Toxicological evaluation of a synthetic QS-21 adjuvant in a tuberculosis vaccine candidate—Demonstrating safety and immunological equivalence to natural reference

ABSTRACT

Synthetic saponin-based adjuvants hold promise for overcoming the supply instability and purity limitations inherent to naturally derived QS-21. However, establishing their toxicological and functional equivalence to the natural reference is a necessary step before clinical use. This study systematically evaluated the safety and immunogenicity of a recombinant tuberculosis vaccine (named LYB007) formulated with a synthetic QS-21-Api compared to an identical formulation using a naturally-derived reference via single- and repeated-dose toxicity assessments in Sprague-Dawley (SD) rats. Both synthetic and naturally-derived formulations exhibited highly comparable safety and immunological profiles. Injection site reactogenicity was self-limiting, with complete histological resolution by Day 58. Systemic physiological adaptations, including transient thermal shifts and acute-phase protein fluctuations, were consistent with the potent immunostimulatory activity of QS-21 and did not have long-term toxicological significance. Notably, both formulations induced robust, uniform (100%) seroconversion with identical kinetics of antibody development and persistence, with peak geometric mean titers (GMTs) consistently attained between Day 29 and Day 43. The strict seronegativity in negative and mock-vaccine control groups confirmed that the humoral response was exclusively driven by the LYB007-adjuvant complex. Histopathological findings, such as splenic germinal center expansion and extramedullary hematopoiesis, were identified as expected immunological adaptations rather than manifestations of systemic toxicity. In conclusion, these findings confirm that the synthetic QS-21-Api possesses a safety and efficacy profile equivalent to the naturally-derived reference. The use of a chemically defined and standardized adjuvant component overcomes key constraints in vaccine scalability and quality control, supporting the development of future high-efficacy subunit vaccines.

Introduction

Adjuvants are essential components in the development of high-performance subunit vaccines, particularly those targeting complex intracellular pathogens such as Mycobacterium tuberculosis. Among the diverse repertoire of immunomodulators, QS-21, a triterpene glycoside isolated from the bark of the South American tree Quillaja saponaria Molina, is considered the benchmark saponin-based adjuvant for modern vaccine development.^1,2 Its unique ability to activate both high-titer humoral responses and potent Th1-type cell-mediated immunity (CMI) has made it a cornerstone of advanced adjuvant systems, such as the AS01 series.^3,4 This system has achieved remarkable clinical success, most notably in the recombinant zoster vaccine Shingrix, which induced humoral and CMI responses that persisted above pre-vaccination levels for at least 9 y in adults ≥ 60 y, with modeling data predicting protection lasting up to 15 y,⁵ and the RTS,S/AS01 (Mosquirix) malaria vaccine, the first WHO-approved malaria vaccine.⁶ Furthermore, there are over 40 other QS-21 adjuvant vaccines being under development. The success of the M72/AS01E tuberculosis vaccine candidate in Phase IIb trials – showing significant protection against progression to active disease – underscores the necessity of QS-21-based platforms in overcoming the limitations of traditional subunit vaccines.⁷

Despite its clinical success, the sustainability of natural QS-21 is severely compromised by the inherent limitations of its extraction and purification process. The primary source of natural QS-21 is the bark of mature Quillaja saponaria trees, which typically require 25 to 30 y to reach maturity for harvest, rendering the supply chain ecologically sensitive and susceptible to extreme volatility.⁸ Furthermore, while crude saponin mixtures may be sufficient for global food and agricultural applications, these extracts are not of sufficient quality to isolate QS-21 on a large scale to meet global vaccine demand. A critical challenge is the complex purification process required to isolate QS-21 from the heterogeneous saponin mixture in the bark, especially from other saponins with similar structures.^9–11 Beyond purity concerns, recent structural insights have revealed that the immunostimulatory activity of QS-21 is inextricably linked to its complex amphiphilic structure, yet the same structure is responsible for dose-limiting toxicity, including hemolysis and local reactogenicity.^12–14 While natural extracts provide a fixed, evolutionary-determined composition, rational design and synthetic production offer a transformative alternative. By systematically modifying the triterpene core and acyl chain, synthetic chemistry enables the development of analogs that retain immunostimulatory activity while reducing the hemolytic and reactogenic moieties responsible for dose-limiting toxicity.¹⁵ These supply chain vulnerabilities and the potential for molecular optimization have necessitated a transition toward synthetic and semi-synthetic QS-21 alternatives.

Evidence suggests that the immunomodulatory potential of QS-21 can be significantly amplified when integrated into combinatorial adjuvant systems. By leveraging synergistic signaling cascades – such as the integration of QS-21 with Toll-like receptor (TLR) agonists [e.g., Monophosphoryl Lipid A (MPLA) or synthetic analogues]—vaccine candidates can achieve a more balanced Th1/Th2 profile and potent adaptive immunity, as evidenced by the success of licensed platforms like AS01 and AS15.^15–18 Simultaneously, chemical formulation engineering has become a critical strategy for addressing the inherent challenges of QS-21, such as its amphiphilic-driven membrane-lytic activity. By sequestering QS-21 within sophisticated delivery systems – including liposomes,^19,20 nanoparticle-based immune-stimulating complexes (ISCOMs),^21–23 or ionizable lipid nanoparticles (LNPs)^24,25—we can effectively mask the membrane-lytic moieties responsible for hemolysis and local reactogenicity while preserving the potent immune-activating warhead.

In the context of a “platform approach” for vaccine development, establishing the toxicological equivalence of synthetic QS-21-based formulations compared to the natural gold standard is a vital prerequisite for clinical translation. This approach is based on that a synthetic variant can serve as a viable, sustainable, and reliable component for the vaccine platform, provided its safety profile remains consistent with the natural reference across systematic inflammatory, hematological, and histopathological profiles. The present study describes a comprehensive single-dose and repeated-dose toxicity assessment in Sprague-Dawley (SD) rats to evaluate the safety and tolerability of a recombinant tuberculosis vaccine candidate formulated with synthetic QS-21. As a standard rodent model with extensive historical control data, SD rats allow clear differentiation between test article-related toxicities and spontaneous strain-specific lesions. By focusing on the kinetics of immune activation and the reversibility of tissue-level reactions, we aim to provide a rigorous scientific basis for the use of synthetic QS-21 as a standardized, quality-controlled alternative to heterogeneous natural extracts. These findings are intended to support the broader development of a QS-21-based vaccine platform and to establish consistent quality and safety benchmarks for future candidates.

