TMPRSS2 Independency for Haemagglutinin Cleavage In Vivo Differentiates Influenza B Virus from Influenza A Virus.

Influenza A and B viruses show clear differences in their host specificity and pandemic potential. Recent studies have revealed that the host protease TMPRSS2 plays an essential role for proteolytic activation of H1, H3, and H7 subtype strains of influenza A virus (IAV) in vivo. IAV possessing a monobasic cleavage site in the haemagglutinin (HA) protein replicates poorly in TMPRSS2 knockout mice owing to insufficient HA cleavage. In the present study, human isolates of influenza B virus (IBV) strains and a mouse-adapted IBV strain were analysed. The data showed that IBV successfully underwent HA cleavage in TMPRSS2 knockout mice, and that the mouse-adapted strain was fully pathogenic to these mice. The present data demonstrate a clear difference between IAV and IBV in their molecular mechanisms for spreading in vivo.

Influenza B virus (IBV) is classified into the same family, Orthomyxoviridae, as influenza A virus (IAV)1. Both viruses cause seasonal influenza, and the clinical manifestations of the infections are indistinguishable. However, they differ in their antigenic diversities and host ranges. IAV has 18 haemagglutinin (HA) and 11 neuraminidase (NA) subtypes, whereas no antigenic subtypes are found for IBV123. Only two antigenically distinct virus lineages, B/Victoria/2/87 [Victoria] and B/Yamagata/16/88 [Yamagata], have been co-circulating worldwide since 198345. Although the natural hosts of IAV are wild aquatic birds, many subtypes of IAV are maintained in various birds and mammals, including humans1. IAV was also detected in bats23. By contrast, IBV mainly circulates in humans, and occasionally in seals6 and pigs7. To date, the molecular bases generating the biological differences between IAV and IBV are poorly understood.

The HA proteins of both IAV and IBV are synthesized as inactive precursors, and their proteolytic cleavage is a prerequisite for the HA conformational changes essential for virus infectivity. Recent studies using TMPRSS2 gene knockout (TMPRSS2 KO) mice have revealed that IAV strains possessing a monobasic cleavage site in HA primarily undergo proteolytic activation by TMPRSS2, a type II transmembrane serine protease (TTSP), in vivo8910. H3N2, H1N1, and H7N9 IAV strains replicated very poorly in TMPRSS2 KO mice owing to insufficient HA cleavage in these mice8910. Therefore, TMPRSS2 is a potential molecular target for the development of anti-IAV drugs. Meanwhile, cell culture analyses demonstrated that TMPRSS2 also has the potential to activate IBV11. In this study, the role of TMPRSS2 for the in vivo replication of IBV was examined.

Results

Human isolate IBV strains are avirulent but spread similarly in wild-type (WT) and TMPRSS2 KO mice

Human TMPRSS2 (hTMPRSS2) has been shown to cleave IBV HA11. First, the cleavage activity of mouse TMPRSS2 (mTMPRSS2) for IBV HA was analysed and compared with that of hTMPRSS2. HeLa cells constitutively expressing mTMPRSS2 or hTMPRSS2 (HeLa/mTM2 and HeLa/hTM2, respectively) and the parental HeLa cells were infected with the B/Aichi/99[V] strain (Fig. 1A) or mouse-adapted (MA)-B/Ibaraki/85[V] strain (Fig. 1B). At 12 and 24 hours post-infection (p.i.), polypeptides for IBV HA were detected by SDS-PAGE and immunoblotting. Cleaved forms of HA (HA1 and HA2) were clearly detected in HeLa/mTM2 and HeLa/hTM2 cells, but not in the parental HeLa cells, infected with either B/Aichi/99[V] or MA-B/Ibaraki/85[V] (Fig. 1A,B). The cleavage efficiency was similar between HeLa/mTM2 and HeLa/hTM2 cells. When the culture media were supplemented with trypsin, cleaved forms of HA were detected in all three cell lines (Fig. 1A,B). These data indicated that mTMPRSS2 cleaves IBV HA as well as hTMPRSS2.

