An anti-TSLP monoclonal antibody for uncontrolled CRSwNP: the DUBHE randomized clinical trial

Abstract

To explore the therapeutic potential of blocking thymic stromal lymphopoietin (TSLP) in patients with chronic rhinosinusitis with nasal polyps (CRSwNP), we conducted a phase 1b/2a, randomized, double-blind, placebo-controlled trial to assess the safety and efficacy of CM326, a monoclonal antibody against TSLP. We enrolled 84 eligible patients with uncontrolled CRSwNP and stratified them based on baseline tissue eosinophil count. Patients are assigned to receive CM326 220 mg (n = 40) or placebo (n = 20) every 2 weeks (Q2W) and CM326 220 mg (n = 20) or placebo (n = 4) every 4 weeks (Q4W) for 16 weeks. Subsequently, all patients continue on CM326 220 mg Q2W or Q4W for an additional 36 weeks, followed by a 12-week follow up. Primary endpoints are safety of CM326 and change from baseline in NPS at week 16 in patients with eosinophilic CRSwNP (ECRSwNP). Main secondary endpoints include the change from baseline in NPS at week 16 in non-eosinophilic CRSwNP (nonECRwNP) and pharmacodynamic markers. Throughout the 64-week study, all treatment-emergent adverse events (TEAEs) are mild or moderate. CM326 Q2W improves NPS in patients with ECRSwNP compared with placebo at week 16 (mean difference [95% CI], −1.2 [−2.3 to −0.1], P = 0.04), with sustained benefits during the open-label and follow-up periods. Notably, peripheral blood and tissue eosinophil counts and concentrations of plasma IL-13 and IL-5 are reduced by week 16 with the treatment of CM326 Q2W versus placebo. A post-hoc analysis demonstrates that all participants with baseline TSLP > 330 fg/mL achieve a substantial reduction in NPS by week 16 with the treatment of CM326 Q2W (mean difference vs. placebo: −1.75 [95%CI, −3.06 to −0.44], P = 0.01). Overall, CM326 is well tolerated and effective in patients with uncontrolled ECRSwNP. A baseline plasma TSLP level of 330 fg/mL may serve as a predictive marker for treatment efficacy of CM326. ClinicalTrials.gov Identifier: NCT05324137.

Anti-TSLP biologics have shown promise for the clinical treatment of type 2 respiratory disease. Here the authors present a clinical trial of CM326, a monoclonal antibody against TSLP, and show the clinical effects in patients with eosinophilic chronic rhinosinusitis with nasal polyps (ECRSwNP), who present with type 2-dominant inflammation.

Introduction

Chronic rhinosinusitis with nasal polyps (CRSwNP) is a chronic inflammation of nasal mucosa and paranasal sinuses, affecting ~2–4% of the population worldwide^1,2. Patients with CRSwNP suffer from symptom burdens such as nasal congestion, rhinorrhea, headache, and loss of smell, all of which reduce their quality of life^3,4. A large proportion of patients with CRSwNP remain uncontrolled despite adequate medical (topical or systemic corticosteroid, SCS) and (or) surgical management⁵. Liao et al. have demonstrated that uncontrolled CRSwNP comprises distinct inflammatory endotypes, including a subset of patients exhibiting eosinophilic inflammation as reflected by elevated interleukin (IL)-4/IL-5 levels and tissue eosinophil infiltration, and a subset displaying neutrophilic inflammation, among others⁶. In our histopathological studies, we have established ≥55 eosinophils/HPF as the diagnostic threshold for eosinophilic CRSwNP (ECRSwNP), a criterion that has also demonstrated clinical utility in identifying uncontrolled status of disease⁷.

Currently, for treatment of uncontrolled CRSwNP, biologics have been developed to neutralize specific targets such as immunoglobulin (Ig) E, IL−4, IL-5, and IL-13, or to block their receptors, without affecting the upstream regulation of these molecules. Thymic stromal lymphopoietin (TSLP) is an epithelial cell-derived cytokine that is synthesised in response to environmental and proinflammatory stimuli⁸. After binding to its receptor, TSLP initiates and amplifies type 2 immune responses by activating dendritic cells, group 2 innate lymphoid cells, and T helper (Th) 2 cells, and upregulating the production of type 2 cytokines such as IL-4, IL-5, IL-9 and IL-13⁹. In asthma, tezepelumab, a monoclonal antibody targeting TSLP, has been shown to prevent exacerbations in patients with high and low Th2 phenotypes, suggesting that blocking TSLP may have a broader physiological effect than just suppressing specific type 2 cytokine signals^10,11. To date, the clinical evidence for TSLP blockade in treating patients with uncontrolled CRSwNP remains under investigation.

CM326 is a humanised monoclonal antibody targeting TSLP with a unique complementarity-determining region (CDR) molecular sequence. Furthermore, CM326 has demonstrated a well-tolerated safety profile, linear pharmacokinetics over the dose range of 22 to 220 mg, and low immunogenicity in healthy adults across two phase 1 studies (NCT04842201, NCT05171348).

