identified and could also be useful as enrichment biomarkers or as markers of a positive biological effect in ALS patients.³⁴

CONCLUSIONS

The results of this study demonstrate that tocilizumab therapy is safe and well tolerated. In individuals who inherit the IL6R C allele, tocilizumab may target IL-6 trans-signaling mediated CNS glial/inflammatory responses. Alterations of CRP, IL-6, and sIL-6R suggest that tocilizumab has a pharmacodynamic effect on inflammatory systems that may be relevant to ALS disease mechanisms. Limitations of the study include the small sample size, apparent insensitivity of PBMC inflammatory gene expression profiles to treatment, lack of tocilizumab measures in the CSF, and lack of inclusion of CRP as an enrichment biomarker. An efficacy trial with a population enriched for carriers of the IL6R C allele and elevated CRP levels in the plasma and CSF is warranted.

ABBREVIATIONS:

ALS

amyotrophic lateral sclerosis

ALSFRS-R

Revised ALS Functional Rating Scale

CNS

central nervous system

C_q

quantification cycle

CRP

C-reactive protein

CSF

cerebrospinal fluid

IL

interleukin

IL-6R

interleukin-6 receptor

LLOD

lower limits of detection

MMP1

matrix metalloproteinase 1

mRNA

messenger ribonucleic acid

MSD

Meso Scale Discovery

PBMC

peripheral blood mononuclear cell

PBS

phosphate-buffered saline

qPCR

quantitative real-time polymerase chain reaction

sIL-6R

CSF or plasma-soluble IL-6R

sGP130

soluble glycoprotein 130

TBS

tris-buffered saline

TNF

tumor necrosis factor