Material and methods

Ethical statement

All animal experimental procedures were conducted in strict accordance with the national laws and regulations governing the use and management of laboratory animals. The GLP-compliant study protocols, which detailed the study objectives, study design, analysis plans and other key information, were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of Chongqing Medleader Bio-Pharm Co., Ltd (Approval Nos: IACUC-T2506401–01, IACUC-T2506410A-01, IACUC-T2506411–01, and IACUC-T2506411A-01). All animal handling procedures strictly adhered to 3Rs principles and welfare standards, executed as approved under continuous institutional oversight. To minimize distress, standardized gentle handling was practiced during dosing, blood sampling, restraint and anatomy. Predefined humane endpoints were established, under which rats exhibiting severe, unrelieved pain or moribundity during observation would be humanely euthanized.

Animals and housing

Wild-type specific pathogen-free (SPF) SD rats (weight: 160–230 g; age: 5–7 weeks) were obtained from Chongqing Jilihui Biotechnology Co., Ltd. Eligible animals were confirmed free of common pathogens, lesions, tumors, malformations, other underlying conditions and laboratory abnormalities prior to study enrollment. Animals were housed in a temperature-controlled environment (20 ~ 26°C) with a 12-hour light/dark cycle and ad libitum access to a standard pelleted diet and sterilized water. Cages were distributed uniformly across the housing area and repositioned weekly to eliminate positional bias. Animals were group-housed with ≤5 rats per cage. All animals were acclimatized for at least 5 d prior to study initiation. Animals showing spontaneous diseases, abnormal physical signs, severe stress, or off-target weights during acclimatization were excluded.

Biological components and reagents

The recombinant tuberculosis fusion protein antigen LYB007 (lyophilized) and the A02 series liposomal adjuvant system were developed in-house. The adjuvant variants were categorized into natural (natural extraction methods) QS-21 and synthetic QS-21-Api (chemical synthesis). Natural-extracted QS-21 is a mixture of QS-21-API and QS-21-xyl (Figure 1), with a ratio of approximately 65:35. Highly purified natural QS-21 (>98%) used in this study is commercially obtained from Nanjing Well Pharmaceutical Technology Co., Ltd., with its structure has been fully confirmed by mass spectrometry, nuclear magnetic resonance, etc. The synthetic QS-21-Api used in this study was prepared via total chemical synthesis, with its chemical structure illustrated in Figure 1. The compositions of the adjuvants were as shown in as Table 1.

Figure 1.

Structures of QS-21 isomers.

Table 1.

Compositions of A02 adjuvants.

Adjuvant	Composition
Natural A02B	QS-21 (94.9 μg/mL), HMLA (98.2 μg/mL), cholesterol (0.52 mg/mL), and DOPC (2.20 mg/mL)
Natural A02C	QS-21 (48.1 μg/mL), HMLA (49.3 μg/mL), cholesterol (0.27 mg/mL), and DOPC (1.12 mg/mL)
Synthetic A02B	QS-21-Api (97.3 μg/mL), HMLA (96.6 μg/mL), cholesterol (0.52 mg/mL), and DOPC (2.20 mg/mL)
Synthetic A02C	QS-21-Api (47.1 μg/mL), HMLA (48.4 μg/mL), cholesterol (0.25 mg/mL), and DOPC (1.07 mg/mL)

Experimental vaccine preparation

All vaccine formulations were prepared under aseptic conditions immediately prior to administration. The lyophilized LYB007 antigen was reconstituted by transferring the entire contents of one vial of liquid adjuvant (0.6 ml/vial) into one vial of lyophilized antigen for reconstitution (solely with the liquid adjuvant). Followed by a gentle swirl to mix until complete homogeneity, 0.5 mL solution will be withdrawn with the syringe as one dose for administration.

1. Single-dose toxicity study preparation (total dose: 80 μg): reconstituted using two vials of adjuvant/antigen to achieve a total volume of 1.0 mL (containing 80 μg antigen).

• LYB007 (A02B): reconstituted antigen in A02B adjuvant [containing 50 μg QS21 and 50 μg Monophosphoryl Lipid (MPL) per adjuvant].

• LYB007 (A02C): reconstituted antigen in A02C adjuvant (a separate batch manufactured at 50% concentration of A02B adjuvant with same volume per vial, 25 μg QS21 and 25 μg MPL per adjuvant).

• Mock vaccine (CA): lyophilized antigen placebo in A02B adjuvant (containing 50 μg QS21 and 50 μg MPL per adjuvant).

• Negative control (CN): 0.9% sodium chloride.

2. Repeat-dose toxicity study preparation: reconstituted using one vial of adjuvant/antigen to achieve a total antigen dose of 40 μg or 10 μg in 0.5 mL, with 50 μg QS21 and 50 μg MPL per adjuvant.

• High dose (HD): 40 μg antigen in A02B adjuvant.

• Low dose (LD): 10 μg antigen in A02B adjuvant.

• Mock vaccine (CA): lyophilized antigen placebo in A02B adjuvant.

• Negative control (CN): 0.9% sodium chloride.

All formulations were gently agitated to ensure complete homogenization and a colorless to light brown opalescent appearance, and were stored at 4°C for immediate use.

Study design

The toxicological profiles of the natural and synthetic formulations were assessed in independent parallel studies. Each study series followed identical protocols to ensure comparability between the two adjuvant sources.

Single-dose toxicity studies

70 eligible rats were stratified by sex and baseline body weight, and assigned to groups via RANK-based randomization using an independent sequence of generated random numbers. As an open-label study, group allocations were known to all personnel involved. For each adjuvant version, 30 SD rats were randomly assigned to CA (mock vaccine, LYB007 antigen placebo + A02B adjuvant), LYB007 (A02B) (LYB007 antigen 80 μg + A02B adjuvant), LYB007 (A02C) (LYB007 antigen 80 μg + A02C adjuvant), with 10 rats per group (5 males and 5 females). 10 SD rats (5 males and 5 females) were randomly assigned to the CN group to serve as a shared control groups for both adjuvant cohorts.