To assess the safety and efficacy of CM326 in patients with uncontrolled CRSwNP, we conducted the DUBHE trial, a 16-week phase 1b/2a, randomized, double-blind, placebo-controlled trial, followed by a 36-week open-label extension period and a 12-week follow-up period. To identify the underlying indication of CM326, patients enrolled in the study were stratified by nasal polyp tissue eosinophil count. The efficacy of CM326 was then evaluated within each stratum.

Results

Participants

Between July 11, 2022, and February 13, 2023, 84 eligible patients were enrolled and randomized. Sixty patients were assigned to CM326 (n = 40, 19 ECRSwNP vs. 21 non-eosinophilic CRSwNP [nonECRSwNP]) or placebo (n = 20, 10 ECRSwNP vs. 10 nonECRSwNP) every 2 weeks (Q2W) and 24 to CM326 (n = 20, 10 ECRSwNP vs. 10 nonECRSwNP) or placebo (n = 4, 2 ECRSwNP vs. 2 nonECRSwNP) every 4 weeks (Q4W). As of the end of the study (March 29, 2024), 80 out of 84 (95.2%) patients completed the 16-week double-blind treatment, and 73 out of 80 (91.3%) patients completed the 36-week open-label treatment (Fig. 1). The median age of the participants was 43.5 (interquartile range, 34.0–55.0) years, 71.4% (n = 60) were male, and 28.6% (n = 24) were female. Fifty-six percent (n = 47) had history of nasal polyp surgery, and 61.9% (n = 52) had history of SCS use. Additionally, 33.3% (n = 28) had comorbid asthma. Demographics and baseline characteristics for each group are summarized in Table 1.

Fig. 1. Patient flow in the DUBHE randomized clinical trial.

* Four enrolled patients in the open-label cohort were not included in further analysis of this study. The cohort of every two weeks was stratified by tissue eosinophil count ( < 55/HPF, ≥55/HPF) and NPS ( < 5, ≥5). The cohort of every four weeks only included patients with NPS ≥ 5 and was stratified by tissue eosinophil count ( < 55/HPF, ≥55/HPF). HPF high-power field; NPS nasal polyp score; Q2W every two weeks; Q4W every four weeks; FAS full analysis set; SS safety set; PDS pharmacodynamic set; IMGS immunogenicity set; mFAS modified full analysis set.

Table 1.

Patient baseline demographics and clinical characteristics

	CM326 Q2W (N = 40)	Placebo Q2W (N = 20)	CM326 Q4W (N = 20)	Placebo Q4W (N = 4)	Total (N = 84)
Age, median (IQR), yr	45.5 (35.5–55.5)	43.5 (38.0–48.5)	34.0 (29.0–47.5)	59.0 (49.0-62.5)	43.5 (34.0–55.0)
Sex, No. (%)
Male	31 (77.5%)	12 (60.0%)	14 (70.0%)	3 (75.0%)	60 (71.4%)
Female	9 (22.5%)	8 (40.0%)	6 (30.0%)	1 (25.0%)	24 (28.6%)
Race, No. (%)
Han	40 (100%)	20 (100%)	19 (95.0%)	4 (100%)	83 (98.8%)
Other	0	0	1 (5.0%)	0	1 (1.2%)
BMI, kg/m2	24.16 (2.79)	23.67 (3.12)	26.14 (4.40)	24.45 (3.63)	24.53 (3.42)
With nasal polyp surgery history, No. (%)	24 (60.0%)	10 (50.0%)	11 (55.0%)	2 (50.0%)	47 (56.0%)
With SCS use history, No. (%)	23 (57.5%)	13 (65.0%)	14 (70.0%)	2 (50.0%)	52 (61.9%)
Duration of CRSwNP, median (IQR), yr	4.0 (1.0-13.0)	4.5 (0.5-9.5)	2.5 (0.0-7.5)	4.5 (0.0–9.5)	4.0 (0.0-10.0)
Blood eosinophils,109 cells/L	0.314 (0.217)	0.333 (0.178)	0.285 (0.224)	0.228 (0.122)	0.307 (0.205)
ECRSwNP, No. (%) a	19 (47.5%)	10 (50.0%)	10 (50.0%)	2 (50.0%)	41 (48.8%)
Tissue eosinophilic count, cells/HPF	104 (147)	78 (77)	105 (120)	129 (169)	99 (126)
Comorbid asthma, No. (%)	15 (37.5%)	4 (20.0%)	7 (35.0%)	2 (50.0%)	28 (33.3%)
Bilateral nasal polyp score, scale 0-8 b	4.9 (1.4)	4.7 (1.4)	5.9 (0.8)	5.3 (0.5)	5.1 (1.3)
Nasal congestion score, scale 0-3 b	2.30 (0.45)	2.34 (0.54)	2.41 (0.45)	2.00 (0.47)	2.32 (0.47)
Lund-Mackay CT score, scale 0-24 b	18.4 (3.3)	17.3 (3.8)	18.5 (3.7)	14.5 (7.9)	18.0 (3.8)
Total symptom score, scale 0-9 b	6.32 (1.49)	5.93 (1.6)	6.90 (1.13)	5.77 (1.79)	6.34 (1.47)
SNOT-22 score, scale 0-110 b	41.2 (19.1)	44.5 (22.0)	42.4 (19.4)	33.3 (17.9)	41.9 (19.7)
UPSIT score, scale 0-40 c	14.0 (7.1)	16.7 (9.6)	12.2 (7.7)	12.8 (4.1)	14.1 (7.9)

Data are based on full analysis set and presented as mean (standard deviation) unless otherwise noted. Analyses were performed in the full analysis set, which included all participants who received at least one dose of study drug after randomization on an intent-to-treat basis.

