Rhinovirus species and tonsillar immune responses.

Background Rhinovirus A and C infections are important contributors to asthma induction and exacerbations. No data exist on the interaction of local immune responses in rhinovirus infection. Therefore, we aimed to determine the tonsillar immune responses according to rhinovirus A, B and C infections. Methods We collected tonsillar samples, nasopharyngeal aspirates and peripheral blood from 42 rhinovirus positive tonsillectomy patients. Fifteen respiratory viruses or their types were investigated from nasopharynx and tonsil tissue, and rhinovirus species were typed. The expression of 10 cytokines and 4 transcription factors (IFN-α, IFN-β, IFN-γ, IL-10, IL-13, IL-17, IL-28, IL-29, IL-37, TGF-β, FOXP3, GATA3, RORC2 and Tbet) were studied from tonsil tissue by quantitative PCR. A standard questionnaire of respiratory symptoms and health was filled by the patient or his/her guardian. The patients were divided into three groups by the determination of rhinovirus species. Results Overall, 16 patients had rhinovirus A, 12 rhinovirus B and 14 rhinovirus C infection. In rhinovirus B positive group there were significantly less men (P = 0.0072), less operated in spring (P = 0.0096) and more operated in fall (P = 0.030) than in rhinovirus A or C groups. Rhinovirus A positive patients had more respiratory symptoms (P = 0.0074) and particularly rhinitis (P = 0.036) on the operation day. There were no significant differences between the groups in virus codetection. In adjusted analysis, rhinovirus C infections were associated with increased IFN-α (P = 0.045) and decreased RORC2 expression (P = 0.025). Conclusions Rhinovirus species associated differently with clinical characteristics and tonsillar cytokine responses.


Background
Human rhinovirus (RV) is a positive-strand RNA-virus in the family Picornaviridae and genus Enterovirus [1,2]. Three species have been found, A and B in the 1950s and C in 2006 after the development of highly sensitive molecular techniques [1,3]. RV is widely known to be a major cause of common cold and upper respiratory illnesses [1,[4][5][6] but it is also proven to cause lower respiratory diseases [1,[4][5][6]. There is evidence that it is a contributor to asthma induction and exacerbations [1,[4][5][6]. Especially RV-A and RV-C seem to cause more severe respiratory illnesses and are dominate over RV-B in patients with asthma or chronic obstructive pulmonary disease [1,2,4,7]. Moreover RV-A and RV-C are common in hospitalized children [5,7,8].
RV infection causes cell destruction and changes in immunological reactions [6,8]. It has influenced the expression of several interferons and cytokines such as IL (interleukin) -4, IL-6, IL-8, IL-13, IL-16 and IFN (interferon)-γ [1,6,8]. Differences between the immune responses of the three species has not been extensively studied. Jong et al. [9] found no significant differences in the cytokine levels of nasopharyngeal aspirate of RVinfected children but they studied only four cytokines (IFN-γ, IL-4, IL-10 and tumor necrosis factor α). It seems that RV-B replicates more slowly and induces less cytokine production than RV-A and RV-C [10]. Mikola et al. Clin Transl Allergy (2019) 9:63 In our previous studies [11,12] we have regarded tonsils as a good in vivo-model for investigating immune responses. Tonsils are local lymphoid tissue and in close contact with infective agents and allergens. In this study our aim was to observe the differences in tonsillar cytokine expression between RV-A, -B and -C species in routine tonsillectomy patients. Most previous studies concern only cytokine expression of nasopharyngeal aspirate in hospitalized patients.

Patients
We enrolled 200 patients who were going through elective tonsillectomy or adenotomy between April 2008 and March 2009 due to clinical indication for the operation in Satakunta Central Hospital, Pori Finland. Written consent to participate in the study from the patient or his/ her guardian was considered as inclusion criteria along with the tonsillectomy. The study protocol was approved by the Ethics Committee of Satakunta Central Hospital.

