50-valent inactivated rhinovirus vaccine is broadly immunogenic in rhesus macaques

As the predominant etiological agent of the common cold, human rhinovirus (HRV) is the leading cause of human infectious disease. Early studies showed monovalent formalin-inactivated HRV vaccine can be protective, and virus-neutralizing antibodies (nAb) correlated with protection. However, co-circulation of many HRV types discouraged further vaccine efforts. We approached this problem straightforwardly. We tested the hypothesis that increasing virus input titers in polyvalent inactivated HRV vaccine will result in broad nAb responses. Here, we show that serum nAb against many rhinovirus types can be induced by polyvalent, inactivated HRVs plus alhydrogel (alum) adjuvant. Using formulations up to 25-valent in mice and 50-valent in rhesus macaques, HRV vaccine immunogenicity was related to sufficient quantity of input antigens, and valency was not a major factor for potency or breadth of the response. We for the first time generated a vaccine capable of inducing nAb responses to numerous and diverse HRV types.


32
HRV causes respiratory illness in billions of people annually, a socioeconomic burden 1 . 33 HRV also causes pneumonia hospitalizations in children and adults and exacerbations of asthma 34 and chronic obstructive pulmonary disease (COPD) 2 . HRV was found to be the second leading

55
There are three species of HRV, A, B, and C. Sequencing methods define 83 A types, 32 56 B types, and 55 C types 20,21 . It is thought there are 150 to 170 serologically distinct HRV types.

57
HRV A and C are associated with asthma exacerbations and with more acute disease than HRV 58 B 22,23 . HRV C was discovered in 2006 and 2007 24-27 and recently cultured in cells 28,29 . Here, we 59 focused on HRV A, the most prevalent species. There are no permissive animal challenge 60 models of HRV virus replication, but mice and cotton rats can recapitulate aspects of HRV 61 pathogenesis 30,31 . The best efficacy model is human challenge. In monovalent vaccine trials, 62 formalin-inactivated HRV-13 was validated prior to clinical testing by assessing induction of 63 nAb in guinea pigs, and a reciprocal serum nAb titer of 2 3 resulting from four doses of a 1:125 64 dilution of the vaccine correlated with vaccine efficacy in humans 9 . Although the nAb titer 65 required for protection is not defined, early studies established inactivated HRV as protective in 66 humans, and immunogenicity in animals informed clinical testing.

68
Results and discussion 69 We first used BALB/c mice to test immunogenicity. We propagated HRVs in H1-HeLa 70 cells and inactivated infectivity using formalin. Sera from naïve mice had no detectable nAb 71 against HRV-16 ( Fig. 1). Alum adjuvant enhanced the nAb response induced by i.m. inactivated 72 HRV-16 ( Fig. 1). There was no effect of valency (comparing 1-, 3-, 5-, 7-, and 10-valent) on the 73 nAb response induced by inactivated HRV-16 or to the 3 types in the 3-valent vaccine (HRV-16, 74 HRV-36, and HRV-78) (Fig. 1). The 50% tissue culture infectious dose (TCID 50 ) titers of the 75 input viruses prior to inactivation (inactivated-TCID 50 ) are provided in Supplemental Table 1. 76 Original antigenic sin can occur when sequential exposure to related virus variants results in 77 biased immunity to the type encountered first 32 . In bivalent HRV-immune mice, we observed 78 modest original antigenic sin following prime vaccination with 10-valent inactivated HRV, and 79 boost vaccination partially alleviated the effect (Supplemental Fig. 1), similar to influenza 80 virus 32 . Collectively, these results prompted us to explore more fully the nAb response to 81 polyvalent HRV vaccine.