^aECRSwNP, defined as the tissue eosinophil count ≥55 eosinophils/high-power field in nasal polyp biopsies within the screening period.

^bhigher scores indicate greater disease severity or worse quality of life.

^chigher scores indicate lower disease severity and a score of 35–40 denotes a normal sense of smell. Q2W every 2 weeks; Q4W every 4 weeks; yr years; ECRSwNP eosinophilic chronic rhinosinusitis with nasal polyps; BMI body mass index; SCS systemic corticosteroid; HPF high-power field; SNOT-22 22-item Sino-nasal Outcome Test; UPSIT University of Pennsylvania Smell Identification Test; IQR interquartile range.

Safety and Immunogenicity

CM326 was safe and well tolerated in both the Q2W and Q4W cohorts. No clinically relevant changes were observed in vital signs, physical examinations, electrocardiograms, or clinical laboratory tests. All treatment-emergent adverse events (TEAEs) occurred throughout the entire study (week 64) were mild or moderate. During the 16-week double-blind period, the overall incidence of TEAEs was comparable across the treatment groups. The most common TEAEs were COVID-19 and upper respiratory tract infection (Supplementary Table 1).

Throughout the entire study period, five patients in the Q2W cohort and one patient in the Q4W cohort experienced treatment-related TEAEs (Table 2), all of which were mild and resolved. Serious TEAEs were reported in one patient in the CM326 Q2W group (unilateral renal calculus and knee ligament injury) and two in the CM326 Q4W group (one with community acquired pneumonia and one with ureterolithiasis), which occurred during the open-label period and were deemed unrelated to treatment. Two patients in the CM326 Q2W group discontinued treatment due to TEAEs: one had treatment-related allergic dermatitis in the double-blind period and the other had treatment-unrelated primary biliary cirrhosis during the open-label period. No patient was detected with positive anti-drug antibody throughout the study, indicating low immunogenicity of CM326.

Table 2.

Treatment-emergent adverse events occurred during the entire study period of 64 weeks

	CM326 Q2W (N = 40 c)	Total CM326 Q2W exposure (N = 58 d)	CM326 Q4W (N = 20 c)	Total CM326 Q4W exposure (N = 24 d)
Any TEAEs a	37 (92.5%)	52 (89.7%)	17 (85.0%)	20 (83.3%)
Drug-related TEAEs	4 (10.0%)	5 (8.6%)	1 (5.0%)	1 (4.2%)
Any serious TEAEs	1 (2.5%)	1 (1.7%)	2 (10.0%)	2 (8.3%)
Any TEAEs leading to death	0	0	0	0
Any TEAEs leading to permanent treatment discontinuation	2 (5.0%)	2 (3.4%)	0	0
TEAEs occurring in ≥ 3 cases in either total CM326 Q2W or Q4W exposure group b
COVID-19	21 (52.5%)	22 (37.9%)	4 (20.0%)	4 (16.7%)
Upper respiratory tract infection	10 (25.0%)	15 (25.9%)	7 (35.0%)	8 (33.3%)
Nasopharyngitis	6 (15.0%)	9 (15.5%)	2 (10.0%)	2 (8.3%)
Suspected COVID-19	4 (10.0%)	4 (6.9%)	2 (10.0%)	2 (8.3%)
Pharyngitis	2 (5.0%)	3 (5.2%)	0	0
Periodontitis	1 (2.5%)	3 (5.2%)	1 (5.0%)	1 (4.2%)
Alanine aminotransferase increased	3 (7.5%)	3 (5.2%)	2 (10.0%)	3 (12.5%)
Aspartate aminotransferase increased	1 (2.5%)	1 (1.7%)	2 (10.0%)	3 (12.5%)
Blood uric acid increased	2 (5.0%)	3 (5.2%)	0	0
Epistaxis	2 (5.0%)	3 (5.2%)	0	0
Oropharyngeal pain	2 (5.0%)	3 (5.2%)	0	0
Mouth ulceration	2 (5.0%)	3 (5.2%)	0	0
Pruritus	2 (5.0%)	3 (5.2%)	0	0

Data are presented as number of patients (%).

^aTreatment-emergent adverse events (TEAEs) were defined as adverse events that emerged after the first treatment or those happened before but exacerbated after first administration;

^bAccording to Medical Dictionary for Regulatory Activities preferred terms.

^cIncluded patients who received ≥1 dose of CM326 during the double-blind period, and TEAEs during the entire 64 weeks are shown.

