Differential induction of type I interferons in macaques by wild‐type measles virus alone or with the hemagglutinin protein of the Edmonston vaccine strain

ABSTRACT

Measles vaccines are highly effective and safe; however, the mechanism(s) underlying their attenuation has not been well understood. In this study, type I IFNs (IFN‐α and IFN‐β) induction in macaques infected with measles virus (MV) strains was examined. Type I IFNs were not induced in macaques infected with wild‐type MV. However, IFN‐α was sharply induced in most macaques infected with recombinant wild‐type MV bearing the hemagglutinin (H) protein of the Edmonston vaccine strain. These results indicate that the H protein of MV vaccine strains may have a role in MV attenuation.

Abbreviations

BAL

bronchoalveolar lavage

DI

defective interfering

dpi

days post infection

EdH‐EGFP2

EGFP‐expressing recombinant wild‐type MV bearing the H protein of the Edmonston vaccine strain

EGFP

enhanced green fluorescent protein

H protein

hemagglutinin protein

MV

measles virus

PBMC

peripheral blood mononuclear cell

pDC

plasmacytoid DCs

SLAM

signaling lymphocyte activation molecule

Measles is highly infectious and remains a major cause of childhood morbidity and mortality worldwide despite the availability of effective vaccines 1. Measles vaccines are generated by successive passages of field isolates of MV in cells of different origins 1 and are highly effective and safe; however, the mechanism(s) underlying their attenuation has not been well understood. One major difference between wild‐type and vaccine strains of MV is the receptor specificity of the H protein in vitro. The H proteins of wild‐type MV strains recognize SLAM, also known as CD150, which is expressed in certain immune system cells 2, and nectin‐4, also known as poliovirus receptor‐like protein 4, which is expressed in epithelial cells in trachea, skin, lung, prostate and stomach, as cellular receptors 3, 4. On the other hand, in addition to recognizing SLAM and nectin‐4 as cellular receptors, the H proteins of vaccine MV strains also recognize CD46, which is ubiquitously expressed on all nucleated human and monkey cells 5, 6.

To examine the contribution of the H protein to MV attenuation, an EdH‐EGFP₂ has been generated using a reverse genetics system based on the pathogenic wild‐type IC‐B strain 7 and cynomolgus monkeys aged between 4 and 5 years (three animals per strain) have been intranasally infected with wild‐type MV (IC323‐EGFP₂) or EdH‐EGFP₂ 8. In IC323‐EGFP₂ and EdH‐EGFP₂, the EGFP gene is between the N and P genes. Interestingly, EdH‐EGFP₂ replicated significantly less in tissues and lymphocytes of infected macaques than in those of wild‐type MV. From these results and because it has been well established that type I IFNs are induced by many viruses and play central roles in the host defense against viral infection, we speculated that type I IFNs may affect the growth of EdH‐EGFP₂ in macaques 9.

We performed all animal experiments in compliance with the guidelines of the National Institute of Infectious Disease (Permission Number 510008). In this study, we examined induction of type I IFNs in macaques with the aim of investigating the mechanism(s) for growth attenuation of EdH‐EGFP₂ in these animals. For this purpose and because it is known that DI RNAs, especially 5′ copy‐back DI RNAs, in virus stocks of MV can induce type I IFN through interaction with the RNA helicases retinoic acid‐inducible gene‐1/melanoma differentiation‐associated gene 5 10-13, we first assessed the presence of DI RNA in virus stocks. We extracted viral RNA from virus stocks using a QIAamp Viral RNA Mini kit (Qiagen, Hilden, Germany) and detected DI RNA using RT‐PCR, as previously described 10. We detected DI RNAs in the Edmonston (laboratory strain) and IC‐V (wild‐type strain isolated in Vero cells) stocks as reported 10, whereas we did not detect DI RNA in the IC323‐EGFP₂ or EdH‐EGFP₂ virus stocks used in this experiment (Fig. 1).

Figure 1

Absence of 5′ copy back DI RNA in MV stocks. Stocks of IC323‐EGFP₂ and EdH‐EGFP₂ used for infection were tested for the absence of 5′ copy back DI RNA. 5′ copy back DI genomes were detected with primers (JM396; 5′‐TATAAGCTTACCAGACAAAGCTGGGAATAGAAACTTCG‐3′ and JM403; 5′‐CGAAGATATTCTGGTGTAAGTCTAGTA‐3′). MV standard genomes were detected using primers (JM396 and JM402; 5′‐TTTATCCAGAATCTCAARTCCGG‐3′). The Edmonston and IC‐V strains, which are known to contain 5′ copy back DI RNA, were used for positive controls.

