Pathogenesis of emerging severe fever withthrombocytopenia syndrome virus in C57/BL6mouse modelCong Jina,1, Mifang Lianga,1, Junyu Ningb,1, Wen Gua, Hong Jiangc, Wei Wua, Fushun Zhanga, Chuan Lia, Quanfu Zhanga,Hua Zhuc, Ting Chenc, Ying Hana, Weilun Zhangc, Shuo Zhanga, Qin Wanga, Lina Suna, Qinzhi Liua, Jiandong Lia,Tao Wanga, Qiang Weic, Shiwen Wanga, Ying Dengb, Chuan Qinc, and Dexin Lia,2

aLaboratory of Viral Hemorrhagic Fever, National Institute for Viral Disease Control and Prevention, China Center for Disease Control Beijing 102206, China;bDepartment of Toxicology, Beijing Center for Disease Control and Prevention, Beijing 100013, China; and cLaboratory of Virus, Institute of Laboratory AnimalScience, Chinese Academy of Medical Sciences, Beijing 100021, China

Edited by Diane E. Griffin, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, and approved May 1, 2012 (received for review December10, 2011)

The discovery of an emerging viral disease, severe fever withthrombocytopenia syndrome (SFTS), caused by SFTS virus (SFTSV),has prompted the need to understandpathogenesis of SFTSV.Weareunique in establishing an infectious model of SFTS in C57/BL6 mice,resulting in hallmark symptoms of thrombocytopenia and leukocyto-penia. Viral RNAandhistopathological changeswere identified in thespleen, liver, andkidney. However, viral replicationwas only found inthe spleen, which suggested the spleen to be the principle targetorgan of SFTSV.Moreover, the number ofmacrophages and plateletswere largely increased in the spleen, and SFTSV colocalized with pla-telets in cytoplasm of macrophages in the red pulp of the spleen. Invitro cellular assays further revealed that SFTSV adhered to mouseplatelets and facilitated the phagocytosis of platelets by mouse pri-mary macrophages, which in combination with in vivo findings, sug-gests that SFTSV-induced thrombocytopenia is caused by clearance ofcirculating virus-bound platelets by splenic macrophages. Thus, thisstudy has elucidated the pathogenic mechanisms of thrombocytope-nia in a mouse model resembling human SFTS disease.

animal model | phlebovirus | pathology | reduction of platelets |macrophage infection

Severe fever with thrombocytopenia syndrome (SFTS) is a re-cently identified emerging viral infectious disease in China that

is caused by a novel phlebovirus in the family Bunyaviridae, SFTSvirus (SFTSV) (1). Clinical features of SFTSpatients include abrupthigh fever, thrombocytopenia, leukocytopenia, and gastrointestinalsymptoms. Laboratory tests commonly show elevated serum levelsof alanine aminotransferase (ALT), aspartate aminotransferase(AST), blood urea nitrogen (BUN), lactate dehydrogenase, crea-tine kinase, creatine kinase MB fraction, as well as elongated acti-vated partial-thromboplastin time (1). These abnormally changedlaboratory parameters are indicative of the pathological lesions thatoccur inmultiple organs and altered homeostasis of the coagulationsystems of SFTS patients. Pathological studies of SFTS are absentbecause patient tissues are rarely donated after death in rural areasof China. Therefore, the causes of illness and death, as well aspathological changes within organs, remain unclear. To systemati-cally investigate the pathogenic mechanisms of SFTS and un-derstand key symptoms, such as thrombocytopenia, infectiousanimal models for SFTSV are urgently needed.

ResultsEstablishment of a SFTSV Pathogenic Mouse Model. In our initialstudy, to identify an infectious animal model that could mimicmost clinical features during SFTSV infection, the susceptibilitiesof three commonly used laboratory rodent strains for phlebovirus(2–4), C57/BL6 mice, BalB/C mice, and Syrian hamsters, wereexamined. The SFTSV strain HB29 was inoculated at 105 TCID50(50% tissue culture infective dose) per mouse or 5 × 105 TCID50per hamster through four different routes of infection, including

intravenous, intramuscular, intraperitoneal, and intracerebralinjections. The results showed that in C57/BL6 mice, inoculationthrough all four routes induced a reduction of white blood cells,and the inoculation intravenously and intramuscularly induceda reduction of platelets (Fig. S1), but abnormal medical signs andweight loss were not observed. Inoculation of SFTSV to BalB/Cmice or hamsters through all of the above routes could not causesignificant changes in weight, white blood cell counts, and plateletcounts. However, pathology examination on day 14 showed that allthese SFTSV-infected rodents developed similar pathologicallesions in liver and kidney [Fig. 1, representative pathology at day14 postintion (p.i.) from C57/BL6 mice]. Based on a comprehen-sive evaluation, we determined that SFTSV-infected adult C57/BL6 mice mimicked the major clinical features of leukocytopeniaand thrombocytopenia in SFTS patients and, therefore, were se-lected for further investigations. Because SFTSV infection in thehuman is thought to be transmitted by tick bite (1), we chose in-tramuscular inoculation for subsequent animal studies.Next, to thoroughly examine the complete disease progression of