A total volume of 1.0 mL was administered via intramuscular (IM) injection across four sites: bilateral triceps brachii (forelimbs) and bilateral quadriceps femoris (hindlimbs), at 0.25 mL per site. Rats were monitored for 14 d post-administration. Operator bias during animal handling and parameter measurements was minimized by regular rotation of study personnel.

Repeat-dose toxicity studies

Independent Good Laboratory Practice (GLP) studies were conducted for each adjuvant version. To comply with sample size requirements from CDE (Technical Guidelines for Repeated Dose Toxicity Studies) and ICH S4/S6(R1) guidelines, a total of 120 rats were randomized into three groups for each adjuvant version (30 main study and 10 satellite animals per group, 50% male and 50% female): CA, LYB007 (low dose, LD) (10 μg antigen + A02B), and LYB007 (high dose, HD) (40 μg antigen + A02B), while 40 SD rats (20 males and 20 females) were randomly assigned to the CN group to serve as a shared control groups for both adjuvant cohorts. Rats were randomized using the same sex- and weight-stratified RANK-based procedure described for the Single-Dose Toxicity Studies, with both studies employing an open-label design. Rats received three IM injections on Days 1, 15, and 29, matching the maximum intended clinical regimen in accordance with ICH M3(R2) guidelines. The vaccine was administered bilaterally into the quadriceps femoris (0.25 mL per site, total 0.5 mL/animal). Following the dosing period, a 4-week recovery period was observed. Operator bias was minimized by regular rotation of technicians.

Clinical signs and local reactogenicity

In the single-dose studies, rats were monitored continuously for 1 hour post-administration and daily thereafter. In the repeat-dose studies, rats were monitored continuously for 3-hour post-administration and twice daily thereafter. Clinical signs, including behavior, posture, and mortality, were recorded. Injection site irritation was evaluated using a standardized 5-point scale (0: no changes; 5: extensive necrosis). Detailed clinical examinations were performed weekly by a veterinarian.

Body weights were recorded at scheduled intervals (e.g., D1, D8, D15, D22, D29, etc.). Average daily food intake per animal was calculated as: (Total mass provided – Mass remaining)/number of animals per cage. In the repeat-dose studies, rectal temperature was monitored pre- and post-administration to detect potential pyrogenic responses as scheduled.

Ophthalmic examination

In the repeat-dose studies, ophthalmic assessments were conducted prior to group assignment, on D29, and on D57, targeting the eyelids, conjunctiva, cornea, iris, and fundus.

Clinical pathology

Blood was collected via the abdominal aorta at the end of the dosing (D30) and recovery (D58) periods. Hematology: analyzed using an automated analyzer for parameters including white blood cell (WBC), red blood cell (RBC), hemoglobin (HGB), and platelet (PLT). Coagulation: evaluated prothrombin time (PT), activated partial thromboplastin time (APTT), and Fibrinogen (Fbg). Serum biochemistry: encompassing aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), gamma-glutamyl transferase (GGT), creatine kinase (CK), blood urea nitrogen (BUN), creatinine (CRE), total proteins, glucose, lipids, and inflammatory markers [C-reactive protein (CRP), immunoglobulins, and complement C3/C4]. Also, the urine laboratory examination is conducted at the end of the dosing (D30) and recovery (D58) periods.

Gross necropsy and histopathology

Terminal necropsy was performed on all surviving animals. Animals were fasted for 12–16 hours prior to necropsy, with free access to water throughout the fasting period. Euthanasia was performed under deep anesthesia, induced via intraperitoneal (i.p.) injection of Zoletil 50 (50 mg/kg; injection volume: 1 mL/kg). Subsequently, animals were euthanized via exsanguination through the abdominal aorta to minimize autolytic changes. A full macroscopic examination was conducted, focusing on the systemic status of major organs and the local injection site. Major organs were weighed, and organ-to-body/brain weight ratios were calculated. Tissues were fixed in 10% neutral-buffered formalin or modified Davidson’s fluid (for eyes and reproductive organs). H&E-stained sections were subjected to microscopic examination. All findings were reviewed by a certified toxicological pathologist through an independent peer-review process.

Immunological analysis

Lymphocyte subsets and serum cytokine were evaluated in repeat-dose studies on D30 and D58. Flow cytometry was utilized to analyze T-lymphocyte subsets [Cluster of Differentiation (CD3+, CD3+CD4+, and CD3+CD8+)] in peripheral blood. Serum cytokine levels [Tumor Necrosis Factor-α (TNF-α), Interferon-γ (IFN-γ), Interleukin-2 (IL-2), and (IL-4) were quantified using the Meso Scale Discovery (MSD) platform.

To measure the antigen-specific IgG response, enzyme-linked immunosorbent assays (ELISAs) were performed. In brief, blood samples were collected and sera were obtained and stored at − 20°C. ELISA plates were coated overnight at 4°C with LYB007 antigen (2 μg/mL, 50 μL per well), washed with phosphate buffered saline with Tween 20 (PBST, 0.05% Tween 20), and blocked at room temperature for 1–4 h. Serially diluted sera (starting at 1:100, followed by 3-fold dilutions) were added and incubated for 1 h, followed by incubation with Horseradish Peroxidase (HRP)-conjugated anti-IgG secondary antibody for 1 h. After washing, substrate solution was added for 10 min in the dark, and the reaction was stopped with 1 M HCl. Absorbance was measured at 450 nm (reference 620 nm). Endpoint titers were defined as the highest serum dilution yielding an absorbance exceeding the mean Optical Density (OD) of naïve serum by at least three standard deviations or 0.15, whichever was greater, based on interpolation from the dilution curve.

Statistical analysis

Data are presented as mean ± SD and analyzed using SPSS 19.0. Normality and variance homogeneity were assessed via the One-Sample Kolmogorov-Smirnov and Levene’s tests, respectively. For normally distributed and homogeneous data, One-Way ANOVA followed by Dunnett’s t-test was employed. For non-parametric data, the Kruskal-Wallis H test followed by the Mann-Whitney U test was applied. Statistical significance was defined at P ≤ .05.

Results

Comparative single-dose toxicity of natural and synthetic adjuvant formulations

Clinical observations and General Necropsy

All rats in both the natural and synthetic adjuvant study series survived until the scheduled termination on Day 15. During the 14 d observation period, no treatment-related clinical signs of systemic toxicity or abnormal behavioral patterns were observed. Gross necropsy performed on Day 15 revealed no macroscopic lesions or treatment-related abnormalities in any of the experimental groups, whether in the LYB007 with natural adjuvant or synthetic adjuvant formulation series.