Figure 1

Cleavage of IBV HA by mTMPRSS2 and hTMPRSS2.

HeLa cells constitutively expressing mTMPRSS2 or hTMPRSS2 (HeLa/mTM2 and HeLa/hTM2, respectively) and the parental HeLa cells were infected with B/Aichi/99[V] (A) or MA-B/Ibaraki/85[V] (B) at an MOI of 1.0 in the presence or absence of trypsin. Mock-infected cells were also prepared. At 12 and 24 hours p.i., IBV HA in cells was detected by SDS-PAGE and immunoblotting.

When challenged with B/Aichi/99[V] and B/Shiga/98[Y], both WT and TMPRSS2 KO mice showed little, if any, body weight loss (data not shown). When challenged with B/Yamagata/88[Y] strain, both WT and TMPRSS2 KO mice showed very mild body weight loss (Fig. 2A). Neither WT nor TMPRSS2 KO mice showed any clinical signs (data not shown). The viral titres in lung lavage fluids and lung homogenates of the mice were analysed. Since these IBV strains have not been adapted to grow in mice, the virus titres were generally low (Fig. 2B,C). Nonetheless, the virus titres in TMPRSS2 KO mice were generally slightly lower than those in WT mice (Fig. 2B,C). However, most of the differences between WT and TMPRSS2 KO mice were not significant (P > 0.05), with only the data for the lung homogenates of B/Yamagata/88[Y] and B/Shiga/98[Y] at 2 days p.i. showing significant differences (P < 0.05) (Fig. 2B). Histopathological analyses demonstrated that the viruses spread moderately in the lungs of both WT and KO mice. Each strain caused slight bronchiolitis at 2 days p.i. and mild alveolitis at 6 days p.i. in both WT and KO mice (Fig. 3). The immunostaining patterns for IBV proteins were similar between WT and KO mice at 2 and 6 days p.i. (Fig. 3). Although these pathological conditions progressed to some extent, they remained mild (Fig. 3). Most importantly, the data revealed no significant differences in the histopathological changes and viral antigen patterns between WT and KO mice infected with IBV.

Figure 2

Body weights of WT and TMPRSS2 KO mice infected with human isolates of IBV and virus titres in these mice.

(A) WT and TMPRSS2 KO mice were intranasally inoculated with 1.6 × 104 PFU of B/Yamagata/88[Y] (n = 6). The body weights were measured daily. Filled and open triangles indicate data of WT and TMPRSS2 KO mice, respectively. Data represent means ± SD. (B,C) WT and TMPRSS2 KO mice (n = 5) were intranasally inoculated with 4.6 × 103, 1.6 × 104, and 1.3 × 104 of B/Aichi/99[V], B/Yamagata/88[Y], and B/Shiga/98[Y], respectively. At 2, 4, and 6 days p.i., the virus titres in lung homogenates (B) and lung lavage fluids (C) were determined. Filled and open bars indicate data of WT and TMPRSS2 KO mice, respectively. Data represent means ± SD. *P < 0.05, significant difference based on a t-test.

Figure 3

Histopathological findings in the lungs of WT and TMPRSS2 KO mice infected with human isolates of IBV, B/Aichi20/99[V] (A), B/Yamagata/88[Y] (B), and B/Shiga/98[Y] (C). Data obtained by hematoxylin-eosin (HE) staining and immunohistochemistry (IHC) for IBV antigens are shown. Original magnification: ×10. The inflammation scores of individual mice (n = 3) are shown: 0, no apparent changes; 1, minimal changes or bronchiolitis; 2, bronchiolitis and/or slight alveolitis; 3, mild alveolitis with neutrophils, monocytes/macrophages, or lymphocytes; 4, moderate alveolitis.

