Physicochemical Fingerprinting of Cyclosporine A Ophthalmic Emulsions as a New Approach Methodology for Clinical Tolerability

Introduction

Dry eye disease (DED) is a multifactorial disorder characterized by loss of tear-film homeostasis and ocular surface inflammation. ¹ Topical cyclosporine A (CsA), a T-cell–targeted immunomodulator, ² is widely used in moderate-to-severe DED and, owing to its lipophilicity, requires emulsion-based delivery systems. ³ Two principal CsA formulation technologies dominate clinical practice: an anionic emulsion (Restasis^®, 0.05% CsA) and a cationic emulsion (Ikervis^®, 0.1% CsA).

Although both deliver the same active compound, their clinical tolerability phenotypes diverge consistently across published data. Cationic emulsions are more frequently associated with instillation-site pain and stinging, whereas anionic emulsions are more frequently associated with transient blurred vision.^4–8 Burning sensation, in contrast, has been reported at broadly comparable frequencies between formulations,^6,8 suggesting that the divergent endpoints reflect formulation-dependent rather than drug-dependent mechanisms.

Conventional evaluation of formulation-dependent ocular tolerability has relied on animal irritation assays or large clinical trials. The current regulatory environment, however, is shifting toward New Approach Methodologies (NAMs) and Integrated Approaches to Testing and Assessment (IATA), which emphasize mechanistic physicochemical and in vitro data to reduce reliance on animal testing.^9,10 Despite this shift, the relationship between laboratory-scale physicochemical fingerprints and patient-reported tolerability phenotypes remains insufficiently defined.

This Commentary proposes a two-phase physicochemical interpretive framework—not as a demonstrated causal mechanism, but as a hypothesis-generating tool for organizing disparate observations across formulations. We hypothesize that rheological behavior may be more relevant to the immediate instillation phase (0–5 min), whereas excipient-phase volatile features detectable by gas chromatography–mass spectrometry (GC–MS) may contribute to sustained post-instillation discomfort (5–30 min). ¹¹ This framework is intended to support, rather than replace, future head-to-head clinical and biological validation studies.

Materials and Methods

Two commercial CsA ophthalmic emulsions, Restasis (0.05% CsA, anionic; Allergan, Irvine, CA, USA) and Ikervis (0.1% CsA, cationic; Santen, Osaka, Japan), were analyzed within their labeled shelf lives. A 0.5% carboxymethylcellulose-sodium (CMC) artificial tear (Refresh Plus^®, Allergan) was included as a nonionic, nonsurfactant, low-nociceptive physicochemical control, not as a clinical comparator. Although the CsA concentrations differ between the two emulsions, the bulk rheological behavior of oil-in-water emulsions is governed primarily by excipient composition and emulsion microstructure rather than by active pharmaceutical ingredient (API) concentration.

Supporting observations

Rheological fingerprints (phase 1 candidate)

At 35°C, both CsA emulsions exhibited non-Newtonian shear-thinning behavior, whereas the CMC reference exhibited near-Newtonian behavior across its measurable range (Fig. 1A). At a shear rate of 0.1 s⁻¹, Restasis viscosity (0.6542 ± 0.0450 Pa·s) was ∼5-fold higher than that of Ikervis (0.1281 ± 0.0320 Pa·s); at 100 s⁻¹, the ratio expanded to >100-fold (0.2520 ± 0.0210 vs. 0.0024 ± 0.0002 Pa·s). At blink-associated high-shear rates, Ikervis (0.0024–0.0042 Pa·s) and the CMC reference (∼0.003–0.004 Pa·s) showed comparable viscosity, consistent with the normal human tear-film range (∼1.5–4 mPa·s). ¹³ Despite this rheological similarity at high shear rates, their reported tolerability profiles differ substantially, providing the central observation underlying the framework: Viscosity alone is insufficient to account for instillation-site discomfort.

OPEN IN VIEWER

FIG. 1.

Two-panel summary of physicochemical fingerprinting observations supporting the proposed two-phase NAM framework. (A) Rheology. Steady-state shear-viscosity profiles of Restasis (0.05% CsA anionic emulsion), Ikervis (0.1% CsA cationic emulsion), and a 0.5% carboxymethylcellulose-sodium artificial-tear reference at 35°C across a shear-rate range of 0.1–100 s⁻¹. Both CsA emulsions exhibit non-Newtonian shear-thinning; Restasis showed consistently higher viscosity, with the viscosity ratio increasing from ≈5-fold at 0.1 s⁻¹ to >100-fold at 100 s⁻¹. The CMC artificial tear shows near-Newtonian behavior at ≥2 s⁻¹, with viscosity comparable to that of Ikervis at high shear and within the normal human tear-film range (shaded band). Data are presented as mean ± SD from triplicate measurements (n = 3 per formulation). ^† CMC data are shown only for shear rates ≥2 s⁻¹; lower-shear values were excluded owing to instrument torque-sensitivity limits. (B) GC–MS chemical fingerprinting. Representative total ion chromatograms of the excipient phases of Ikervis (top), Restasis (middle), and the CMC artificial-tear reference (bottom). In Ikervis, the chromatographic peak at retention time of 13.024 min was putatively assigned to dimethyl palmitamine (long-chain volatile tertiary amine, C16; m/z 58, 269)—a candidate thermal degradation product or structural analog of cetalkonium chloride. In Restasis, the dominant peak at 12.722 min was putatively assigned to glycerol (m/z 61, 43). In the CMC reference, no signal consistent with a long-chain tertiary amine was detected in the 12–14 min region; a minor peak at ≈10.6 min was putatively assigned to N,1-dimethylhexylamine (short-chain amine, C6; m/z 58). Compound assignments were based on putative identification through NIST 23 GC–MS library matching (≥70% match-score threshold) and have not been confirmed against authentic analytical standards. CsA, cyclosporine A; GC–MS, gas chromatography–mass spectrometry; SD, standard deviation.