Body weight and food consumption

A consistent transient physiological response was observed across both study series following intramuscular administration. Animals in all adjuvant-treated groups (mock vaccine, LYB007-A02B, and LYB007-A02C) exhibited a temporary weight decline on Day 2, which was followed by a return to stable growth trajectories (Figure 2 (A–D)). While statistically significant reductions in body weight were observed sporadically, these effects were fully resolved by Day 8, with no significant differences (P > .05) compared to the negative control group observed thereafter.

Figure 2.

Impact of single-dose LYB007 on body weight and food consumption in SD rats. (A–D) individual and average body weight trajectories for male (A,C) and female (B,D) SD rats over a 14 d observation period. (E–H) detailed food consumption patterns calculated for specific intervals, of which data are expressed as estimated means (SD based on individual animals are not provided because food consumption was recorded on a per-cage basis). Solid lines represent cohorts receiving synthetic adjuvant formulations, while dashed lines represent natural adjuvant controls. Shapes represent treatment groups (circle: CN, negative control; square: CA, mock vaccine; regular triangle: low-dose LYB007; inverted triangle: high-dose LYB007). Symbol fill status represents biological sex (all solid symbols: males; all open symbols: females). Line styles represent adjuvant types (solid lines: synthetic adjuvant; dashed lines: natural adjuvant). n = 10/group (5 males and 5 females) for each adjuvant cohort and the shared control (CN) group.

A dose-dependent reduction in food intake was consistently documented during the initial 4 d post-administration window (Figure 2 (E–H)). Full adjuvant dose (A02B) & mock groups: These groups experienced the most pronounced decrease in food consumption, with peak reductions of ~30% to nearly 60% within the first 48 hours compared to the negative control. Half adjuvant dose (A02C) groups: Consistent with the reduced adjuvant load, these groups demonstrated a more moderate and transient impact on food intake. Importantly, food consumption for all groups returned to levels comparable to the negative control by D7–D8, demonstrating the rapid resolution of the acute inflammatory stimulus.

Results of repeated-dose toxicity studies: evaluation of LYB007 vaccine candidate with synthetic/natural adjuvants

Systemic safety, body weight and body temperature

Both the synthetic and natural adjuvant LYB007 candidate groups exhibited consistent systemic safety profiles throughout the repeated-dose study. Apart from transient injection-site swelling – which occurred in all treatment groups (with adjuvant) and resolved spontaneously within 1–5 d, no mortality, morbidity, or behavioral abnormalities were recorded.

Consistent with the immunostimulatory nature of the formulations, transient and reversible fluctuations in body temperature, food consumption, and body weight gain were observed across all treated cohorts. Specifically, all treatment groups (including the mock vaccine control and LYB007 low/high dose groups) exhibited a mild increase in body temperature approximately 3–6 hours post-injection on Day 1, Day 15, and Day 29 (Figure 3). In the synthetic QS-21-Api cohort, these increases ranged from 2.16% to 3.78%, while the natural QS-21 cohort showed similar elevations of 2.13% to 4.32% (P ≤ .05). In most instances, these febrile responses stabilized within 24 hours and fully returned to baseline levels comparable to the negative control group by the following day.

Figure 3.

Characterization of localized reactogenicity and transient systemic thermoregulation. (A–B) incidence of localized swelling at the injection site following the 1st, 2nd, and 3rd administrations of synthetic (A) and natural (B) adjuvant formulations. (C–F) dynamic changes in rectal temperature were recorded across three dosing cycles. CN, negative control; CA, mock vaccine; LYB007 (LD/HD), low-dose and high-dose groups; syn-Adj/Nat-adj, synthetic/natural adjuvant groups. n = 40/group (20 males and 20 females) for each adjuvant cohort and the shared control (CN) group.

Parallel to these thermal shifts, transient fluctuations in body weight and food consumption were observed in both cohorts (Figure 4). A temporary reduction in body weight gain was observed in males of all treatment groups during the D1–D8 interval (23.0%–26.8% reduction; P ≤ .01) compared with the negative control. Concurrently, transient decreases in food intake were recorded following the initial and second doses in both the synthetic QS-21-Api (D15–D16; 19.01%–30.42% reduction; P ≤ .01) and natural QS-21 (D1–D2 and D15–D16; 18.9%–59.8% reduction; P ≤ .01) groups. Crucially, these effects lacked long-term toxicological significance, as both food consumption and body weight gain normalized rapidly following the third dose and throughout the subsequent recovery period.

Figure 4.

Longitudinal assessment of body weight and nutritional status during repeated-dose toxicity study of LYB007. (A–D) body weight gain trajectories for male (A,C) and female (B,D) SD rats across the dosing phase (D1–D30) and recovery phase (D31–D57). (E–H) interval-based food consumption (g/rat/day) in males and females. Solid lines (A,B) represent synthetic adjuvant groups, while dashed lines (C,D) represent natural adjuvant groups. (E–H). Shapes represent treatment groups (circle: CN, negative control; square: CA, mock vaccine; regular triangle: low-dose LYB007; inverted triangle: high-dose LYB007). Symbol fill status represents biological sex (all solid symbols: males; all open symbols: females). Line styles represent adjuvant types (solid lines: synthetic adjuvant; dashed lines: natural adjuvant). n = 40/group (20 males and 20 females) for each adjuvant cohort and the shared control (CN) group.

Local reactions were observed in male and female animals dosed intramuscularly (IM) with both the natural and synthetic LYB007 formulations. At the end of the dosing phase, irritation scores ranged from 0.10 to 0.70 across all groups (Table 2). According to the standardized irritation grading criteria (0.0–0.4 for “None”; 0.5–1.4 for “Minimal”), the administration of the vaccine candidates induced, at most, a “Minimal” level of muscle irritation. The synthetic LYB007 variant demonstrated a local tolerance profile clinically equivalent to that of the natural reference standard. Importantly, these findings were self-limiting. Following the recovery period (Day 58), a complete resolution of all local irritation was observed in every cohort, with mean irritation scores returning to 0 (grading as “None”).