^dIncluded patients who received ≥ 1 dose of CM326 during the double-blind period and patients who switched from placebo to CM326 during the open-label period. The sum of the TEAEs in the former during the entire 64 weeks and the TEAEs in the latter during the open-label and follow-up periods are shown. COVID-19 coronavirus disease; Q2W every 2 weeks; Q4W every 4 weeks.

Primary and secondary efficacy outcomes

For the primary efficacy endpoint, the improvement in endoscopic bilateral endoscopic nasal polyp score (NPS) at week 16 in patients with ECRSwNP was greater with CM326 treatment than with placebo in the Q2W cohort (mean difference, −1.2 [95% CI, −2.3 to −0.1], P = 0.04) (Fig. 2a, and Supplementary Fig. 1). The improvement was observed at as early as the first assessment (week 4) (Fig. 2a). The mean reduction of NPS among patients with ECRSwNP who were receiving CM326 compared to those receiving placebo in the Q4W cohort was −1.1 (95% CI, −3.1, 0.9), but the difference was not statistically significant (P = 0.24, Fig. 2b). For the secondary efficacy endpoint, improvement in NPS was not observed in patients with nonECRSwNP at week 16 (Fig. 2c, d).

Fig. 2. Change from baseline over time in NPS in patients with ECRSwNP or nonECRSwNP.

a, b Change from baseline over time in NPS in patients with eosinophilic chronic rhinosinusitis with nasal polyps (ECRSwNP) in the Q2W (a) and Q4W cohort (b); c, d Change from baseline over time in NPS in patients with non-eosinophilic chronic rhinosinusitis with nasal polyps (nonECRSwNP) in the Q2W (c) and Q4W cohort (d). Error bars denote standard error. Baseline values are presented as mean (standard deviation). NPS nasal polyp score; Q2W every 2 weeks; Q4W every 4 weeks. Group differences within 16 weeks were tested using t-test. P values were two sided and nominal, without adjustments for multiplicity. No between comparison was performed during the open-label and follow-up periods as all patients received CM326 during the open-label period.

Exploratory efficacy outcomes

Continued use of CM326 Q2W and Q4W resulted in sustained NPS improvement in patients with ECRSwNP during the open-label period (Fig. 2a, b). For those who switched from placebo to CM326, a prompt reduction was observed in NPS, which resembled that seen in the CM326 arm during the double-blind period. Notably, the efficacy was also maintained for up to 12 weeks after treatment discontinuation (week 64). However, continued use of CM326 beyond week 16 appeared to have no effect on reducing NPS in patients with nonECRSwNP (Fig. 2c, d).

Patients with ECRSwNP receiving CM326 Q2W also showed continuous improvements over time compared with placebo within 16 weeks. This was evident in improved health-related quality of life, as reflected by a decreased 22-item Sino-Nasal Outcome Test (SNOT-22) score (mean difference at week 16, −12.5 [95% CI, −23.2 to −1.8], P = 0.02); reduced sinus inflammation, as reflected by a decreased Lund-Mackay CT score (mean difference at week 16, −5.1 [95% CI, −9.1 to −1.2], P = 0.01); improved overall nasal symptoms, as reflected by a decreased total symptom score (TSS, mean difference at week 16, −1.14 [95% CI, −2.32 to 0.04], P = 0.06); and enhanced sense of smell, as reflected by an increased University of Pennsylvania Smell Identification Test (UPSIT) score (mean difference at week 16, 4.7 [95% CI, −0.4 to 9.8], P = 0.07) (Supplementary Fig. 2). These improvements were maintained during the open-label period and persisted up to 12 weeks after treatment discontinuation (Supplementary Fig. 2). In contrast, no notable improvements in any of the efficacy endpoints were observed in patients with nonECRSwNP (Supplementary Fig. 3). Exploratory efficacy outcomes of patients with ECRSwNP and nonECRSwNP in the Q4W cohort are shown in Supplementary Figs. 4 and 5 respectively.

Pharmacodynamic (PD) markers

Compared with placebo, CM326 Q2W treatment led to a decrease in peripheral blood eosinophil count (10⁹cells/L) during the double-blind period (mean difference: −0.13 [95% CI, −0.21 to −0.04], P = 0.004 at week 16), and this decrease was sustained throughout the open-label and follow-up periods (Fig. 3a). At week 16, the tissue eosinophil count (cells/ high-power field [HPF]) also showed a greater reduction from baseline in the CM326 Q2W group versus placebo group (mean difference: −81.10 [95%CI, −171.43 to 9.24], P = 0.047, Fig. 3b, c). CM326 Q2W treatment decreased the plasma concentration of the type 2 cytokines IL-5 and IL-13 (fg/mL) compared with placebo at week 16 (mean difference: −527.76 [95% CI, −1136.19 to 80.66], P = 0.011; −23.79 [95%CI, −43.53 to −4.06], P = 0.044) (Fig. 3d, e). However, there were no clear differences between CM326 and placebo in terms of concentrations of serum total IgE, serum thymus and activation-regulated chemokine (TARC), and serum periostin throughout the study in the Q2W cohort (Fig. 3f–h). The concentration of plasma free TSLP (fg/mL) significantly decreased within 16 weeks in the CM326 Q2W group versus placebo (mean difference: −336.94 [95%CI, −429.27 to −244.61], P < 0.0001). The reduction was sustained throughout 52 weeks of treatment and rebounded after treatment discontinuation (Fig. 3i). PD marker changes for the Q4W cohort are shown in Supplementary Fig. 6.