SFTSV infection in C57/BL6 mice, a group of 10 mice with in-tramuscular inoculation of 105 TCID50 SFTSV per mouse anda group of 5 mock-infected mice were killed at each time point ofdays 1, 3, 7, 14, 21, and 28 p.i. for analysis of body weight, tem-perature, various laboratory parameters, and pathology. Similarly,as in initial studies, SFTSV inoculation caused significantly reducedplatelet count on day 3 p.i., as well as reduced white blood cellcount on day 1 and day 3 p.i. (Table 1). Additionally, serum ASTelevated on day 1 and lasted till day 7, and serum ALT and BUNelevated during 7–14 d (Table 1). Therefore, we showed that in-tramuscular inoculation of SFTSV intoC57/BL6mice could inducesymptoms that mimic the thrombocytopenia, leukocytopenia, anddysfunction in liver and kidney that are key clinical presentations ofSFTS patients. However, fever and weight loss were not induced inSFTSV-infected C57/BL6 mice (Table 1), and all tested mice sur-vived. To evaluate if the immunocompromised animals could de-velop a more severe disease and even fatal outcome, we used im-munosuppressive drug mitomycin C, which is known to inhibit thehematopoiesis in bone marrow (5), to treat SFTSV-infected mice.We found that with mitomycin C application, 50% of theSFTSV-infected mice died between day 9 and day 10, and theremaining surviving mice experienced a significant loss of body

Author contributions: C.J., M.L., J.N., Q. Wei, Y.D., C.Q., and D.L. designed research; C.J.,M.L., J.N., W.G., H.J., W.W., F.Z., C.L., Q.Z., H.Z., T.C., Y.H., W.Z., S.Z., Q. Wang, L.S., Q.L., J.L.,T.W., and S.W. performed research; C.J., M.L., J.N., W.G., Y.H., and W.W. analyzed data;and C.J., M.L., J.N., and D.L. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1C.J., M.L., and J.N. contributed equally to this work.2To whom correspondence should be addressed. E-mail: lidx@chinacdc.cn.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1120246109/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1120246109 PNAS | June 19, 2012 | vol. 109 | no. 25 | 10053–10058

MICRO

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Aug

ust 8

, 202

1

mass compared with uninfected mice administered mitomycinC in parallel (Fig. S2).To define the infectious status of the immunecompetent mouse

model, the dissemination and replication of SFTSV in blood andtissues were examined by detection of both viral RNA copies andinfectious viral titers (1, 6). The results showed that blood viralload peaked on day 1 p.i. and substantially decreased on day 3 p.i.(Table 1). Dynamic virus distribution in eight tissues, including thespleen, liver, kidney, lung, intestine, heart, muscle, and brain, werealso analyzed. Viral RNA was detected in three organs, includingthe spleen, liver and kidney, and the viral RNA copies per milli-gram of tissue in the spleen were significantly higher than in theliver or the kidney. Infectious viral titers were only detected in thespleen and kidney. This difference between the viral copies andinfectious titers might be a result of the small amount of virus inliver being below the limit of detection for assays to determineinfectious titers. We also noticed that there was a marked increaseof viral load in spleen on day 3 p.i., and the virus RNA copiescould be detected at a low level till to day 21 p.i. (Table 1). Thisfinding suggested active viral replication occurred and remained atcertain level in spleen. In the liver and kidney, the viral RNAcopies initially were elevated on day 1 p.i. and then cleared by day7 p.i. On the whole, the dynamic analysis of SFTSV disseminationin the blood and tissues showed that intramuscularly inoculatedSFTSV entered the blood on day 1 p.i. and then were enriched inspleen on day 3 p.i., where the virus underwent a short term of

replication, followed by clearance to a low level, thereafter. Thesustained detection of low viral RNA copies in the spleen suggestsa possible long-term harboring of SFTSV in spleen, contrastingthe fast kinetics of viral clearance in other organs.To further define immune responses induced by SFTSV in-

fection in C57/BL6 mice, SFTSV-specific IgM and IgG antibodiesin sera were tested by Luminex assays (7). The data showed thatSFTSV-specific IgM peaked on day 7 p.i., and the SFTSV-specificIgG started to increase on day 7 p.i. and peaked on day 28 p.i.(Table 1). The levels of neutralizing antibodies and spleniccellular responses in SFTSV-inoculated mice were significantlyenhanced over mock-infected mice on day 28 p.i. (Table 1). Thisevidence of efficiently elevated humoral and cellular immuneresponses, combined with evidence of sustained viral replicationin mice, support that C57/BL6 mice experienced a true infectionwith SFTSV and that they provide a useful parallel to SFTSpatients for studying SFTSV pathogenesis.The dose-dependent effect of SFTSV inoculation was also

assessed by intramuscular inoculation of a low dose of 103 TCID50SFTSV per C57/BL6 mouse. Low dose of SFTSV did not inducea significant decrease of circulating platelets on day 3, but causeda significantly decreased white cell count on day 1 (Table S1).Similarly, this low dose of SFTSV induced viremia on day 1, and thevirus load peaked in spleen on day 3, but the levels of viral loadwereless than the levels observed in high dose of SFTSV infection(Table S1). Viral copies were hardly detected in liver or kidney