Study protocol
From the enrolled patients we collected samples from tonsils, nasopharyngeal aspirate and peripheral blood. The tonsillectomy was performed according to clinical routine. Tonsil samples were cut into 3-4 mm pieces and stored in RNAlater RNA stabilization reagent (Qiagen, Hilden, Germany), incubated + 4 °C until next working day and after removal of non-absorbed reagent stored in − 80 °C. Nasopharyngeal aspirates were collected at the beginning of the operation during anesthesia using a standardized procedure [11]. For the viral analysis, part of the tonsil and nasopharyngeal aspirate (NPS) were stored in dry tubes at − 80 °C [11]. Serum 25(OH) D measurement and serum IgE measurements for food allergen and aeroallergen screening (Phadiatop Combi ® , Phadia, Uppsala, Sweden) were made from peripheral blood samples. Patient or his/her guardian also filled a standard questionnaire concerning their respiratory symptoms 30 days prior the operation and allergic diseases (Additional file 1: Table S1).

Statistical analysis
Data was analyzed using JMP Pro version 12.0.1 software (SAS Institute Inc. Cary, NC, USA). Due to skewed distribution continuous variables are described as medians and interquartile ranges and were analyzed using Kruskal-Wallis test. Categorical variables are expressed as frequencies and percentages and were analyzed using Chi square test or Fisher´s exact test (when counts < 5). Before regression analyses, cytokine and transcription factor values were log-transformed because of positively skewed distributions. Clinical, viral and immunological differences between study groups were analyzed using unadjusted and multivariable linear model analysis. The adjustments for immunologic analyses included clinical factors and virus infections which significantly differed between the groups (sex, season of the surgery spring or fall, any respiratory symptoms on the operation day, rhinitis on the operation day) and age. Backward stepwise method was used for the final adjustment model separately for each cytokine and transcription factor. Only statistically significant factors were kept in the model. P < 0.05 was considered statistically significant.

Recruitment and study population
Of 200 patients 143 had viral and immunological analyzes made from their tonsils. Fifty-seven patients were excluded because of poor quality of tonsil samples ( Fig. 1). Of those 143 patients, fifty-seven samples were rhinovirus positive in primary PCR test. Of them rhinovirus sequencing was successful from 42 samples (Fig. 1).

Patient characteristics
Median age of study subjects was 9.3 years, 55% were males, and 42% had food or aeroallergen sensitization ( Table 1). Main indications for tonsillectomy were tonsillar hypertrophy in 43% subjects and recurrently infected tonsils in 19% subjects (Table 1). RV-A and -C groups had male dominance. RV-A epidemic occurred in spring and RV-C epidemics in fall, and RV-A infected subjects were most often symptomatic (Table 1).