Fig. 3. Change from baseline in pharmacodynamic biomarkers in the Q2W cohort.

a Peripheral blood eosinophil count; b Tissue eosinophil count; c Hematoxylin and eosin stain of biopsied nasal polyp tissue at baseline and week 16 from a patient (magnification: ×400); d Plasma interleukin-5 (IL-5); e Plasma interleukin-13 (IL-13); f Total serum immunoglobulin E (IgE); g Serum thymus and activation-regulated chemokine (TARC); h Serum periostin; i Plasma thymic stromal lymphopoietin (TSLP). HPF high-power field; Q2W every 2 weeks. Error bars denote standard error. Baseline values are presented as mean (standard deviation). Biomarkers were analyzed under pharmacodynamic set using Wilcoxon rank-sum test. P values were two sided and nominal, without adjustments for multiplicity. No between comparison was performed during the open-label and follow-up periods as all patients received CM326 during the open-label period.

Biomarker predictive of response to CM326 treatment

Next, an attempt was made to roughly identify the suitable patients for CM326 treatment within the Q2W cohort irrespective of tissue eosinophil level. This was achieved through a post-hoc analysis that utilized baseline plasma TSLP levels as a biomarker. In the end, the area under the curve (AUC) of the receiver operating characteristic (ROC) curve turned out to be 0.77 (Supplementary Fig. 7). The optimal cut-off was chosen as 330 fg/ml, with a specificity of 0.71 and a sensitivity of 0.89. In the Q2W cohort, patients with high baseline TSLP ( > 330 fg/mL) achieved a substantial reduction in NPS at week 16 (mean difference vs. placebo: −1.75 [95%CI, −3.06 to −0.44], P = 0.01), while those with low baseline TSLP ( ≤ 330 fg/mL) did not show obvious improvement (mean difference vs. placebo: 0.03 [95%CI, −0.73 to 0.79], P = 0.93) (Supplementary Table 2). Additionally, patients with high baseline plasma TSLP level also showed greater improvements in different efficacy endpoints, including NCS, TSS, UPSIT score, SNOT-22 score, and Lund-Mackay CT score. A greater magnitude of CM326 effect on PD markers, such as plasma IL-13 and eosinophil counts in peripheral blood and tissue, was also observed in patients with higher baseline TSLP (Supplementary Table 3). Patients with baseline plasma TSLP ≤ 330 fg/mL showed no significant improvement in either the efficacy endpoints or PD endpoints (Supplementary Tables 2–3).

Discussion

We report the safety and efficacy of CM326, a monoclonal antibody targeting TSLP, in patients with uncontrolled CRSwNP. In this study, the addition of CM326 (220 mg, Q2W) to mometasone furoate nasal spray showed a favourable safety profile and resulted in improvements in NPS, SNOT-22 score, and Lund-Mackay CT score at week 16 in patients with ECRSwNP. The efficacy of CM326 was maintained for up to 52 weeks and did not diminish within 12 weeks after treatment discontinuation. Compared with the placebo group, CM326 treatment also reduced peripheral blood and tissue eosinophil counts.

In vitro comparative analysis demonstrated that CM326 was more effective than tezepelumab at suppressing Th2-driven inflammation, as evidenced by reduced JAK/STAT5 activation-induced cell proliferation and TARC release¹². In pulmonary allergic cynomolgus monkey model challenged with A. suum, CM326 improved lung function and restored the inflammatory microenvironment toward normal levels¹². CM326 is also currently being investigated in asthma (NCT05774340) and atopic dermatitis (NCT05671432; NCT05671445).

Our study has demonstrated the safety and tolerability of CM326 in patients with CRSwNP. All TEAEs occurred throughout the entire study (week 64) were mild or moderate. The most common TEAE observed in all trial groups was COVID-19 infection, which is largely attributed to the pandemic of COVID-19 in China during the treatment period. Besides COVID-19 and suspected COVID-19, upper respiratory tract infection occurred in 23 patients during the total CM326 exposure, consistent with common AEs in asthma patients treated with tezepelumab¹³.