The mouse-adapted B/Ibaraki/2/85 strain is fully activated proteolytically in vivo and shows full pneumopathogenicity in TMPRSS2 KO mice

Next, the replication capacity and virulence of the MA-B/Ibaraki/85[V] strain were analysed. Since the adaptation to mice must cause mutations in the viral genome, the genomic sequences of the parental B/Ibaraki/85[V] and MA-B/Ibaraki/85[V] strains were determined. The genomic sequence of the parental B/Ibaraki/85[V] strain was deposited in the GISAID database with the following GISAID Isolate ID numbers: EPI_ISL_219717; PB2: EPI749072; PB1: EPI749073; PA: EPI749074; HA: EPI749075; NP: EPI749076; NA: EPI749077; MP: EPI749078; NS: EPI749079. The mutations acquired by MA-B/Ibaraki/85[V] are summarized in Table 1. Seven amino acid changes were found in the HA protein. The residues were mapped on the IBV HA structure (PDB 2RFU). All of these residues were located in the globular head domain at positions far distant from the cleavage site (Fig. 4). Both WT and KO mice showed severe body weight loss and required euthanasia, when challenged with MA-B/Ibaraki/85[V]. Similar kinetics of body weight loss were observed for WT and KO mice (Fig. 5A). The survival rates were also similar between WT and TMPRSS2 KO mice (Fig. 5B). At 2 days p.i., the lungs of both WT and KO mice showed similar levels of bronchiolitis with slight alveolitis, and viral antigens were detected throughout the bronchial epithelia (Fig. 5C). At 6 days p.i., moderate alveolitis had progressed in the lungs of both WT and KO mice with viral antigens spreading throughout the lungs (Fig. 5C). The virus titres in the lung lavage fluids and lung homogenates from both WT and KO mice were high during the observation period, although the titres in the lung lavage fluids in WT mice tended to be several-fold higher than those in KO mice (Fig. 6A). However, the differences were not significant (P < 0.05). The extent of activation in vivo of the progeny viruses was analysed. Even in KO mice, the infectivity titres were not restored significantly by in vitro trypsin treatment (Table 2), indicating that the majority of MA-B/Ibaraki/85[V] was produced in an almost fully-activated form in the lungs of both WT and KO mice. To obtain evidence of HA cleavage in vivo, lung lavage fluids collected from infected WT and KO mice (n = 3 for each) were analysed by SDS-PAGE and immunoblotting. Generally, the HA signal intensities were greater in WT mice than in KO mice (Fig. 6B), consistent with the greater amounts of progeny viruses in WT mice compared with KO mice (Fig. 6A). However, it was clearly shown that HA was already cleaved into subunits (HA1 and HA2) in both WT and KO mice (Fig. 6B). The NP signals were also greater in WT mice than in TMPRSS2 KO mice (Fig. 6B). The extents of cleavage estimated by the signal intensities of HA0 and the cleaved HA subunits (HA1 and HA2) were similar in WT and TMPRSS2 KO mice (Fig. 6C). Taken together, in both WT and KO mice, MA-B/Ibaraki/85[V], whose HA possesses a monobasic cleavage site, was almost fully activated proteolytically and replicated in multiple steps in the lungs, resulting in pneumonia and exhibiting high pathogenicity.

Table 1

Nucleotide and amino acid differences between the parental (PT) B/Ibaraki/95[V] and MA-B/Ibaraki/85[V] strains.

Gene/segment	Nucleotide			Protein	Amino acid
	Positiona	PT	MA		Position	PT	MA
PB2	226	A	G	PB2	76	R	G
PB1		None
PA	779	G	A	PA	260	G	E
	1160	A	G		387	Y	C
HA	250	G	A	HA	84	G	R
	466	G	G, A		156	G	G, R
	480	T	C		160	synonymous
	541	A	A, G		181	N	N, D
	610	A	A, G		204	T	T, A
	634	G	G, A		212	D	D, N
	1303	T	C		435	F	L
	1317	A	A, G		439	I	I, M
NP	1176	A	T	NP	392	synonymous
	1251	A	G		417	synonymous
	1543	G	A		515	G	N
	1544	G	A		515	G	N
NA	1200	T	C	NA	400	synonymous
	1371	C	T		457	synonymous
M	847	G, A	G	BM2	34	G, E	G
NS	112	A	G	NS1	38	R	G

aThe number of the first nucleotide of the initiation codons for PB1, PB2, PA, HA, NP, NA, M1, and NS1 proteins is set to ‘nucleotide position 1′. When two nucleotides are detected at one position, both nucleotides and two predicted amino acids are shown.