WP

RP

Sco

re

Days PI

Cellularity in the red pulp

Cel

l num

ber/2

0x fie

ld

Days PI

Spleen megakaryocytes

Impa

ired

glom

erul

us(%

)

Days PI

Live

r inj

ury (

%)

Days PI

Liver pathology

1 3 7 14 21 280

2

4

6

1 3 7 14 21 280

5

10

15

1 3 7 14 21 280

20

40

60

80

100

1 3 7 14 21 280

10

20

30

Kidney pathology

* *

* * * * *

**

*

Mock SFTSV

1 3 7 14 21 280

10

20

30

Days PI

BM megakaryocytes

Cel

l num

ber/2

0x fie

ld * * * *

A

B

C

D

E

WPRP

RP

WP

RP

Mock mice SFTSV-infected mice

Fig. 1. Dynamic profile of pathologicalchanges in mice infected with SFTSV.Representative H&E-stained tissue sec-tions from SFTSV-infected mice andmock mice are shown. Areas of interest(AOI) are enlarged, and quantitativegraphs are presented at the right. (A)Decreased cellularity in the red pulp (RP)of the spleen at day 1 p.i.; no obviouschanges are visualized in the white pulp(WP). (B) Increased megakaryocytes inthe spleen at day 3 p.i.. Arrowheads in-dicate megakaryocytes. (C) Increasedmegakaryocytes in the bone marrow atday 3 p.i. Arrowheads indicate mega-karyocytes. (D) In liver at day 14 p.i.,arrowheads show hepatocyte necrosiswith shrinking nucleic or hepatocytedegradation with balloon-like emptycytoplasm. (E) In the kidney at day 14p.i., arrowheads show impaired renalcapsules. In A and B, the images and AOIare 100× and 200×, respectively. In C–E,the images and AOI are 200× and 400×,respectively. In quantitative graphs, reddots and blue dots indicate SFTSV-infec-ted mice and mock mice, respectively,and lines indicate the means. *P < 0.05.

10054 | www.pnas.org/cgi/doi/10.1073/pnas.1120246109 Jin et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 8

, 202

1

tissue. Low-dose SFTSV induced similar levels of virus specific IgMand IgG as high dose of virus (Table 1). Because a low dose ofSFTSV induced insignificant clinical signs, low viral load, and highhumoral immune responses, this finding suggests that the viral loadis associated with the severity of the disease and that a low amountof SFTSV could result in an asymptomatic infection.

Pathological Lesions in SFTSV-Infected Mice. Pathological changeswithin organs were evaluated in the eight tissues indicated above,as well as in the bone marrow, by H&E staining of tissue samplescollected at various time points postintramuscular inoculation.During SFTSV infection, pathological changes were mainlyidentified in the spleen and bone marrow within 3 d p.i. Thelymphocyte cellularity of the red pulp was visually decreased inspleens of SFTSV infected mice during the first week afterinoculation and gradually recovered 14 d later (Fig. 1A). Thisobservation of lymphocyte depletion in the red pulp may co-incide with the systemic decrease of white blood cells. In addi-tion, at an early stage of SFTSV infection, a marked increase ofmegakaryocytes was observed in the spleen, a secondary hema-poietic organ inmice, based on their cellularmorphology (Fig. 1B).This finding suggests extramedullary hematopoiesis, which hasbeen reported to occur in conjunction with decreased lymphocytecellularity of the red pulp (8). Similarly, megakaryocytes in the

bone marrow, the principle organ for hematopoiesis, also signifi-cantly increased at the early stage of infection as they did in thespleen (Fig. 1C). Because megakaryocytes are progenitor cells forplatelets, we presume that the rapid increase of megakaryocytes inthe hemopoietic organs of the spleen and bonemarrow functionedto compensate for the depletion of circulating platelets. Mega-karyocytes have extended survival; therefore, once they have pro-liferated, they could persist in organs for a long period (9). Duringthe late phase of SFTSV infection, pathological changes werenoted in liver and kidney. The primary lesions in liver consisted ofballooning degeneration of hepatocytes and scattered necrosis, thelatter indicated by appearing multifocal pyknosis, karyorrhexis,and karyolysis (Fig. 1D). TUNEL assays on liver tissues detectedfew TUNEL-positive nuclei, indicating that hepatocyte apoptosiswas not induced. The kidney showed glomerular hypercellularity,mesangial thickening, and congestion in Bowman’s space, butinfiltration of inflammatory cells was absent (Fig. 1E). Patho-logical changes within the liver and kidney tissues were maximalon day 14 p.i. but had nearly recovered on day 28 p.i. Therefore,the pathological features observed in the spleen and bone mar-row during the early stage of infection were consistent with thehematological changes of thrombocytopenia and leukocytopenia.The transient pathologic changes observed in kidney and liver at

Table 1. Establishment of a SFTSV pathogenic mouse model

Tests (unit)