Excipient-phase chemical fingerprints (phase 2 candidate)

GC–MS revealed distinct chemical signatures across the three formulations (Fig. 1B). In Ikervis, a peak with a retention time 13.024 min was putatively identified as a long-chain volatile tertiary amine species (dimethyl palmitamine, C16; m/z 58, 269)—a candidate thermal degradation product or structural analog of cetalkonium chloride (CKC), the cationic surfactant present in Ikervis. In Restasis, the dominant peak at 12.722 min was putatively assigned to glycerol (m/z 61, 43). In the CMC reference, no signal consistent with a long-chain tertiary amine was detected in the 12–14 min region; only a minor peak at ∼10.6 min, putatively assigned to N,1-dimethylhexylamine (short-chain amine, C6; m/z 58), was observed. All assignments are based on NIST 23 library matching and have not been confirmed against authentic standards; the potential for thermal degradation or fragmentation artifacts inherent to GC–MS analysis of surfactant-like species is acknowledged.

Mapping to published tolerability data

Phase 1: Restasis is associated with transient blurred vision (1%–5% per FDA prescribing information), ⁶ consistent with its higher viscosity profile. Ikervis is associated with a lower incidence of visual disturbance (2.1% in the SANSIKA RCT). ⁴ Phase 2: Ikervis-associated instillation-site pain has been reported at 29.2% at 6 months (SANSIKA), ⁴ 12.1% in a pooled Phase III analysis (n = 734), ⁷ and 7.8% at 12 months of continuous treatment in the SANSIKA open-label extension ⁵ —a trajectory that is consistent with corneal nerve adaptation. Restasis-associated eye pain is reported at 1%–5% according to the FDA label. ⁶ The long-chain tertiary amine signal detected exclusively in Ikervis represents a candidate formulation-specific excipient-phase feature whose relevance to this pain pattern is hypothesized—not demonstrated.

Internal specificity check

Burning sensation was reported at broadly comparable frequencies across formulations (∼17.5% Ikervis, EMA SmPC ⁸ ; 17% Restasis, FDA prescribing information). ⁶ Because this nonspecific endpoint did not show a formulation-dependent directional difference, it functions as an internal specificity control: The divergent patterns in pain and blurred vision against the comparable burning rates support the specificity of the proposed two-phase framework rather than reflecting a generalized irritation difference.

Implications and Limitations

The principal contribution of this Commentary is not the establishment of clinical causality but the proposal of a tractable physicochemical framework that may help interpret formulation-dependent tolerability patterns and guide future targeted validation.

Future Directions

Within the limits set above, the framework defines a tractable validation agenda. Future work will, where feasible, employ (1) authentic-standard GC–MS confirmation of the long-chain tertiary amine signal, (2) orthogonal high-resolution mass spectrometry, (3) in vitro nociception assays—including TRPV1- and TRPA1-related models—to test whether the candidate compound or post-instillation formulation activates ocular surface nociceptive pathways, and (4) ex vivo corneal models that better approximate the dynamic tear-film environment. Eventual head-to-head clinical evaluation under harmonized endpoints would be the definitive test of the framework. The intent of this Commentary is to render the proposed mechanism testable, not to assert that it is true.

Conclusions

This Commentary proposes a two-phase physicochemical interpretive framework—rheology-governed early phase and a GC–MS-detectable excipient-phase contributions to a sustained phase—as a NAM/IATA-aligned, hypothesis-generating tool for ophthalmic formulation tolerability. Preliminary observations on Restasis, Ikervis, and a CMC physicochemical reference are consistent with the framework but do not establish its mechanistic validity. GC–MS fingerprinting may serve as a useful complementary tool alongside rheology in future formulation tolerability assessment. Targeted validation through authentic-standard analytical confirmation and in vitro nociception models remains essential.

Author’s Contributions

J.K.: Conceptualization, data curation, formal analysis, investigation, methodology, visualization, writing—original draft, and writing—review and editing.

Data Availability Statement

The original rheological and GC–MS data supporting the conclusions of this article are available from the corresponding author upon reasonable request.

Supplemental Material