Figure 4

Predicted locations of the mutated amino acid residues in MA-B/Ibaraki/85[V] HA.

Based on the structure of B/HongKong/8/73 (PDB 2RFU), the predicted locations of the mutated amino acid residues in MA-B/Ibaraki/85[V] HA are shown. The mutated residues are shown in magenta in a sphere model. The cyan circle indicates the position of the cleavage site.

Figure 5

Body weights, survival, and histopathological findings of WT and TMPRSS2 KO mice infected with MA-B/Ibaraki/85[V].

(A) WT and TMPRSS2 KO mice were intranasally inoculated with 2.0 × 103, 2.0 × 104, and 2.0 × 105 PFU of MA-B/Ibaraki/85[V] (n = 5 or 6). The body weights were measured daily. Filled and open triangles indicate data of WT and TMPRSS2 KO mice, respectively. Data represent means ± SD. (B) Survival curves of MA-B/Ibaraki/85[V]-infected mice. Blue and red lines indicate data of WT and TMPRSS2 KO mice, respectively. (C) Histopathological findings in the lungs of WT and TMPRSS2 KO mice infected with MA-B/Ibaraki/85[V]. Data obtained by HE staining and IHC for IBV antigens are shown. Original magnification: ×10. The inflammation scores of individual mice (n = 3) are shown.

Figure 6

Virus titres and HA cleavage in WT and TMPRSS2 KO mice infected with MA-B/Ibaraki/85[V].

(A) WT and TMPRSS2 KO mice (n = 3) were intranasally inoculated with 6.7 × 103 PFU of MA-B/Ibaraki/85[V]. Lung lavage fluids and lung homogenates at 2, 4, and 6 days p.i. were subjected to virus titration. Filled and open bars indicate data of WT and TMPRSS2 KO mice, respectively. Data represent means ± SD. (B) The HA and NP protein levels in lung lavage fluids at 2 dpi were analysed by SDS-PAGE and immunoblotting. Each lane corresponds to data from an individual mouse. The same amounts of lung lavage fluids were loaded. (C) Quantification of cleavage by measuring the chemiluminescent signals for the cleaved HA subunit (HA1 and HA2) signals to total HA (HA1, HA2, and HA0) signals.

Table 2

The extent of activation in vivo of the progeny viruses in lung lavage fluids of mice infected with MA-B/Ibaraki/85[V].

	Titre (PFU/ml)
	Mouse	n = 3	Trypsin+	Trypsin−	Activation (%)a
2 days p.i.	WT	Means	7.2 × 106	6.8 × 106	94.4
		SD	3.3 × 106	1.9 × 106
	KO	Means	3.2 × 106	4.7 × 106	146.9
		SD	1.3 × 106	0.3 × 106
4 days p.i.	WT	Means	5.7 × 106	1.3 × 106	22.8
		SD	1.9 × 106	0.3 × 106
	KO	Means	1.7 × 106	7.7 × 105	45.3
		SD	1.6 × 106	6.5 × 105
6 days p.i.	WT	Means	2.4 × 106	7.2 × 105	30.0
		SD	2.3 × 106	2.8 × 105
	KO	Means	2.9 × 105	1.9 × 105	65.5
		SD	3.1 × 105	1.5 × 105

aThe percentage of virus titres activated in vivo (Trypsin−) in total virus titres activated ex vivo (Trypsin+).