Table 2.

Macroscopic irritation scores of injection sites during the dosing and recovery phases.

Time	location	n	group
			CN	CA	LYB007 (LD)	LYB007 (HD)
Synthetic adjuvant
D30	Left hind limb	20	0.25	0.40	0.35	0.60
	Right hind limb	20	0.10	0.50	0.35	0.45
D58	Left hind limb	10	0	0	0	0
	Right hind limb	10	0	0	0	0
Natural adjuvant
D30	Left hind limb	20	0.25	0.30	0.60	0.50
	Right hind limb	20	0.10	0.25	0.70	0.45
D58	Left hind limb	10	0	0	0	0
	Right hind limb	10	0	0	0	0

Data represent the average irritation scores Scoring criteria: 0.0–0.4, no irritation; 0.5–1.4, mild irritation. CN, negative control; CA, mock vaccine; LY007 (LD/HD), low-dose and high-dose adjuvant groups.

Comprehensive ophthalmological examinations were performed at baseline, Day 29 and the day prior to the end of the recovery period (Day 57). The examinations encompassed a detailed bilateral assessment of the ocular structures, including the eyelids, conjunctiva, cornea, sclera, iris, and anterior chamber. Retinal integrity, vascular patterns and neurological responses were also evaluated. Across both the synthetic and natural QS-21 cohorts, no treatment-related abnormalities or clinically significant changes were observed in any of the examined structures at any time point. All ocular tissues maintained a normal physiological appearance, and no signs of inflammation, opacity, or degenerative changes were recorded.

Clinical pathology

The clinical pathology profiles for both synthetic QS-21-Api and natural QS-21-based LYB007 formulations revealed patterns consistent with a robust, transient systemic inflammatory response. The high degree of concordance between the two cohorts confirms the safety equivalence of the synthetic adjuvant platform.

Hematology: inflammatory Leukogram

Consistent with an inflammatory leukogram, significant elevations in total white blood cell (WBC) counts were observed compared with the CN group, at the end of the dosing phase (Day 30). This was primarily driven by marked increases in neutrophils (NEUT#) and eosinophils (EO#) across both genders. While transient reductions in red blood cell mass (HGB, HCT, RET#) and platelet counts (PLT) were observed, these findings were of limited magnitude, lacked consistent dose-dependency, and returning to the CN-comparable levels by the end of the recovery phase (P > .05, Figure 5 (A,B)).

Figure 5.

Hematological and coagulation profiles, and systemic acute phase responses during repeated administrations of LYB007. (A–B) Comprehensive hematology assessment including counts of neutrophils, reticulocytes, lymphocytes, and platelets at Day 30 and Day 58. (C–D) Coagulation parameters of APTT, PT, and Fbg are shown for the synthetic and natural adjuvant groups at Day 30 and Day 58. (E–F) Acute phase response monitoring of Fbg concentrations at Day 2 and Day 7. APTT, activated partial thromboplastin time; PT, prothrombin time; Fbg, fibrinogen; CN, negative control; CA, mock vaccine; LYB007 (LD/HD), low-dose and high-dose groups; M, male; F, female. n = 40/group (20 males and 20 females) for each adjuvant cohort and the shared control (CN) group. The dotted horizontal line represents the background reference range.

Coagulation parameters

The coagulation profile indicated a transient, mild hypercoagulable state consistent with systemic vaccine-induced inflammation (Figure 5 (C,D)). Significant elevations in Fbg and APTT were observed compared with the CN group, at the end of the dosing phase (Day 30). These shifts are characteristic of an activated coagulation cascade during the acute inflammatory phase. Crucially, all coagulation markers were completely reversible, returning to the CN-comparable levels by Day 58 (P > .05), indicating no cumulative impact on coagulation homeostasis.

Acute phase response

To assess the kinetics of the systemic inflammatory response, fibrinogen (Fbg) levels were monitored as a primary positive acute phase reactant (Figure 5 (E,F)). Following the initial administration, both synthetic and natural LYB007 candidate groups exhibited a rapid, a significant increase in Fbg levels compared with the CN group within 24 hours (Day 2). This immediate response was observed across both sexes, with increases ranging from approximately 52.8% to 281.9% relative to the negative control group. This prompt induction of fibrinogen confirms activation of the hepatic acute phase response, a predictable and transient manifestation of the immunostimulatory activity of the QS-21 adjuvant platform. Notably, the inflammatory reaction was self-limiting across all cohorts; by Day 7, Fbg levels had returned to, the CN-comparable levels, indicating no prolonged or cumulative inflammatory burden.

Metabolic adaptation and complement dynamics

The clinical chemistry profiles of both synthetic and natural adjuvant LYB007 formulations demonstrated a signature pattern of vaccine-induced systemic response, characterized by acute phase protein shifts and immunological engagement (Figure 6). Both cohorts exhibited consistent biochemical evidence of a robust acute phase response at Day 30. This was characterized by significant reductions in serum albumin (ALB) and the albumin: globulin (A/G) ratio, alongside reciprocal elevations in acute-phase reactants such as C-reactive protein (CRP) and globulin (GLB) fractions. These shifts reflect the physiological redirection of hepatic protein synthesis toward inflammatory defense mechanisms, a hallmark of potent adjuvant-mediated immune stimulation. The complement system which is a critical effector of the immune response, showed highly parallel behavior across groups. We observed a significant and sustained consumption of Complement C4 (a key component of the classical complement pathway) from Day 30 through the recovery phase (Day 58). This reduction serves as a biomarker of humoral immune activation, corroborated by concurrent elevations in IgG and IgM levels observed in the LYB007-treated cohorts. These changes were non-toxicological, self-limiting, and reflective of active immunological memory formation. Transient elevations in total bilirubin (TBIL), creatinine (CRE), and AST were observed in both groups. However, these biochemical fluctuations occurred without evidence of structural organ damage: (1) liver and kidney histopathology remained normal, indicating that these changes represent transient physiological shifts rather than organ toxicity; (2) the observed decreases in blood glucose (GLU) are consistent with the increased energy demands of the immune system during the acute inflammatory phase; (3) all metabolic and hepatic markers, including Na+, glucose, and bilirubin, trended toward or returned to baseline levels by the end of the recovery phase (Day 58), confirming the transient nature of these responses.

Figure 6.