Meanwhile, our study provides clinical evidence on the efficacy of CM326 (220 mg, Q2W) treatment for patients with ECRSwNP. In patients with ECRSwNP, the mean decrease from baseline in NPS at week 16 in the CM326 Q2W group versus the control group of 1.2 points exceeds the criterion (1 point) suggested by position paper to define a response to biologic treatment^2,5. The improvement in NPS by CM326 is comparable with dupilumab, mepolizumab, and omalizumab treatment, all of which target specific type 2 inflammatory biomarkers^14–16. Moreover, subgroup analysis revealed that patients, regardless of nasal polyp surgery or comorbid asthma, show a trend of NPS reduction, while patients with previous nasal polyp surgery might have numerically greater NPS reduction than those without (Supplementary Fig. 8). In addition, the improvement in SNOT-22 score of −12.5 (95% CI, −23.2 to −1.8), though moderate relative to the baseline value of 43.9, is greater than the minimal clinically important difference of 8.9 points, indicating improved quality of life that can be experienced by the patient. The slight improvement in Lund-Mackay CT score of −5.1 (95% CI, −9.1 to −1.2) is regarded clinically significant considering the baseline value of 18.6. Given that the sample size of the Q4W cohort in this trial was primarily set for safety assessment, the efficacy outcomes derived from this cohort are preliminary. Nonetheless, further investigation is warranted to evaluate the efficacy of the extended dosing interval.

A post-hoc analysis of the NAVIGATOR trial has also shown that treatment with anti-TSLP monoclonal antibody tezepelumab reduced the SNOT-22 score at weeks 28 and 52 in patients with severe asthma and comorbid CRSwNP¹⁷. Specifically, the improvement of SNOT-22 score was −10.58 points (tezepelumab vs. placebo) at week 52, which is similar with the improvement of −12.5 observed in ECRSwNP patients of our Q2W cohort. However, the NAVIGATOR trial focused on patients with asthma, but not those with uncontrolled CRSwNP. There are also a few ongoing trials investigating TSLP signaling blockage therapies in patients with CRSwNP (NCT06164704, NCT06036927, NCT05891483). Moreover, several cases have shown that tezepelumab can improve CRSwNP-related symptoms in patients with comorbid aspirin-exacerbated respiratory disease or asthma^18,19. WAYPOINT, a phase 3 trial published very recently has demonstrated the efficacy of tezepelumab in patients with severe, uncontrolled CRSwNP²⁰. It’s worth noticing that the population enrolled in WAYPOINT and our study are different. Randomization of the WAYPOINT study was monitored to ensure that 50%–70% of the study population will have co-morbid asthma/ aspirin-exacerbated respiratory disease, and at least 50% will have had prior surgery for CRSwNP, while our study was stratified to ensure the 1:1 ratio of ECRSwNP and nonECRSwNP. In asthma, tezepelumab has been proven to be clinically effective irrespective of type 2 inflammation²¹. It has been reported that Asian countries have more patients with nonECRSwNP than Western countries¹³, and that both eosinophilic and non-eosinophilic inflammatory phenotypes are seen in patients with uncontrolled CRSwNP. In this regard, we tried to demonstrate the efficacy of anti-TSLP treatment in overall uncontrolled CRSwNP population. As demonstrated that, the degree of tissue eosinophil infiltration reflects the local inflammation level of the nasal lesion and has been associated with the prognosis after surgery and severity of CRSwNP^22–24. In the meanwhile, an instructional video (Supplementary Movie 1) was also provided to all centres to ensure that the process of obtaining nasal polyp biopsy tissue was standardized and safe. In terms of efficacy of CM326 in patients with nonECRSwNP, we did not observe improvement in efficacy endpoints. This divergence might suggest a differential mechanism of TSLP involved in the pathogenesis of CRSwNP compared to asthma. The use of different criteria to characterize type 2 inflammation via tissue or peripheral blood eosinophil could provide an alternative explanation.

A post-hoc analysis was conducted to assess the potential of TSLP, the target of CM326, in identifying patients most likely to benefit from CM326 treatment. In the Q2W cohort, which comprises both ECRSwNP and nonECRSwNP patients, individuals with high baseline plasma TSLP levels ( > 330 fg/mL) demonstrated significant improvements in NPS, NCS, Lund-Mackay CT score, and TSS following treatment with CM326, in comparison to those receiving a placebo. Additionally, these patients experienced reductions in both systemic and local PD markers. Conversely, in patients with low baseline plasma TSLP levels ( ≤ 330 fg/mL), CM326 treatment did not significantly improve either objective or subjective efficacy outcomes, nor did it reduce the levels of PD markers compared with placebo. Utilizing plasma TSLP levels enables the identification of a distinct group of patients who are likely to benefit from CM326 treatment. To our best knowledge, this study pioneers the establishment of plasma TSLP as a predictor in anti-TSLP treatment, which needs to be further validated in an independent cohort.

As a proof-of-concept trial, this study has some limitations. Multiplicity adjustments were not designed, and the number of participants (84 patients) was small, therefore, all P values were nominal. During the open-label period, the benefits derived from the long-term regular use of mometasone furoate nasal spray may have also contributed to the improvements in both the CM326 and control groups, which cannot be entirely ruled out. The majority of the recruited patients in this study were from Han population. Further studies with larger sample sizes are in preparation to investigate the safety and efficacy of CM326.