Days postinoculation†

Mock‡Day 1 Day 3 Day 7 Day 14 Day 21 Day 28

Clinical signsFever (°C) 37.2 ± 0.6 36.5 ± 1.1 36.8 ± 0.8 36.8 ± 0.9 37.4 ± 0.8 37.8 ± 0.4 37.7 ± 0.7Platelet count (109/L) 862.2 ± 250.4 597.4 ± 194.0*** 932.4 ± 225.5 898.2 ± 345.9 961.2 ± 189.5 928.7 ± 195.2 990.9 ± 167.4White blood cellcount (109/L)

6 ± 2.1*** 8.4 ± 1.8* 13.9 ± 2.0 13.4 ± 1.9 13.8 ± 1.9 13.6 ± 2.8 12.1 ± 2.2

AST(U/L) 95.3 ± 8.7*** 83.1 ± 12.7*** 49.0 ± 5.3* 40.8 ± 5.5 34.5 ± 4.7 N/A 29.4 ± 2.3ALT (U/L) 1.9 ± 0.9 1.4 ± 0.4 8.9 ± 0.4*** 9 ± 1.2*** 3.5 ± 1.8 N/A 1.6 ± 0.3BUN (μM) 3.3 ± 0.5 5.4 ± 1.2* 6.4 ± 1.1** 5.6 ± 1.4** 4.1 ± 1.7 N/A 1.9 ± 0.2

Virus detectionBlood virus load(copies/mL)

74,034.3 ± 21,012.6*** 6,393.1 ± 2,662.1** 81.2 ± 30.6 74.0 ± 31.9 86.1 ± 62.4 78.9 ± 57.9 80.5 ± 37.9

Blood infectioustiters (Log10

TCID50/mL)

4.3 ± 0.8 3.2 ± 0.6 <3 <3 <3 N/A <3

Spleen virus load(copies/mg)

608.7 ± 350.4*** 9,134.6 ± 3,469.7*** 231.5 ± 169.3*** 26.0 ± 20.3* 109.4 ± 44.8** N/A 0.03 ± 0.02

Spleen infectioustiters (Log10

TCID50/0.1 g)

4.2 ± 0.7 5.6 ± 0.7 3.8 ± 0.5 <3 <3 N/A <3

Liver virus load(copies/mg)

1.3 ± 0.8** 1.4 ± 1.2** 0.08 ± 0.02 0.19 ± 0.18 0.13 ± 0.07 N/A 0.12 ± 0.11

Liver infectioustiters (Log10

TCID50/0.1g)

<3 <3 <3 <3 <3 N/A <3

Kidney virus load(copies/mg)

219.1 ± 105.6*** 61.9 ± 37.5*** 2.2 ± 0.9 2.3 ± 1.7 3.8 ± 2.4 N/A 1.4 ± 0.5

Kidney infectioustiters (Log10

TCID50/0.1g)

3.8 ± 0.5 3.2 ± 0.6 <3 <3 <3 N/A <3

Immune responsesSerum IgM (MFI)§ 15.0 ± 7.2 625.1 ± 542.0*** 2,136.8 ± 589.8*** 591.1 ± 255.1*** 296.2 ± 117.0** 128.7 ± 47.5 17.7 ± 10.1Serum IgG (MFI)§ 4.6 ± 1.5 51.7 ± 41.3 1,173.5 ± 272.7*** 3,420.3 ± 1,011.8*** 4,237.7 ± 248.8*** 3,944.1 ± 464.6*** 4.4 ± 2.0Neutralizing antibody(titer)

N/A N/A N/A N/A N/A 38.9 ± 29.3*** 1.03 ± 0.03

Cellular responses(spots/106 cells){

N/A N/A N/A N/A N/A 281.4 ± 137.1*** 3.7 ± 2.9

N/A, data not available. <3 for virus titer were considered below the limit of detection. ***P < 0.001, **P < 0.01, *P < 0.05, SFTSV-inoculated mice vs. mockmice.†At each time point, 10 mice each inoculated with 105 TCID50 SFTSV were analyzed and the results were presented as mean ± SD.‡Results from 30 mock mice were presented as mean ± SD.§Serum SFTSV-specific IgM and IgG responses were evaluated by Luminex immunoassay and presented as medium fluorescent index (MFI).{Cellular responses were evaluated by ELISPOT assay to test IFN-γ releasing from splenic cells and presented as spots/106 cells.

Jin et al. PNAS | June 19, 2012 | vol. 109 | no. 25 | 10055

MICRO

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Aug

ust 8

, 202

1

the later stage of infection were indicative of acute glomerularnephritis and acute hepatitis with self-limiting outcomes.