Biochemical profiles and systemic immune homeostasis during repeated administrations of LYB007. (A–E) Assessment of hepatic and renal functional integrity, including levels of ALT, AST, total bilirubin, albumin, and creatinine at Day 30 and Day 58. (F–H) Monitoring of systemic immune parameters, including serum IgG concentrations (F) and complement components C3 and C4 (G,H); ALT, alanine aminotransferase; AST, aspartate aminotransferase; CN, negative control; CA, mock vaccine; LYB007 (LD/HD), low-dose and high-dose groups; M, male; F, female. n = 40/group (20 males and 20 females) for each adjuvant cohort and the shared control (CN) group. The dotted horizontal line represents the background reference range.

Urinalysis

Urinalysis parameters were evaluated to monitor renal filtration and excretion homeostasis throughout the dosing and recovery phases. Both the synthetic and natural LYB007 candidate groups demonstrated an overall favorable urinary safety profile. Minor, transient shifts were occasionally observed, including isolated alterations in urine specific gravity (SG) and pH in the natural adjuvant high-dose male group (Day 30). These changes were of small magnitude, lacked dose-dependency, and were not corroborated by concurrent changes in renal clinical chemistry (e.g., creatinine, blood urea nitrogen) or renal histopathology.

Regarding qualitative parameters – including urine color, ketones, occult blood, protein, nitrite, and white blood cells – sporadic detections occurred across all cohorts, including the negative control and mock-vaccine groups. The occurrences were scattered and did not follow a specific dose-response pattern. Given the consistent findings in the control groups, these observations are attributed to the normal physiological variability and baseline fluctuations inherent in SD rat populations rather than any vaccine-induced renal impairment.

Organ weight, macroscopic, and microscopic observations

Organ weight changes, macroscopic postmortem findings, and microscopic histopathological findings were generally comparable between the synthetic and natural LYB007-adjuvanted vaccine studies (Table 3). Macroscopic findings associated with the administration of LYB007 were variably present and localized primarily to the injection sites (bilateral quadriceps femoris), characterized by focal dark red discolorations. These findings correlated with interstitial hemorrhage observed during microscopic evaluation. By the end of the recovery phase (Day 58), these macroscopic findings had fully resolved, indicating complete macroscopic recovery of the injection sites. Vaccine administration resulted in a consistent increase in absolute and relative (to body and brain weight) spleen weights in both sexes across both adjuvant cohorts at the end of the dosing phase (Day 30). Increases in spleen weight correlated with macroscopic splenomegaly. There were no test-substance-related significant weight changes in other organs that correlated with histopathological findings on Day 30 and Day 58. Microscopic findings of the injection sites include myofiber necrosis, edema, and mixed cell inflammation. This inflammation was characterized by an infiltrate of neutrophils, lymphocytes, and macrophages, admixed with interstitial hemorrhage and cellular debris. These findings were considered a physiological response to the potent immunostimulatory activity of the QS-21 adjuvant platform. At the end of the recovery phase (Day 58), there was significant attenuation of these findings; the inflammatory infiltrate shifted toward a more chronic phenotype primarily composed of lymphocytes and mononuclear cells, with a marked reduction in necrosis and hemorrhage, confirming the regenerative and self-limiting nature of the local reaction. Microscopically, increases in spleen weights are correlated with mild to moderate extramedullary hematopoiesis (EMH) in the red pulp, which is characterized by increased cellularity of differentiated myeloid cells. This systemic finding is an expected immunological adaptation to adjuvant-mediated immune activation and does not represent toxicological pathology.

Table 3.

Absolute and relative organ weights in SD rats following repeated administration of LYB007.

		CN n = 40	Synthetic Adjuvant Group			Natural Adjuvant Group
			CA n = 40	LYB007 (LD) n = 40	LYB007 (HD) n = 40	CA n = 40	LYB007 (LD) n = 40	LYB007 (HD) n = 40
Female
D30
Spleen weight (g)	Mean	0.5773	0.7343**	0.7866**	0.6842*	0.8246**	0.7549**	0.7004*
	Diff%	/	27.20↑	36.25↑	18.52↑	42.84↑	30.76↑	21.32↑
Spleen/BW ratio (%)	Mean	0.2297	0.2954**	0.3226**	0.2850**	0.3188**	0.2982**	0.2903**
	Diff%	/	28.60↑	40.44↑	24.07↑	38.79↑	29.82↑	26.38↑
Spleen/BR ratio (g/g)	Mean	0.2988	0.3851**	0.4057**	0.3556*	0.4318**	0.3841**	0.3777**
	Diff%	/	28.88↑	35.78↑	19.01↑	44.51↑	28.55↑	26.41↑
Liver weight (g)	Mean	6.6207	7.8509**	7.3993*	7.3132*	8.4934**	7.6073**	7.4334**
	Diff%	/	18.58↑	11.76↑	10.46↑	28.29↑	14.90↑	12.28↑
Liver/BW ratio (%)	Mean	2.6382	3.1459**	3.0408**	3.0443**	3.2790**	3.0116**	3.0763**
	Diff%	/	19.24↑	15.26↑	15.39↑	24.29↑	14.15↑	16.61↑
Liver/BR ratio (g/g)	Mean	3.4262	4.1175**	3.8231*	3.7978*	4.4471**	3.8752**	4.0153**
	Diff%	/	20.18↑	11.58↑	10.85↑	29.80↑	13.10↑	17.19↑
D58
Spleen/BW ratio (%)	Mean	0.1771	/	/	/	0.2161*	0.1882	0.1801
	Diff%	/	/	/	/	22.02↑	6.27↑	1.69↑
Uterine weight (g)	Mean	0.5033	/	/	/	0.7546**	0.5320	0.5634
	Diff%	/	/	/	/	49.93↑	5.70↑	11.94↑
Uterine/BW ratio (%)	Mean	0.1750	/	/	/	0.2717*	0.1824	0.1898
	Diff%	/	/	/	/	55.26↑	4.23↑	8.46↑
Uterine/BR ratio (g/g)	Mean	0.7278	/	/	/	1.0718**	0.7911	0.8512
	Diff%	/	/	/	/	47.27↑	8.70↑	16.96↑
Male
D30
Heart/BW ratio (%)	Mean	0.3242	0.3544**	0.3370	0.3417	/	/	/
	Diff%	/	9.32↑	3.95↑	5.40↑	/	/	/
Spleen weight (g)	Mean	0.7494	0.9547**	0.9105*	0.8713	0.9460**	0.8290	0.9228*
	Diff%	/	27.40↑	21.50↑	16.27↑	26.23↑	10.62↑	23.14↑
Spleen/BW ratio (%)	Mean	0.1912	0.2572**	0.2423**	0.2432**	0.2575**	0.2314*	0.2481**
	Diff%	/	34.52↑	26.73↑	27.20↑	34.68↑	21.03↑	29.76↑
Spleen/BR ratio (g/g)	Mean	0.3658	0.4573**	0.4432*	0.4250	0.4652**	0.4076	0.4503*
	Diff%	/	25.01↑	21.16↑	16.18↑	27.17↑	11.43↑	23.10↑
Liver/BW ratio (%)	Mean	2.8188	3.0645**	3.0017	2.9434	/	/	/
	Diff%	/	8.72↑	6.49↑	4.42↑	/	/	/
Adrenal gland/BW ratio (%)	Mean	0.0131	0.0159**	0.0167**	0.0158**	0.0156**	0.0175**	0.0155**
	Diff%	/	21.37↑	27.48↑	20.61↑	19.08↑	33.59↑	18.32↑
D58
Spleen/BW ratio (%)	Mean	0.1464	0.1871*	0.1686	0.1824*	0.1589	0.1824**	0.1679
	Diff%	/	27.80↑	15.16↑	24.59↑	8.54↑	24.59↑	14.69↑