Discussion

Our data show a clear difference in the in vivo protease specificity between IAV and IBV. Many studies have been performed to identify the proteases responsible for HA activation, and many candidate proteases have been reported12. Several TTSPs, such as TMPRSS21314, HAT13, TMPRSS41516, matriptase1718, MSPL19, and DESC119, were also shown to activate IAV strains, although certain TTSPs may activate specific HA subtypes. Recent studies have provided concrete evidence that among these proteases, TMPRSS2 is the main protease for monobasic HA cleavage in IAV strains in vivo8910. Meanwhile, high-pathogenic (HP) avian IAV strains possessing a multibasic HA cleavage site use furin and pro-protein convertase 5/62021. As these proteases are ubiquitously expressed in the trans-Golgi network, HP strains have the potential to spread in a variety of organs and tissues, causing lethal infections in domestic poultry. Thus, HP IAV strains show high virulence, even toward TMPRSS2 KO mice810. Compared with efforts for IAV, only limited data are available for HA cleavage in IBV strains. Previous studies have demonstrated different levels of trypsin-dependency between IAV and IBV for growth in specific MDCK cell lines2223. Noma et al.22 suggested that IBV HA has higher sensitivity to an endoprotease(s), currently unidentified, than IAV HA, or that specific MDCK cell lines express an endoprotease that only activates IBV HA. Further analyses using cell cultures demonstrated that TMPRSS2 activates IBV HA as well as IAV HA11. Therefore, IBV strains may utilize TMPRSS2 for HA cleavage in vivo, similar to IAV strains. Nevertheless, our present data clearly demonstrate that TMPRSS2 is dispensable for HA cleavage and pathogenicity of IBV in mice. However, the virus titres in TMPRSS2 KO mice were slightly lower than those in WT mice. Therefore, potential anti-IAV drugs targeting TMPRSS2 are less effective for IBV. Under certain experimental conditions, H3N2 also showed the ability to use a TMPRSS2-independent HA activation mechanism through loss of an oligosaccharide at the HA stalk region24. The loss of this oligosaccharide likely improved the accessibility of an alternative host protease to the loop containing the HA cleavage site of IAV. The HA protein of IBV also possesses oligosaccharides at the stalk region and near the cleavage site25, but our data in the present study show that they do not limit the activation protease to TMPRSS2. The stalk oligosaccharide of IBV HA attached in vivo may be incompatible with the HA cleavage by TMPRSS2. Because the protease specificity of viruses determines virus tropism and pathogenicity2627, identification of the protease(s) responsible for IBV HA cleavage may reveal a molecular basis for the different biological features between IAV and IBV, and contribute to the development of anti-IBV drugs.

Materials and Methods

Ethics statement

All experiments with animals were performed in strict accordance with the Animal Experimentation Guidelines of the National Institute of Infectious Diseases, and the protocol was approved by the Institutional Animal Care and Use Committee of the institute.

Mice

TMPRSS2 KO mice were reported previously10. Briefly, the mice were generated from TMPRSS2 gene KO C57BL/6 ES cells (KOMP Repository Knockout Mouse Project; Product ID: VG13341) and have a complete C57BL/6 genetic background. WT C57BL/6 mice were purchased from Japan SLC.

Viruses and cells

Four IBV human isolates, two from the Victoria lineage, B/Aichi20/99 (B/Aichi/99[V]) and B/Ibaraki/2/85 (B/Ibaraki/85[V]), and two from the Yamagata lineage, B/Yamagata/16/88 (B/Yamagata/88[Y]) and B/Shiga/T30/98 (B/Shiga/98[Y]), and a mouse-adapted (MA)-B/Ibaraki/85[V], were used. The human strains were isolated in MDCK cells or embryonated eggs. The adaptation process of the MA-B/Ibaraki/85[V] strain was reported previously28. HeLa cells constitutively expressing mouse TMPRSS2 and human TMPRSS2 (HeLa/mTM2 and HeLa/hTM2) were reported previously10.