Identification and Colocalization of SFTSV, Macrophages, and Plateletsin the SFTSV-Infected Spleen. The findings that viral RNA wasenriched in the spleen and that remarkable pathological changesoccurred in the spleen during the early stage of infection suggestedthat the spleen was the primary target organ of SFTSV. Therefore,we performed an immunohistochemistry assay using a polyclonalantibody against SFTSV virions to detect target cells for viralreplication, using spleens collected on day 3 p.i. The resultsshowed that, in the spleen of SFTSV-infected mice, virus-positivecells were large monocytes, suggested by their large cytoplasm,lightly stained nuclei, and their scattered location throughout thered pulp (Fig. 2A). Immunehistochemistry staining for SFTSVwasnegative in bonemarrow, kidney, liver, lung, gastrointestinal tract,heart, muscle, and brain. Based on themorphology and location ofvirus-positive monocytes in the spleen, macrophages were hy-pothesized to be the target cell of SFTSV. Therefore, splenicmacrophages were identified by staining for the macrophage-specific membrane marker F4/80 (10). These results showeda substantially increased number of macrophages in the red pulpof the spleen on day 3 p.i. It is known that splenic macrophages areresponsible for clearing old or impaired platelets from blood (11,12). Based on these observations, we hypothesized that the over-whelmingly increased number of macrophages appearing in thespleen function to clear platelets excessively from blood, resultingthrombocytopenia. We expected that, if this were the case, moreplatelets would be deposited in the spleens of SFTSV-infectedmice. Therefore, splenic sections from tissues isolated day 3 p.i.were probed for platelets using an antibody against CD62P, anactivation marker of platelets (11). The results showed an in-creased deposition of CD62P+ platelets in cytoplasm of themacrophage-like monocytes located in the red pulp of SFTSV-infected spleens (Fig. 2A), which coincided with increased mac-rophage density in the spleen.To further clarify whether the SFTSV colocalized monocytes

were macrophages and whether the increased splenic macrophageswere engaged in platelet clearance, confocal microscopy was usedto examine the colocalization of SFTSV, macrophages, and plate-lets. The results showed that almost all increased CD62P+ plateletsin spleen colocalized with F4/80+ macrophages and, moreover,

SFTSV was found to colocalize with CD62P+ platelets and F4/80+macrophages in spleen (Fig. 2B). Therefore, splenic macrophageswere identified as target cells for SFTSV replication, and the in-creased numbers of splenic macrophages during the early stage ofSFTSV infection appear to promote thrompocytopenia as a resultof their responsibility of phagocytosing platelets.

SFTSV Adherence on Platelets Enhances Phagocytosis of Platelets byMacrophages. To investigate the mechanistic interactions betweenSFTSV, platelets, and macrophages, mouse primary macrophageswere used to test virus infection in vitro. The results showed thatSFTSV could efficiently infect and replicate in mouse primarymacrophages (Fig. 3A). Because some Bunyaviruses have beenreported to bind to platelets (13), we hypothesized that platelet-binding by SFTSV could promote their activation and phagocy-tosis by macrophages. To clarify this hypothesis, serially dilutedSFTSV virions were cocultured with a fixed amount of mouseplatelets, after which unbound virions were washed away. By usingquantitative real-time PCR to determine viral copies in plateletpellets, a gradient of viruses were found to adhere on plateletscorresponding to the serial dilutions (Fig. 3B, Left). The virionsthat had adhered to platelets were further cultured in medium forstudies to address their infectivity after platelet binding. Ourresults showed that the platelet-adherent virus dissociated into thesupernatants after culturing for 1 d (Fig. 3B, Right). Additionally,the total amount of virus detected in the platelets and super-natants did not increase over time, suggesting that no virus repli-cation occurred in platelets.Moreover, mouse primary macrophages were mixed with

mouse platelets that were previously cultured with or withoutSFTSV, respectively. After 30 min of culturing, macrophageswere washed to remove unbound platelets. The results showedthat, although normal platelets could adhere on the surface ofuninfected macrophages, they were rarely internalized (Fig. 3C,Top). In contrast, platelets with adherent SFTSV were rapidlyphagocytosed by uninfected macrophages (Fig. 3C, Middle). TheSFTSV-infected mouse macrophages did not show a significantdifference in their phagocytosis of platelets (Fig. 3C, Bottom).We also tested the macrophage phagocytosis of virus-adheredplatelets using human macrophage cell line THP-1, differenti-ated by phorbol-12-myristate-13-acetate (11, 12, 14). Similarly,we found that SFTSV could infect human THP-1 cells and

BMock mice SFTSV-infected mice SFTSV

Macrophage

Platelet

ASFTSV

Macrophages

Platelets

Merge

Fig. 2. Identification and colocalization of SFTSV,macrophages, and platelets in SFTSV-infectedspleen. (A) SFTSV, macrophages, and platelets wereidentified by immuno-histochemistry in the spleenof Mock mice or SFTSV-infected mice at day 3.Representative images and enlarged AOI are takenat an original magnification of 200× and 400×, re-spectively. (B) Confocal microscopy to examinecolocalization of SFTSV (in green), macrophages (inred), and platelets (in blue) in the SFTSV-infectedspleen collected on day 3 p.i. The representativefluorescent images and embedded AOI were origi-nally taken at 200× and 400×, respectively.

10056 | www.pnas.org/cgi/doi/10.1073/pnas.1120246109 Jin et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 8

, 202

1

adhered on human platelets. Furthermore, only SFTSV-adheredplatelets but not normal platelets could be phagocytosed byTHP-1 cells, and SFTSV-infected THP-1 cells were incapable ofphagocytosing normal platelets (Fig. S3). Therefore, the adher-ence of SFTSV on platelets is necessary for platelets to bephagocytosed by macrophages. Although we observed thatmacrophages could be directly infected by SFTSV, we expectthat the phagocytosis of virus-decorated circulating plateletswould deliver larger amounts of virus into macrophages thandirect infection, while concurrently promoting the internalizationof platelets.