Data are expressed as mean and percentage difference (Diff%) compared to the CN group. BW, body weight; BR, brain weight; CN, negative control; CA, mock vaccine; LYB007 (LD/HD), low-dose and high-dose groups. n = 40/group (20 males and 20 females) for each adjuvant cohort and the shared control (CN) group. **P ≤ .01, *P ≤ .05, indicate significant differences compared to the CN group. ↑indicate an increase relative to the CN group.

Lymphocyte subsets and cytokines

Lymphocyte subset analysis (specifically CD3+ T-cell populations) and serum cytokine profiles (IFN-γ, IL-2, TNF-α, IL-4) showed consistent patterns across both adjuvant cohorts, while no detectable levels of IFN-γ and IL-2 were found in any rats (Figure 7). Minor, statistically significant fluctuations in CD3+ percentages and cytokine levels were noted transiently during the dosing phase (Day 30).

Figure 7.

Immunophenotypic profile and serum cytokine kinetics following repeated administration of LYB007. (A–C) flow cytometric analysis of peripheral blood T cell subpopulations, including CD3+ T cells, CD4+ T cells (CD4+), CD8+ T cells at Day 30 and Day 58, and calculated CD4+/CD8+ ratios across study groups (D). (E–F) Serum concentrations of cytokine TNF-α and IL-4 measured by MSD assay. CN, negative control; CA, mock vaccine; LYB007 (LD/HD), low-dose and high-dose groups. n = 40/group for each adjuvant cohort and the shared control (CN) group.

These changes, characterized by low-level shifts in circulating cytokines and T-cell populations, are considered typical markers of systemic immune mobilization following potent adjuvant exposure. The lack of dose-dependency and the absence of correlating pathological clinical signs confirm these fluctuations are transient physiological adaptations rather than indications of immunotoxicity.

Antigen-specific humoral response

The induction of antigen-specific IgG antibodies served as the primary measure of adjuvant efficacy. Crucially, all rats in the negative control and mock-vaccine control groups remained seronegative for antigen-specific IgG antibodies throughout the study (Table 4). This absence of background response in control cohorts underscores the specificity of the humoral response induced by the LYB007-adjuvanted vaccines.

Table 4.

Seroconversion rates following repeated immunizations with LYB007.

Groups	Synthetic adjuvant					Natural adjuvant
	D1	D15	D29	D43	D58	D1	D15	D29	D43	D58
CN	0/10	0/10	0/10	0/10	0/10	0/10	0/10	0/10	0/10	0/10
CA	0/10	0/10	0/10	0/10	0/10	0/10	0/10	0/10	0/10	0/10
LYB007 (LD)	0/10	10/10	10/10	10/10	10/10	0/10	10/10	10/10	10/10	10/10
LYB007 (HD)	0/10	10/10	10/10	10/10	10/10	0/10	10/10	10/10	10/10	10/10

CN, negative control; CA, mock vaccine; LYB007 (LD/HD), low- and high-dose adjuvant groups. Measured in satellite groups (n = 10/group).

In contrast, both the synthetic and natural LYB007 groups demonstrated robust and equivalent seroconversion. All treated rats (10/10) achieved seroconversion (Table 4). The profiling of serum antigen-specific antibody titers illustrated in Figure 8 demonstrates that detectable antibody titers emerged as early as Day 15, and peak geometric mean titers (GMT) were consistently attained between Day 29 and Day 43, indicating a rapid and potent primary immune response. By the end of the recovery phase, although antibody titers declined from the peak observed on Day 43, they remained elevated, consistent with the expected profile of a durable humoral response.

Figure 8.

Serum antigen-specific antibody titers induced by LYB007. (A) Antigen-specific antibody level of natural adjuvant group. (B) Antigen-specific antibody level of synthetic adjuvant group. Data are presented as geometric mean titers (GMT) for each group at indicated time points, with scatters showing individual antibody titers per animal. CN, negative control; CA, mock vaccine; LYB007 (LD/HD), low-dose and high-dose groups. n = 40/group.

Discussion

The nonclinical safety evaluation program for the LYB007 recombinant tuberculosis vaccine candidate (as described in our previous study²⁴) demonstrates that both synthetic and natural QS-21-based adjuvant formulations are well-tolerated in SD rats when administered IM in a single 80ug dose or repeated 10 μg/40 μg every two weeks (3 total doses). The two adjuvant formulations showed closely comparable profiles across all assessed endpoints, supporting the synthetic QS-21-Api as a viable and functionally equivalent alternative to the naturally derived standard.

