Identification of the causes of poor oral vaccine immunogenicity in low-income countries might lead to more effective vaccines. We measured mucosal and systemic immune parameters at the time of vaccination with oral poliovirus vaccine (OPV) in 292 Indian infants aged 6–11 months, including plasma cytokines, leukocyte counts, fecal biomarkers of environmental enteropathy and peripheral blood T-cell phenotype, focused on gut-homing regulatory CD4+ populations. We did not find a distinct immune phenotype associated with OPV immunogenicity, although viral pathogens were more prevalent in stool at the time of immunization among infants who failed to seroconvert (63.9% vs. 45.6%, p = 0.002). Using a machine-learning approach, we could predict seroconversion a priori using immune parameters and infection status with a median 58% accuracy (cross-validation IQR: 50–69%) compared with 50% expected by chance. Better identification of immune predictors of OPV immunogenicity is likely to require sampling of mucosal tissue and improved oral poliovirus infection models.

确定低收入国家口服疫苗免疫原性差的原因可能会导致更有效的疫苗。我们测量了 292 名 6-11 个月大的印度婴儿在接种口服脊髓灰质炎病毒疫苗 (OPV) 时的黏膜和全身免疫参数，包括血浆细胞因子、白细胞计数、环境肠病的粪便生物标志物和外周血 T 细胞表型，重点关注关于肠道归巢调节 CD4+ 群体。我们没有发现与 OPV 免疫原性相关的明显免疫表型，尽管在未能血清转化的婴儿中，在免疫接种时粪便中的病毒病原体更为普遍（63.9% 对 45.6%，p = 0.002)。使用机器学习方法，我们可以使用免疫参数和感染状态先验预测血清转化，中位数准确率为 58%（交叉验证 IQR：50-69%），而偶然预期的准确率为 50%。更好地识别 OPV 免疫原性的免疫预测因子可能需要对粘膜组织进行取样和改进口腔脊髓灰质炎病毒感染模型。