DiscussionSFTSV is a novel pathogenic phlebovirus in the Bunyaviridaefamily (1). Other pathogenic phleboviruses, such as rift valleyfever virus (RVFV) and Punta Toro virus, mainly cause hepaticinjury and viral antigens are detected in mouse liver lobe duringthe early stage of the infection (15). In SFTSV-infected mice, thespleen was the principle target organ of SFTSV at the early in-fection stage, serving as a place for virus replication and showingmarked pathological changes. In addition to the spleen, the liverand the kidney were also targeted by SFTSV. Although virusreplication was not found in the liver or the kidney, elevatedserum levels of AST, ALT, and BUN indicated dysfunction inthese organs, and pathological changes were observed in the latephase of SFTSV infection that were consistent with the clinicalpresentation of elevated transaminases and renal symptoms ofoliguria and anuria in SFTS patients. However, in this patho-genic mouse model, expected temperature elevations were notobserved, and no lesions in the heart were observed, which wasexpected because some SFTS patients had elevated myocardialenzymes. Thus, using SFTSV infection of C57/BL6 mice, wehave established a mouse model that mimics most major clinicalfeatures of SFTS patients, but still has certain limitations.In our initial studies, we found SFTSV could induce reduction

of white blood cells and platelets in C57/BL6 mice, but not inBalB/C mice or Syrian hamsters, which might be because of thevaried genetic backgrounds of these rodent strains. In fact, sev-eral phleboviruses were reported to have selective susceptibilityin animal hosts because of the existence of genetic determinants,such as during RVFV infection (16, 17) and Punta Toro virusinfection (18). Inoculation of virus through the subcutaneous andintramuscular routes have been commonly used for animalstudies of bunyaviruses (4, 19, 20). In our study, we observed thatthe effect of intramuscular injection of SFTSV was comparableto subcutaneous infection (Fig. S1). Therefore, by testing variousrodent strains and infection routes, we identified intramuscularinfection of 105 TCID50 SFTSV in C57/BL6 mice to be a modelsuitable for further investigation on pathogenesis of SFTSVinfection.Thrombocytopenia is a common clinical presentation of he-

matological changes during infections of viral hemorrhagic fever(VHF) viruses in family of Bunyaviridae. However, the mecha-nisms of thrombocytopenia in VHFs, and especially in bunyavi-rus infection, have been elusive. Possible mechanisms forbunyavirus-induced thrombocytopenia have been suggested toinclude increased platelet consumption in damaged tissues (13,21, 22), decreased survival time of platelets (23, 24), or infectionof megakaryocytes, resulting in decreased platelet production(23, 25). Intriguingly, we found in this study that SFTSV colo-calized with phagocytosed platelets in cytoplasm of splenicmacrophages. Furthermore, using in vitro cellular assays, ourdata suggested that SFTSV could adhere on platelets and directthe platelets to be recognized and phagocytosed by macrophagesin the red pulp of spleen. In support of this mechanism, we foundthat SFTSV antigen was not identified in bone marrow, and thisexcludes the possibility of decreased production of platelets bySFTSV-infected hematopoietic progenitor cells. We also ex-cluded the possibility of increased platelet consumption indamaged tissues. During thrombocytopenia by this mechanism,platelets are activated through encounter with extracellularmatrix, which is normally sequestered beneath an intact endo-thelium. Once activated, these platelets adhere to the extracel-lular matrix, resulting in platelet aggregation (26) and leading todecreased circulating platelets and increased deposition of pla-telets in tissues. However, our results showed that the platelet-positive staining was only localized to the spleen and not withinother organs of SFTSV-infected mice. In this case, the decreasein circulating platelets during SFTSV infection is less likely to bea result of generalized deposition in tissues. Therefore, the en-hanced clearance of virus-bound platelets promoted by splenicmacrophages appears to be the major cause of thrombocytope-nia in SFTSV-infected mice, which differs from the currently

SFTSV|Nuclear

Platelet Mouse primary macrophage Merge

+SFTSV

+SFTSV

A

B

C

Platelet|Nuclear

102101 103 104 105 106

Vira

lcop

ies/

107 p

late

lets

Virus amount (TCID50)0 1 3

0

5.0×5

1.0×6

1.5×6 Platelet pelletsPlatelet supernatant

Days

Vir

alc

op

ies

0 1 3 7 100

1.0×8

2.0×8

3.0×8

4.0×8

Days

Vir

alc

op

ies

0

2.0×4

4.0×4

6.0×42.0×54.0×56.0×5

Macrophage cellsMacrophage supernatant

Fig. 3. Phagocytosis of SFTSV-bound platelets by macrophages. (A) Theconfocal image shows viral N protein (in green) in mouse primary macro-phages (in blue). The graph shows dynamic virus replication in mouse pri-mary macrophages and the virus released into culturing supernatants. (B)Real-time PCR quantification detected a gradient of adherence of SFTSV onmouse platelets corresponding to the amount of virus added into platelets.The graph on the right shows the virus amount detected in cultured virus-adhered platelets as well as in culturing supernatants at the indicated timepoints. (C) Confocal images show mouse primary macrophages coculturedwith normal platelets (Top), mouse primary macrophages cocultured withplatelets with adherent SFTSV (Middle), mouse macrophages previouslyinfected by SFTSV for 3 d before coculturing with normal platelets (Bottom).All images were taken at an original magnification of 400×. Quantificationof macrophage phagocytosis of fluorescent platelets is shown in Fig. S2D.