The administration of LYB007 elicited localized and systemic reactogenicity consistent with the potent immunostimulatory profile of the QS-21 adjuvant platform. Injection site reactions – manifesting as transient myofiber necrosis, edema, and mixed-cell inflammatory infiltrates – are expected sequelae of amphiphilic saponin-based adjuvants.^26–28 The severity and incidence of these findings, which were statistically comparable between synthetic and natural cohorts, were self-limiting. The complete resolution of injection site edema and the shift toward a chronic, regenerative inflammatory phenotype by the end of the 58 d recovery period provide strong evidence for the lack of long-term tissue damage or fibrotic sequelae.

The systemic inflammatory markers observed – such as the rapid elevation of Fbg and the sustained consumption of complement C4–were particularly notable during the dosing phase (Day 30). These shifts are consistent with anticipated physiological responses to adjuvant-mediated immune stimulation and were not accompanied by signs of pathological over-activation or cytokine storm.^29–31 The reduction in complement C4 concentrations strongly correlates with the concurrent rise in serum IgG levels, providing direct biochemical evidence that the adjuvant successfully promotes activation of the humoral immune system.^32,33 Importantly, these markers returned to baseline levels by the recovery phase (Day 58), demonstrating the response’s transient nature. Furthermore, histopathological observations, including splenic expansion and extramedullary hematopoiesis, underwent an orderly transition from an active, proliferative phase to a remodeling phase. The orderly resolution of these histopathological changes suggests that the synthetic adjuvant achieves potent immunostimulation without inducing disproportionate or irreversible tissue responses.

The immunological data provide the most compelling evidence for functional equivalence. The strict seronegativity in the negative and mock-vaccine control groups throughout the study (D1–D58) confirms that humoral responses were exclusively driven by the LYB007-adjuvant/antigen complex. Both formulations induced robust, uniform (100%) seroconversion with identical kinetics; peak geometric mean titers (GMT) were consistently attained between Day 29 and Day 43. This precise temporal alignment indicates that the synthetic QS-21-Api successfully replicates the potency of the natural counterpart in modulating B-cell maturation and plasma cell expansion, facilitating the formation of durable humoral immune memory.

The requirement for robust Th1-biased immunity is a defining challenge in tuberculosis vaccine development. M. tuberculosis is an intracellular pathogen that persists within macrophages, necessitating a sustained cell-mediated response – coordinated primarily by T-lymphocytes – alongside potent humoral immunity to prevent progression to active disease.^34,35 Our analysis of lymphocyte subsets revealed that the adjuvanted LYB007 vaccine induced balanced T-cell mobilization. Although we observed transient fluctuations in circulating CD3+ T-cell populations during the dosing phase, these shifts were consistent with the systemic immune mobilization typical of potent adjuvant engagement.³ In the context of tuberculosis, the ability of the adjuvanted LYB007 to expand CD3+ T-cell subsets is critical; these cells serve as the foundational reservoir for antigen-specific T-cell responses necessary to control mycobacterial replication. Unlike traditional subunit vaccines, which often fail to recruit sufficient T-cell help due to the lack of intrinsic danger signals and the rapid clearance of soluble antigen,^36–38 LYB007, formulated with both adjuvants, could elicit robust immune responses. These results indicate that the synthetic formulation elicited T-cell responses comparable to those induced by the natural standard, supporting its potential as a component of effective tuberculosis vaccines.

Consistent with the adaptive immunity required for tuberculosis vaccine efficacy, LYB007 administration triggered systemic immune mobilization. Macroscopic enlargement of the spleen, histopathologically correlated with increased germinal center activity and extramedullary hematopoiesis in the red pulp, reflects the physiological expansion of myeloid and lymphoid lineages.^39,40 These systemic adaptations are not indicators of toxicological insult but rather the desired pharmacological manifestations of effective adjuvant-mediated immune engagement.

While our findings provide robust evidence for the toxicological and functional equivalence of the synthetic QS-21-Api, several limitations warrant consideration. First, the study utilized the SD rats as a primary nonclinical model. Although rats are highly suitable for detecting adjuvant-induced histopathological and hematological changes, they may exhibit species-specific differences in immune cell mobilization and cytokine sensitivity compared to humans. Future studies in non-human primates or humanized mouse models could provide further insights into the cross-species translational potential of the adjuvanted LYB007 vaccine. Second, while we observed long-term immune persistence through Day 58, the observation period was limited to a single recovery phase. Longitudinal studies assessing the durability of antibody responses beyond this window would be beneficial to confirm the stability of the memory response. Lastly, the current study focused on systemic and local safety profiles using a single adjuvant-to-antigen ratio; exploring the safety margin at higher concentrations or alternative administration schedules could further define the therapeutic index of this synthetic platform. Finally, as this study focuses on the safety of the vaccine and adjuvant, there is a lack of a comprehensive antigen-specific immunological readout, particularly for the cellular response. Therefore, we are undertaking a series of immunological studies to fully assess the immunological similarities and differences (if any) between the two adjuvant formulations.

The transition to a synthetic adjuvant platform addresses critical vulnerabilities in the traditional QS-21 supply chain, including ecological sensitivity and the risk of impurities. By enabling precise control over the triterpene core and acyl chain structure, synthetic production allows for rational design – the decoupling of potent immunostimulation from dose-limiting toxicities. In summary, the synthetic QS-21-Api demonstrated a safety and immunogenicity profile equivalent to the naturally derived reference. These results support its use as a chemically defined, quality-controlled adjuvant component in recombinant subunit vaccine development.

Biography

Yan Liu, Ph.D. Associate Professor of the College of Pharmacy, Chongqing Medical University, China. She serves as an external expert for the Chongqing Drug Technical Review and Inspection Center, a member of the Expert Panel for Investigation and Diagnosis of Adverse Reactions Following Immunization in Chongqing, and a health impact assessment expert for Banan District, Chongqing. Her research interests focus on clinical pharmacology and new drug registration. Over the past five years, she has published 9 SCI-indexed papers and 2 articles in Chinese core journals, and holds 3 authorized invention patents. She teaches several pharmacy courses, including Pharmacist Communication, Introduction to Clinical Pharmacy, and Pharmacotherapeutics. She has also contributed to the compilation of multiple textbooks, including Introduction to Basic Medicine, A Concise Reader of Pharmaceutical Sciences, and New Drug Evaluation and Registration.

Funding Statement

The author(s) reported there is no funding associated with the work featured in this article.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data supporting the findings of this study are available from the corresponding author, Yan Liu, upon reasonable request.