In conclusion, we conducted a clinical trial to investigate the safety and efficacy of blocking TSLP in patients with uncontrolled CRSwNP. Our results demonstrate a favourable safety profile for CM326. The study also indicates that patients with plasma TSLP levels exceeding 330 fg/mL may potentially derive greater benefits from anti-TSLP biologic therapy. These findings will pave the way for future clinical investigation of CM326 in patients with CRSwNP.

Methods

Study design and oversight

This trial was conducted in full accordance with Good Clinical Practice (GCP) guidelines, the Declaration of Helsinki, and the laws and regulations of the local in China. The trial was approved by independent ethics committees of Beijing Tongren Hospital of Capital Medical University and each participating center. All patients provided written informed consent. The trial protocol and statistical analysis plan are provided in the Supplementary Information.

This randomized, double-blind, placebo-controlled trial was conducted at 18 centres in China. The trial was reregistered on March 18, 2022 (https://clinicaltrials.gov/study/NCT05324137), before the commencement of patient recruitment. The graphic study design of the trial was presented Supplementary Fig. 9, and the detailed trial design has been published previously²⁵. This study was initially set up to employ multiple ascending doses, however, in light of the safety data from two parallel phase 1 studies (NCT04842201, NCT05171348), patients were only enrolled in the high dose (220 mg) groups to avoid ineffective low dose exposure. Following a 4-week screening/run-in period, eligible patients were sequentially randomized to receive either subcutaneous CM326 220 mg or placebo (2:1) Q2W and to receive either CM326 220 mg or placebo (5:1) Q4W during the 16-week double-blind period. After the double-blind period, all patients in each cohort received CM326 220 mg Q2W or Q4W for 36 weeks. Subsequently, patients were followed up for 12 weeks. Throughout the entire study, all patients received mometasone furoate nasal spray 100 μg daily in each nostril as background treatment. Saline nasal lavage, systemic antibiotics, and short-course SCS were permitted as needed.

Because uncontrolled CRSwNP was composed of various cellular phenotype and severity, patients in the Q2W cohort were stratified by tissue eosinophil count at screening and NPS at baseline ( < 5 vs. ≥5), while the Q4W cohort enrolled only patients with NPS ≥ 5 points and was stratified by tissue eosinophil count. ECRSwNP was defined as the tissue eosinophil count ≥55 eosinophils/HPF in nasal polyp biopsies. The cut-off value was pre-specified and selected based on our previous evidence due to its high accuracy rate (96.9%) for predicting recurrence⁷, which was further validated by a meta-analysis²⁶. The number of patients was balanced between strata. Stratified randomization was used with a block size of 3 in each stratum of the Q2W group and a block size of 6 in each stratum of the Q4W group. The interactive web response randomization system (IWRS) was applied for patient randomization and study drug dispensing, with schedules generated by SAS software and imported into the IWRS by a system engineer. Participants and investigators were blinded throughout the double-blind treatment period. Clinical data was collected using Taimei eCollect V5. All statistical analyses were performed using SAS version 9.4.

Participants

Eligible patients were aged 18–70 with bilateral nasal polyps and symptoms of chronic rhinosinusitis despite intranasal corticosteroid therapy before randomization. They had received SCS in the preceding 2 years or previous nasal polyp surgery 6 months before the run-in period. Patients were required to have a NPS of at least 3 at screening, with a minimum score of 1 for each nostril. The NPS is the sum of the left and right nostril scores ranging from 0 to 8 (0–4 for each nostril) based on polyp size as evaluated by nasal endoscopy. Patients must also have at least two of the following symptoms: nasal congestion score (NCS) of at least 2 (moderate to severe) at screening with an average weekly score of at least 1 (mild to severe) at randomization, decreased/loss of smell or rhinorrhea. Patients who had a forced expiratory volume in 1 second (FEV₁) of 60% or lower than the predicted normal were excluded. Detailed inclusion and exclusion criteria are listed in the Supplementary Method.

Outcomes

Primary endpoints were the safety of CM326 in patients with CRSwNP and the change from baseline in NPS at week 16 in patients with ECRSwNP. NPS was scored as proposed by Gevaert et al.²⁷ Safety measurements included TEAEs, vital signs, physical examinations, electrocardiogram, and clinical laboratory tests. Secondary endpoints were the change from baseline in NPS at week 16 in patients with nonECRSwNP, PD markers and immunogenicity of CM326 in patients with ECRSwNP and nonECRSwNP. PD markers consisted of serum/plasma biomarkers including TARC, IgE, IL-5, IL-13, periostin, and eosinophil counts in peripheral blood and nasal polyp tissue.

Exploratory efficacy endpoints included changes from baseline in objective and subjective CRSwNP symptom-related outcomes including NPS, NCS, TSS (including nasal congestion, loss of smell, and anterior/posterior nasal discharge), UPSIT score, and SNOT-22 score at each evaluation. Additional measures included changes from baseline in Lund-Mackay CT score at week 16 and plasma TSLP levels at each evaluation.