Jin et al. PNAS | June 19, 2012 | vol. 109 | no. 25 | 10057

MICRO

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Aug

ust 8

, 202

1

known mechanisms of bunyavirus-induced thrombocytopenia. Itwas reported that pathogenic hantaviruses, such as Andes virusor Hantaan virus, could bind quiescent platelets through β3integrin, and this binding promoted the adherence of platelets onendothelial cells (13). However, so far it is not clear by whichmechanism SFTSV-bound platelets are recognized and internal-ized by macrophages.Many VHF viruses in the bunyaviridae family, such as RVFV

and Hantaan virus, can also infect monocytes/macrophages (27,28). Because monocytes are primary target cells for VHF viruses,these cells probably modulate either the spreading or contain-ment of viral infection into cells in other organs. In this study, wefound that SFTSV can infect and replicate in macrophagesin vivo and in vitro, but the splenic viral load decreased to a lowlevel later in infection. This finding suggests that although virus iscapable of hijacking macrophages for replication, macrophagescould also limit the growth of the virus and eventually clear thevirus. Therefore, in immunocompetent individuals, SFTSV shouldbe limited and cleared from the host. In contrast, when the im-mune system is impaired, the virus would be expected to effi-ciently proliferate in the host and result in multiorgan dysfunctionand death. Along these lines, the most severe SFTS patients arethe elderly (1), and some lethal SFTS cases have reported ahistory of early application of dexamethasone, which acts to re-press immune functions (29). In fact, in our study, weight lossand death were not observed in immunocompetent adult mice,but occurred in immunocompromised mice. Additionally, ourfinding that low-dose SFTSV infection results in reduced clinicalseverity compared with high dose of infection suggests that thevirus amount is critical to disease severity. Combining of all theseobservations, it appears that an immunocompromised host thatis unable to limit the replication of SFTSV and has high virusburden would be inclined to develop a severe disease status.Because SFTSV can directly infect macrophages and is harboredwithin splenic macrophages for long periods, the role of macro-phages in limiting virus replication should be further investigatedto clarify the potential pathogenic mechanisms of SFTSV and tounderstand how the host immune system limits virus replication.The present study is unique in providing a reliable mouse

model to investigate the pathogenesis of SFTSV infection and

has demonstrated that pathological changes occur in threeSFTSV-target organs, including the spleen, liver, and kidney, atearly and late stages of the virus infection. In particular, thisstudy has revealed splenic macrophages to be target cells forinfection. Additionally, we have shown that macrophages likelyare involved in the mechanism leading to thrombocytopenia, themajor clinical hallmark symptom of SFTSV infection in humans.These findings relating to pathogenesis of SFTSV will help usbetter understand this new viral disease and open up the field forseveral lines of future studies.

Materials and MethodsSFTSV strain HB29was isolated from a SFTS patient by plaque purification andobtained after five passages in Vero E6 cells. The SFTSV infectious animalexperiments were conducted under biosafety level 3 (BSL3) containmentin accordance with institutional guidelines. C57/BL6 mice (n = 10) were i.m.inoculated with 105 TCID50 SFTSV, and five mock mice were used in parallelas controls. At each time point, the rectal temperature and weight wereacquired, then animals were exsanguinated and tissues were collectedimmediately. Taqman one-step real-time RT-PCR reactions were performedto test viral RNA copies in blood and tissues as previously described (6). In-fectious titers in blood and tissues were determined as previously reported(1, 20). Luminex assay was used to quantify SFTSV-specific serum IgG andIgM antibodies. Neutralizing antibodies were tested by micro-serum neu-tralization test as previously described (1). Serum AST and ALT were detec-ted by ELISA kits (USCN Life Science Inc.), and BUN was detected by a Ureaassay kit (Abcam). Spenic cellular responses were tested by ELISPOT IFN-γ setanalysis (BD). Pathological lesions were examined by H&E stained 4-μM thickparaffin-embedded tissue sections. Further, macrophage infection, plateletadherence, and macrophage phagocytosis of platelet assays were performedas previously described (11, 13). Detailed description of above experiments,as well as the score system for pathological lesions and experimental pro-cedures for immunohistochemistry and immunofluorescence are provided inSI Materials and Methods.

ACKNOWLEDGMENTS. This work was supported by China Mega-Project forInfectious Diseases Grant 2011ZX10004-001 from the Ministry of Scienceand Technology and Ministry of Health; National Key Program on BasicResearch Project (973 Program) Grant 2011CB504700 from the Ministry ofScience and Technology; and National Natural Science Foundation of ChinaGrant 81101301.

1. Yu XJ, et al. (2011) Fever with thrombocytopenia associated with a novel bunyavirusin China. N Engl J Med 364:1523–1532.