To examine the interferon responses elicited by IC323‐EGFP₂ and EdH‐EGFP₂ in macaques, we compared the transcription of IFN‐α and IFN‐β genes in PBMCs of infected monkeys, as previously reported 14. We collected PBMCs on 0, 3, 7, and 10 dpi, and stored them in RNAprotect Animal Blood Tubes (Qiagen) at −30°C. We collected tissues of inguinal lymph nodes 7 days prior to infection and 10 dpi, and stored them in RNAlater solution (Qiagen) at −30°C. We collected tissues of lung at 10 dpi and stored them in RNAlater solution at −30°C. We collected plasma 7 days prior to infection, and on 0, 3, 7, and 10 dpi, and stored them at −80°C. We collected BAL fluids 10 dpi and stored them at −80°C. We isolated total RNA from PBMCs and tissues by using an RNeasy mini kit and RNase‐free DNase (Qiagen) according to the manufacturer's protocol and reverse transcribed using oligo (dT) primer and PCR amplified with a Thermal Cycler Dice TP800 (Takara, Tokyo, Japan) by using FastStart SYBR Green Master (Roche, Mannheim, Germany). We found that IFN‐α and IFN‐β transcription was transiently down‐regulated on Day 3 in macaques infected with IC323‐EGFP₂ (No. 5058, 5062, and 5069), (Fig. 2a, b); the levels of IFN‐α and IFN‐β transcription returned to the baseline by Day 7. IFN‐α and IFN‐β transcription was gradually induced from Day 0 to Day 7 in macaques infected with EdH‐EGFP₂ (No. 5056, 5057, and 5068). By Day 10, the levels of IFN‐α and IFN‐β transcription had decreased in all monkeys infected with both strains.

Figure 2

IFN‐α/β mRNA expression in PBMCs, inguinal lymph nodes and lungs. (a) IFN‐α and (b) IFN‐β mRNA expression in PBMCs from monkeys infected with IC323‐EGFP₂ or EdH‐EGFP₂ were measured by RT‐qPCR. PBMCs were collected on 0, 3, 7 and 10 dpi. (c) IFN‐α and (d) IFN‐β mRNA expression in inguinal lymph nodes and lungs from monkeys infected with IC323‐EGFP₂ or EdH‐EGFP₂ were measured by RT‐qPCR. Inguinal lymph nodes were excised 7 days prior to infection and on 10 dpi, and lungs were excised on 10 dpi. Three monkeys (No. 5058, 5062 and 5069) were infected with IC323‐EGFP₂ and three (No. 5056, 5057, and 5068) with EdH‐EGFP₂. For amplification of IFN‐α mRNA, IFN‐α F primer 5′‐GCCTGAAGGACAGACATGACTTT‐3′ and IFN‐α R primer 5′‐GGATGGTTTGAGCCTTTTGG‐3′ were used. For amplification of IFN‐β mRNA, IFN‐β F primer 5′‐TGCCTCAAGGACAGGATGAA‐3′ and IFN‐β R primer 5′‐ATGGTCCAGGCACAGTGACT‐3′ were used. For amplification of the 18S rRNA segment, the 18S sense primer 5′‐TCAAGAACGAAAGTCGGAGG‐3′ and 18S antisense primer 5′‐GGACATCTAAGGGCATCACA‐3′ were used. For determining the relative amounts of IFN‐α/β mRNA, the amounts of IFN‐α/β mRNA in cynomolgus monkey PBMCs infected with Sendai virus (Cantell strain), which is commonly used to induce IFN‐α/β in vitro, were set to 10¹.

We also examined Type I IFN responses elicited by IC323‐EGFP₂ and EdH‐EGFP₂ in several tissues of infected monkeys. However, the levels of IFN‐α and IFN‐β transcription in inguinal lymph nodes did not change significantly between 7 days before infection and Day 10 in any monkeys (Fig. 2c, d). Day 10 may be too late to detect changes in levels of IFN‐α and IFN‐β transcription. We detected similar levels of IFN‐α and IFN‐β transcription in lungs of monkeys infected with both strains (Fig. 2c, d).