Infection of mice with human isolates of IBV

WT and TMPRSS2 KO mice (6–7-week-old) were infected intranasally with 4.6 × 103, 1.6 × 104, and 1.3 × 104 PFU of B/Aichi/99[V], B/Yamagata/88[Y], and B/Shiga/98[Y], respectively. For mock infection, the same volume (20 μL) of PBS was used. The body weight and clinical signs were monitored daily. For histopathological analysis, WT and TMPRSS2 KO mice were euthanized and autopsied at 2 days p.i. (n = 3) or 6 days p.i. (n = 3). The viral antigens of IBV were detected using a mouse antiserum against IBV NP (Lot. N179). This antiserum was obtained from BALB/c mice immunized intraperitoneally with inactivated IBV virions (Victoria lineage) using the Sigma adjuvant system (Sigma-Aldrich). The inflammation levels in individual mice were scored as follows: 0, no apparent changes; 1, minimal changes or bronchiolitis; 2, bronchiolitis and/or slight alveolitis; 3, mild alveolitis with neutrophils, monocytes/macrophages, or lymphocytes; 4, moderate alveolitis. The cut-off in body weight loss for euthanasia was 25%. Lung lavage fluids and lung homogenates were collected at 2, 4, and 6 days p.i. and subjected to a standard plaque assay.

Infection of mice with MA-B/Ibaraki/85[V]

WT and TMPRSS2 KO mice (6–7-week-old) were infected intranasally with 2.0 × 103, 2.0 × 104, and 2.0 × 105 PFU of MA-B/Ibaraki/85[V]. These doses of MA-B/Ibaraki/85[V] correspond to 30, 300, and 3,000 mouse lethal dose 50 (MLD50) for WT mice. The body weight and clinical signs were monitored daily. For histopathological analysis, similar mice were also infected intranasally with 6.7 × 103 PFU of MA-B/Ibaraki/85[V] and autopsied at 2 days p.i. (n = 3) or 6 days p.i. (n = 3).

Analysis of infectious virus titres activated in vivo

WT and TMPRSS2 KO mice (n = 3) were infected with 2.0 × 103 PFU of MA-B/Ibaraki/85[V] and lung lavage fluids were collected at 2, 4, and 6 days p.i. Monolayers of MDCK cells were infected with serially diluted lung lavage fluid and lung homogenate samples for 1 hour at 4 °C, washed twice with PBS, overlaid with MEM/1% agarose, and incubated for 24 hours at 37 °C to allow the viruses to enter the cells. Trypsin was omitted to avoid HA cleavage before virus entry. At 24 hours p.i., the cell monolayers were additionally overlaid with MEM/1% agarose supplemented with 4.0 μg/mL of trypsin to allow plaque formation. To determine the infectious titres including viruses that had not been activated in vivo but possessed an infectious potential, virus samples were treated with 2.0 μg/mL of trypsin, and subjected to a standard plaque assay.

Immunoblotting

WT and TMPRSS2 KO mice were infected with 6.7 × 103 PFU of MA-B/Ibaraki/85[V] (n = 3) or mock-infected (n = 1). Lung lavage fluids were collected at 2 days p.i. HeLa/mTMPRSS2, HeLa/hTMPRSS2, and parental HeLa cells were infected with B/Yamagata/88[Y], B/Shiga/98[Y], MA-B/Ibaraki/85[V], and parental B/Ibaraki/85[V] at a multiplicity of infection (MOI) of 1.0, and collected at 12 and 24 hours p.i. The lung lavage fluids and cells were lysed with lysis buffer to make a final solution containing 150 mM NaCl, 50 mM Tris-HCl (pH 7.5), 4 mM EDTA, 0.1% sodium deoxycholate, 1% Nonidet P-40, and 0.1% SDS. The polypeptides were separated by SDS-PAGE in 10–20% polyacrylamide gels (e-PAGEL, E-T/R1020L; ATTO), and blotted onto PVDF membranes (WSE-4051; ATTO). Rabbit antisera raised against IBV HA (#11053-RP01; Sino Biological Inc.) and NP (#GTX128539; GeneTex Inc.) were used for detection of HA and NP, respectively.

Additional Information

How to cite this article: Sakai, K. et al. TMPRSS2 Independency for Haemagglutinin Cleavage In Vivo Differentiates Influenza B Virus from Influenza A Virus. Sci. Rep. 6, 29430; doi: 10.1038/srep29430 (2016).