Measurement of serum/plasma PD markers

Following the manufacturer’s protocols, TARC (SDN00, R&D Systems, USA), and periostin (EHPOSTN, Thermo Fisher Scientific, USA) were quantified using enzyme-linked immunosorbent assay; IL-5 (K151J3S, MSD, USA), IL-13 (K151Y9S, MSD, USA), and TSLP (K151D3S, MSD, USA) were analyzed via Meso Scale Discovery (MSD) assay. Total serum IgE concentration was quantified using a sandwich immunoassay with Ruthenium (Ru)-labeled and biotin-labeled goat anti-human IgE antibodies, both derived from affinity-purified polyclonal antibody (A80-108A, Bethyl, Germany). Streptavidin-coated MSD plates were incubated with biotinylated anti-IgE antibody for capture antibody immobilization. Following a wash step, serum samples were added to allow IgE binding. Subsequently, Ru-labeled anti-IgE detection antibody was added. Electrochemiluminescent signals were then generated using MSD read buffer and measured using an MSD instrument. Serum IgE concentrations were determined using a standard curve, with a lower limit of 0.031 ng/mL.

Statistical analysis

In general, a sample size of 20–100 patients is suggested for a phase 1 clinical trial^28,29. Based on this, the sample size was set to be 15 patients with a ratio of 2:1 per stratum in 220 mg Q2W group and 12 patients with a ratio of 5:1 per stratum in 220 mg Q4W group, and the total sample size was 84 with 60 patients receiving 220 mg CM326 and 24 patients receiving placebo, which is sufficient for safety evaluation of CM326 in patients with uncontrolled CRSwNP. There was no adjustment for multiple comparisons, so the results should not be used to infer definitive treatment effects.

For efficacy measurements, under full analysis set, values and their changes from baseline within the 16 weeks after treatment were descriptively summarized by visit and group with last observation carried forward (LOCF) imputation except for Lund-Mackay CT score, where baseline observation carried forward (BOCF) imputation was employed. Data after rescue therapy were considered missing. Efficacy analyses within 16 weeks were conducted using t-test. For the open-label and follow-up periods, efficacy outcomes, under modified full analysis set, were descriptively summarized without imputation for missing data. Biomarkers were analyzed under pharmacodynamic set using Wilcoxon rank-sum test. All statistics for safety and immunogenicity were descriptively analyzed under safety set and immunogenicity set respectively. The definition of different analysis sets and details of statistical analyses are shown in Supplementary Methods and statistical analysis plan.

Post-hoc analysis

A modified criterion according to the recently updated guideline of biologics in CRSwNP² was applied to define an excellent response to CM326 treatment. Patients who met all the following criteria at week 16 of CM326 Q2W treatment were considered responders: a) a reduction of at least 1 point from baseline in the NPS; b) a decrease in SNOT22 score of at least 8.9 points from baseline; c) an improvement from baseline in the UPSIT score. An ROC curve analysis was used to determine the baseline plasma TSLP level threshold for predicting response to CM326 treatment. Afterwards, changes from baseline of efficacy endpoints and PD biomarkers at week 16 in the Q2W cohort were reanalyzed in patients stratified by predicted baseline plasma TSLP level.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

C.W. and L.Z. had full access to the data and take responsibility for the integrity of the data and the accuracy of the data analysis. G.Z. contributed to statistical analysis. L.Z. contributed to obtaining funding. L.Z., C.W., M.X., Feng Lan, B.Y., and S.S. contributed to supervision. M.X., Feng Lan, B.Y., S.S., S.L., Lijia W., X.S., L.J., Y.J., J.X., J.C., L.S., J.Y., Y.Y., H.F., G.T., Q.Z., S.Q., X.W., X.L., Yu X., S.Y., Z.X., Feng Liu, Q.L., Y.Z., Yan X., Lin W., G.Y., H.Y., G.Z., B.C., C.W., and L.Z. contributed to concept and study design, acquisition, analysis, or interpretation of data, drafting of the manuscript, critical revision of the manuscript for important intellectual content, and administrative, technical, or material support.

Peer review

Peer review information

Nature Communications thanks Matteo Gelardi, who co-reviewed with Rossana Giancaspro; Larry Borish, Nicholas Campion and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The data supporting the findings of this trial are available within the manuscript and its Supplementary Information files. All requests for further data sharing will be reviewed by the leading clinical center, Beijing TongRen Hospital, to evaluate whether the request is subject to any intellectual property or confidentiality obligations. Requests may be submitted to dr.luozhang@139.com. A signed data access agreement with the leading center is required before accessing shared data. Requests will be responded to in 3 months. The protocol and statistical analysis plan are provided in the Supplementary Information. The protocol has been published previously²⁵. Source data are provided with this paper.

Competing interests

B.C. is a shareholder of Keymed Biosciences (Chengdu) Co., Ltd. and also an inventor on patents (CN 112876564 B, CN 114887053 A, WO2021104053A1, EP4067377A1, US20230029835A1, JP2023503700A, KR1020220119394A, AU2020390926A1). G.Z. is an employee and shareholder of Keymed Biosciences (Chengdu) Co., Ltd. H.Y. is an employee of Keymed Biosciences (Chengdu) Co., Ltd. All other authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-63682-x.