2. Pifat DY, Smith JF (1987) Punta Toro virus infection of C57BL/6J mice: A model forphlebovirus-induced disease. Microb Pathog 3:409–422.

3. Cusi MG, et al. (2005) Development of a mouse model for the study of Toscana viruspathogenesis. Virology 333:66–73.

4. Fisher AF, et al. (2003) Induction of severe disease in hamsters by two sandfly fevergroup viruses, Punta toro and Gabek Forest (Phlebovirus, Bunyaviridae), similar tothat caused by Rift Valley fever virus. Am J Trop Med Hyg 69:269–276.

5. Molyneux G, et al. (2005) The haemotoxicity of mitomycin in a repeat dose study inthe female CD-1 mouse. Int J Exp Pathol 86:415–430.

6. Sun Y, et al. (2012) Early diagnosis of novel SFTS bunyavirus infection by quantitativereal-time RT-PCR assay. J Clin Virol 53:48–53.

7. Dasso J, Lee J, Bach H, Mage RG (2002) A comparison of ELISA and flow microsphere-based assays for quantification of immunoglobulins. J Immunol Methods 263:23–33.

8. Elmore SA (2006) Enhanced histopathology of the spleen. Toxicol Pathol 34:648–655.9. Kuter DJ (1996) The physiology of platelet production. Stem Cells 14(Suppl 1):88–101.10. Morris L, Graham CF, Gordon S (1991) Macrophages in haemopoietic and other tissues

of the developing mouse detected by the monoclonal antibody F4/80. Development112:517–526.

11. Hoffmeister KM, et al. (2003) The clearance mechanism of chilled blood platelets. Cell112:87–97.

12. Badlou BA, Wu YP, Smid WM, Akkerman JW (2006) Platelet binding and phagocytosisby macrophages. Transfusion 46:1432–1443.

13. Gavrilovskaya IN, Gorbunova EE, Mackow ER (2010) Pathogenic hantaviruses directthe adherence of quiescent platelets to infected endothelial cells. J Virol 84:4832–4839.

14. Park EK, et al. (2007) Optimized THP-1 differentiation is required for the detection ofresponses to weak stimuli. Inflamm Res 56:45–50.

15. Morrill JC, et al. (2010) Rapid accumulation of virulent rift valley Fever virus in micefrom an attenuated virus carrying a single nucleotide substitution in the mRNA. PLoSONE 5:e9986.

16. Ritter M, et al. (2000) Resistance to Rift Valley fever virus in Rattus norvegicus: Genetic

variability within certain ‘inbred’ strains. J Gen Virol 81:2683–2688.17. do Valle TZ, et al. (2010) A new mouse model reveals a critical role for host innate

immunity in resistance to Rift Valley fever. J Immunol 185:6146–6156.18. Ashley SL, et al. (2011) Host genetic variation in susceptibility to Punta Toro virus.

Virus Res 157:71–75.19. Botten J, et al. (2000) Experimental infection model for Sin Nombre hantavirus in the

deer mouse (Peromyscus maniculatus). Proc Natl Acad Sci USA 97:10578–10583.20. Milazzo ML, Eyzaguirre EJ, Molina CP, Fulhorst CF (2002) Maporal viral infection in

the Syrian golden hamster: A model of hantavirus pulmonary syndrome. J Infect Dis

186:1390–1395.21. Lee M, et al. (1989) Coagulopathy in hemorrhagic fever with renal syndrome (Korean

hemorrhagic fever). Rev Infect Dis 11(Suppl 4):S877–S883.22. Guang MY, Liu GZ, Cosgriff TM (1989) Hemorrhage in hemorrhagic fever with renal

syndrome in China. Rev Infect Dis 11(Suppl 4):S884–S890.23. Lee M, et al. (1986) Procoagulant activity and thrombelastography in Korean hem-

orrhagic fever. J Korean Med Sci 1:53–58.24. Cosgriff TM, et al. (1991) Platelet dysfunction contributes to the haemostatic defect in

haemorrhagic fever with renal syndrome. Trans R Soc Trop Med Hyg 85:660–663.25. Swanepoel R, et al. (1989) The clinical pathology of Crimean-Congo hemorrhagic

fever. Rev Infect Dis 11(Suppl 4):S794–S800.26. Diegelmann RF, Evans MC (2004) Wound healing: An overview of acute, fibrotic and

delayed healing. Front Biosci 9:283–289.27. Nagai T, et al. (1985) Isolation of haemorrhagic fever with renal syndrome virus from

leukocytes of rats and virus replication in cultures of rat and human macrophages.

J Gen Virol 66:1271–1278.28. Lewis RM, Morrill JC, Jahrling PB, Cosgriff TM (1989) Replication of hemorrhagic fever

viruses in monocytic cells. Rev Infect Dis 11(Suppl 4):S736–S742.29. Gai Z, et al. (2012) Person-to-person transmission of severe fever with thrombocyto-

penia syndrome bunyavirus through blood contact. Clin Infect Dis 54:249–252.

10058 | www.pnas.org/cgi/doi/10.1073/pnas.1120246109 Jin et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 8

, 202

1