Next, we examined concentrations of IFN‐α in plasma and BAL using a VeriKine cynomolgus/rhesus IFN‐α serum ELISA kit (PBL, Piscataway, NJ, USA) according to the manufacturer's protocol. Plasma concentrations of IFN‐α were not significantly changed in wild‐type IC323‐EGFP₂‐infected macaques (No. 5058, 5062 and 5069); however, we observed slight induction in one macaque (no. 5062) on Day 7 (Fig. 3a). On the other hand, on Day 7 plasma concentrations of IFN‐α increased sharply four‐to five‐fold in two (No. 5056 and 5057) of the three macaques infected with EdH‐EGFP₂ and then declined by Day 10. To confirm IFN‐α induction in EdH‐EGFP₂‐infected macaques, we examined plasma collected in former experiments in which we had infected macaques with recombinant MV strains. In the first group, we used two 10‐year‐old macaques (No. 4568 and 4569). We infected one of these macaques (No. 4568) with wild‐type MV (IC323‐EGFP), in which the EGFP gene is between the leader sequence and the N gene 8, and the other (No. 4569) with IC323‐EGFP₂. In the second group, we used seven juvenile (1 year old) macaques (No. 4848, 4849, 4850, 4858, 4859, 4860 and 4865). We infected three of them (No. 4850, 4860 and 4865) with IC323‐EGFP₂ and the other four (No. 4848, 4849, 4858 and 4859) with EdH‐EGFP₂. Again, plasma concentrations of IFN‐α increased sharply in two EdH‐EGFP₂‐infected macaques (No. 4848 and 4849) on Day 7 but did not do so in IC323‐EGFP‐ and IC323‐EGFP₂‐infected macaques (No. 4568, 4569, 4850, 4860 and 4865) (Fig. 3b). Macaques No. 4860, 4865, 4858 and 4859 and macaques No. 4850, 4848 and 4849 were killed on Days 3 and 7, respectively. Samples from these animals were therefore not available thereafter. We did not observe induction of IFN‐α in BAL of any of the infected monkeys on Day 10 (Fig. 3a).

Figure 3

Plasma and BAL concentrations of IFN‐α. (a) Plasma and BAL concentrations of IFN‐α were measured by ELISA in monkeys infected with IC323‐EGFP₂ or EdH‐EGFP₂. Plasma was collected 7 days prior to infection and on 0, 3, 7 and 10 dpi. BAL was collected at 10 dpi. Three monkeys (No. 5058, 5062 and 5069) were infected with IC323‐EGFP₂ and three (No. 5056, 5057 and 5068 with EdH‐EGFP₂. (b) Plasma concentrations of IFN‐α were measured by ELISA in monkeys infected with IC323‐EGFP, IC323‐EGFP₂ or EdH‐EGFP₂. Plasma was collected on 0, 3, 7 and 10 dpi. One monkey (No. 4568) was infected with IC323‐EGFP, four (No. 4569, 4850, 4860, and 4865) with IC323‐EGFP₂, and four (No. 4848, 4849, 4858, and 4859) with EdH‐EGFP₂. The sensitivity of this assay is 0.30 pg/mL. nd, not done.

Although many studies have demonstrated IFNs production by MV in vitro, little is known about IFNs production in measles patients. In one clinical study using a sensitive radioimmunoassay, it was found that IFN‐α was not induced in plasma of measles patients 15. In another clinical study, Yu et al. found that IFN‐α expression was suppressed in PBMCs of measles patients 16. In an in vivo study using macaques, Devaux et al. reported that expression of type I IFN genes was well regulated 14. In addition, Shivakoti et al. recently found that type I IFNs are not induced in macaques infected by wild‐type MV 17. We found that IFN‐α is not induced in macaques infected with wild‐type MV (Fig. 3a, b). Our results are consistent with those of previous clinical 15, 16 and in vivo studies using monkeys 14, 17. These results suggest that MV can circumvent host IFNs production, this possibly being mediated by the C and V proteins 18. Likewise, little is known about IFNs production in measles vaccinees. In a previous clinical study, IFN was induced after measles vaccination 19. We found that IFN‐α is sharply induced in plasma of macaques infected with EdH‐EGFP₂ (Fig. 3). Interestingly, it has been shown that large amounts of IFN‐α are rapidly produced from pDCs after infection with the Edmonston strain, mostly independent of viral infection cycles 20. Because pDCs express CD46 but not SLAM 20, pDCs in macaques would be infected with EdH‐EGFP₂ via a CD46‐mediated pathway and may produce large amounts of IFN‐α in plasma. In summary, we found that IFN‐α is induced in macaques infected with wild‐type MV bearing the H protein of the Edmonston vaccine strain but not in those infected with wild‐type MV. Our results suggest that the H protein of vaccine strains of MV have a role in MV attenuation.

ACKNOWLEDGMENTS

We thank T. Ohkura, N. Nagata, and Y. Ami for their ongoing support. We also thank M. Ayata and M. Okuwaki for critical reading and valuable comments. This work was supported in part by grants‐in‐aid (No. 19041012 and 16K08804) from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

DISCLOSURE

The authors have no conflicts of interest associated with this study.