Discussion
The clinical consequences of rhinovirus A, B and C infections have previously been proven to differ from each other [1]. Thus, it is evident that immunological responses due to these distinct species are diverse. Bearing these things in mind, we conducted our study with routine tonsillectomy patients, which has three main findings. First, rhinovirus species A, B and C were rather equally distributed among relatively asymptomatic tonsillectomy patients and were mainly found in nasopharyngeal aspirate compared to the low persistence rate in tonsils. Second, rhinovirus species were differently associated with tonsillar cytokine responses. Rhinovirus C affected patients had increased IFN-α and decreased RORC2 expression in tonsils when compared to other rhinovirus species (Figs. 2 and 3). Third, rhinovirus species were differently associated with clinical characteristics. The similar incidence rates of RV species was a bit unexpected since RV-B has usually been slightly less prevalent than RV-A and RV-C in healthy subjects [5,7,8]. However, this difference has been more pronounced in children with severe wheezing and asthma, in whom RV-A and RV-C species have also associated with more severe symptoms [2,7,8]. Our children study subjects were rather young (median 9 years old) and generally healthy which is likely to favor more equal distribution of RV species. Symptomatic and asymptomatic infections are also more common in children than in adults [1,11]. We found no difference between rhinovirus A, B or C positive tonsillectomy patients in terms of viral coinfections. It has been shown that viral codetection is common especially in children [16,17]. Morikawa et al. [16] noted that the coinfection rates of RV-A and C are high, but it remains unclear whether they found differences between the coinfection rates of RV-A, RV-B and RV-C. Miller et al. [18] found more coinfections with RV-A than with RV-C in hospitalized children. We found male sex to be dominant in RV-A and RV-C infected patients. This is in line with a previous study with male sex dominating in RV-A and RV-C infections in hospitalized children [14]. Reason for that might be in sex hormones as androgens promote Th1 (T helper type 1 cell) responses and estrogen and progesterone Th2 responses [19]. The seasonality of all the rhinovirus types seemed to follow the seasonality described in literature [2,8,16,20,21]: there was a major peak in fall and a smaller peak in spring, but the rhinoviruses were detected throughout the whole year. RV-C and RV-B were found mostly from patients operated in fall whereas RV-A was most commonly found in spring.
The most interesting and novel finding was the higher level of tonsillar IFN-α production in patients with RV-A and RV-C infection, especially in those with RV-C, compared with RV-B infected patients. Type I interferons, including IFN-α, are important antiviral cytokines and for preventing inappropriate Th2 response [8,22]. Considering the ability of RV-A and RV-C to cause more severe infection than RV-B [1,2], it is plausible that they may also induce stronger interferon responses. In fact, strong IFN-γ responses have been associated with reduced virus shedding, and exposure to IFN-α or IFN-γ has limited RV infection in vitro [6,21]. Our findings are in agreement with these previous findings and extend the increased expression of type I interferon to local lymphoid tissue level in healthy individuals. Moreover, some previous findings suggest that RV-C may cause slightly more severe infection than RV-A [1] which also fit to our finding.
The rate of respiratory symptoms on the operation day was fairly small (33%) thus our patients are representing mostly asymptomatic subjects. More than half had symptoms 2 to 4 weeks prior the operation which might reflect virus shedding from a past symptomatic infection. It may take even 5-6 weeks from rhinovirus to disappear from nasal mucus [25]. Even in healthy children rhinovirus have been found in 15 to 35% in asymptomatic individuals [2,20,21,26,27] giving an explanation of why only a small proportion of rhinovirus infected patients had symptoms on the operation day. Majority of the patients having respiratory symptoms on the operation day had RV-A or RV-C infection and only one had RV-B infection supporting the stronger clinical importance of RV-A and RV-C. The strength of the study was that it consisted relatively healthy tonsillectomy patients, and thereby our results suggest normal cytokine responses due to symptomatic or asymptomatics RV infection. Many previous studies concern hospitalized patients having an acute infection [7,9,14,21,28]. Also, our study population was constructed with both children and adults whereas often studies concern only children. We also used single cytokine PCR and patients as well as their preceding symptoms were well-characterized. There were also some limitations. Sample size was relatively small and because of subpopulation of study subjects the study was not powered to detect significant changes in outcome measures. We studied only viruses, not bacteria. As there was no control group the observations are not clearly associated with the presence of an acute infection. Our results can only be generalized for healthy individuals. Then, the immune response due to RV-infection is a dynamic process and the relevance of history of previous RV infections remains obscure. The samples were taken at a single time point as the tonsils can be removed only once and therefore it is difficult to assess whether the observed changes were time-dependent or uniform over longer period of time.

Conclusions
To our best knowledge, this was the first time to study immune responses between rhinovirus species at the local lymphoid tissue level. We found that rhinovirus species were associated differently with clinical characteristics and tonsillar cytokine responses. The most interesting finding was the increase of IFN-α and Tbet and decrease of RORC2 in rhinovirus C affected patients suggesting that rhinovirus C infection has greater effect on tonsillar Th1 and Th17 responses than rhinovirus A or B infection. Our results encourage to study immune responses at the local lymphoid tissue level by using tonsils.