82
In 1975, it was reported that two different 10-valent inactivated HRV preparations 83 induced nAb titers to only 30-40% of the input virus types in recipient subjects 33 . However, the 84 input titers of viruses prior to inactivation ranged from 10 1.5 to 10 5.5 TCID 50 per ml, and these 85 were then diluted 10-fold to generate 10-valent 1.0 ml doses given i.m. as prime and boost with 86 no adjuvant 33 . We hypothesized that low input antigen doses are responsible for poor nAb 87 responses to 10-valent inactivated HRV. We reconstituted the 1975 10-valent vaccine, as closely 88 as possible with available HRV types, over a 10 1 to 10 5 inactivated-TCID 50 per vaccine dose, and 89 we compared it to a 10-valent vaccine of the same types with input titers ranging from > 10 5 to > 90 10 7 inactivated-TCID 50 per dose. The reconstituted 1975 vaccine resulted in no detectable nAb 91 after prime vaccination and, following boost vaccination, nAb to the five types that had the 92 highest input titers (Fig. 2). The high titer vaccine resulted in nAb to 5 of 10 types after prime 93 vaccination and all 10 types after the boost (Fig. 2). Following the boost vaccinations, there 94 appeared to be a 10 4 inactivated-TCID 50 per vaccine dose threshold for the induction of nAb in 95 this model (Fig. 2b). Above this titer, there was no correlation between input load and nAb  Table 2) to accommodate the volume adjustment. The 10-valent inactivated 102 HRV vaccine induced nAb to 100% of input types following the prime and the boost (Fig. 3a). 103 The nAb induced by 10-valent inactivated HRV were persisting at 203 days post-boost 104 (Supplemental Fig. 2). The 25-valent inactivated HRV prime vaccination induced nAb to 18 of 105 25 (72%) virus types, and the 25-valent boost resulted in nAb against 24 of the 25 types (96%) 106 (Fig 3b). The average nAb titer resulting from prime + boost was 2 7 for 10-valent and 2 6.8 for 25-107 valent. The data demonstrate broad neutralization of diverse HRV types with a straightforward 108 vaccine approach.

109
In order to increase vaccine valency, we chose rhesus macaques (RMs) and a 1.0 ml i.m.  Table 3). The 25-valent 114 vaccine induced nAb to 96% (RM A) and 100% (RM B) of input viruses following the prime 115 vaccination (Fig. 4a). The 50-valent vaccine induced nAb to 90% (RM C) and 82% (RM D) of 116 input viruses following the prime vaccination (Fig. 4c). The breadth of nAb following prime 117 vaccination in RM was superior to what we observed in mice, which may have been due to 118 animal species differences and/or higher inactivated-TCID 50 input titers in the RM vaccines.

119
Following boost vaccination, there were serum nAb titers against 100% of the types in 25-valent 120 HRV-vaccinated RMs (Fig. 4b) and 98% (49 out of 50) of the virus types in 50-valent HRV-121 vaccinated RMs (Fig. 4d). The average nAb titer resulting from prime + boost in RMs was 2 9.3 122 for 25-valent and 2 8.6 for 50-valent. The nAb responses were type-specific, not cross- inactivated-TCID 50 per type per dose will be useful. Therefore, HRV stock titers ≥ 10 7 TCID 50 128 per ml are required for a potential 83-valent HRV A formulation in a 0.5 ml dose containing 129 alum adjuvant. The HRV stocks used in our vaccinations were produced in H1-HeLa cells, a 130 good substrate for HRV replication but not suitable for vaccine manufacturing. We compared the 131 infectious yield of 10 HRV types in H1-HeLa and WI-38, which can be qualified for vaccine 132 production. Adequate yields were obtained from WI-38 cells (Supplemental Fig. 4). Injectable 133 vaccines require defined purity. As proof of principle, we purified three HRV types by high 134 performance liquid chromatography and found uncompromised immunogenicity of trivalent 135 inactivated purified HRV in mice (Supplemental Fig. 5). 136 Forty years ago, the prospects for a polyvalent HRV vaccine were dour for good  Advancing valency may be applicable to vaccines for other antigenically variable pathogens.  In mice, peripheral blood was collected into microcentrifuge tubes from the submandibular vein.

244
Samples were incubated at room temperature for 20 min to clot. The tubes were centrifuged 7500 245 × g for 10 min to separate serum. The serum samples were pooled from mice of each group and 246 stored at −80 °C until used. Phlebotomy involving RMs was performed under either ketamine 247 (10 mg/kg) or Telazol (4 mg/kg) anesthesia on fasting animals. Following anesthesia with 248 ketamine or Telazol, the animals were bled from the femoral vein. Yerkes blood collection 249 guidelines were followed and no more than 10 ml/kg/28 days of blood was collected. After       for each group, and nAb titers (y-axis) were measured against the indicated types in the vaccines.

304
The dashed line represents LOD. Undetectable nAb were assigned LOD/2, and some symbols 305 below LOD were nudged for visualization. Three independent experiments using low input titers 306 showed similar results. There was a statistically significant association between input TCID 50 307 virus titer and a detectable nAb response following prime (P = 0.01) and boost (P = 0.03) 308 vaccination (Fisher's exact test).

311
The inactivated-TCID 50 input titers per dose are specified in Supplemental Table 2