An Animal Model That Mimics Human Herpesvirus 6B Pathogenesis

Human herpesvirus 6B (HHV-6B) is a ubiquitous virus that establishes lifelong latent infection only in humans, and the infection can reactivate, with severe complications that cause major problems. A small-animal model of HHV-6B infection has thus been desired for research regarding the pathogenicity of HHV-6B and the development of antiviral agents. We generated humanized mice by transplantation with human hematopoietic stem cells, and here, we modified the model by providing an additional transfer of human mononuclear cells, providing the proper conditions for efficient HHV-6B infection. This is the first humanized mouse model to mimic HHV-6B pathogenesis, and it has great potential for research into the in vivo pathogenesis of HHV-6B.

ABSTRACT

Human herpesvirus 6B (HHV-6B), a T-lymphotropic virus, infects almost exclusively humans. An animal model of HHV-6B has not been available. Here, we report the first animal model to mimic HHV-6B pathogenesis; the model is based on humanized mice in which human immune cells were engrafted and maintained. For HHV-6B replication, adequate human T-cell activation (which becomes susceptible to HHV-6B) is necessary in this murine model. Here, we found that an additional transfer of human mononuclear cells to humanized mice resulted in an explosive proliferation of human activated T cells, which could be representative of graft-versus-host disease (GVHD) because the primary transfer of human cells was not sufficient to increase the number and ratio of human T cells. Mice infected with HHV-6B became weak and/or died approximately 7 to 14 days later. Quantitative PCR analysis revealed that the spleen and lungs were the major sites of HHV-6B replication in this model, and this was corroborated by the detection of viral proteins in these organs. Histological analysis also revealed the presence of megakaryocytes, indicating HHV-6B infection. Multiplex analysis of cytokines/chemokines in sera from the infected mice showed secretions of human cytokines/chemokines as reported for both in vitro infection and clinical samples, indicating that the secreted cytokines could affect pathogenesis. This is the first animal model showing HHV-6B pathogenesis, and it will be useful for elucidating the pathogenicity of HHV-6B, which is related to GVHD and idiopathic pneumonia syndrome.

IMPORTANCE Human herpesvirus 6B (HHV-6B) is a ubiquitous virus that establishes lifelong latent infection only in humans, and the infection can reactivate, with severe complications that cause major problems. A small-animal model of HHV-6B infection has thus been desired for research regarding the pathogenicity of HHV-6B and the development of antiviral agents. We generated humanized mice by transplantation with human hematopoietic stem cells, and here, we modified the model by providing an additional transfer of human mononuclear cells, providing the proper conditions for efficient HHV-6B infection. This is the first humanized mouse model to mimic HHV-6B pathogenesis, and it has great potential for research into the in vivo pathogenesis of HHV-6B.

INTRODUCTION

Human herpesvirus 6B (HHV-6B) is a T-lymphotropic betaherpesvirus (1–3) whose cellular receptor is human CD134 (also called OX40), which is expressed in activated human T cells (4). Primary infection by HHV-6B usually causes a common childhood febrile illness known as exanthem subitum (5). Febrile seizures and acute encephalitis develop in some cases, and approximately 50% of patients experience severe neurological sequelae (6). As is also the case for other members of the herpesvirus family, HHV-6B establishes a lifelong latent infection (7). Its reactivation has been reported in 40% to 70% of patients after allogeneic hematopoietic stem cell transplantation (allo-HSCT), with severe complications such as delayed engraftment, graft-versus-host disease (GVHD), and encephalitis (7–11). However, the detailed pathogenicity of HHV-6B in disease remains unknown due to the lack of an appropriate animal model (12).

Although a variety of animal models have been explored for HHV-6B infection, only limited success has been reported. In early research, viral replication was detected in the peripheral blood mononuclear cells (PBMCs) and spleens of Old World primates, including African green monkeys and cynomolgus macaques (13). Another nonhuman primate (NHP), the common marmoset, was used in recent studies; the relationship between HHV-6B infection and the acceleration of the development of multiple-sclerosis-like neuroinflammatory disease has been shown (14, 15). Specialized facilities and high costs are involved in NHP research, and we speculated that a small-animal model of HHV-6B infection would provide a solution for further investigations of HHV-6B pathogenesis.

Humanized mouse models, which have been generated by human CD34⁺ hematopoietic stem and progenitor cell (hHSPC) transplantation into lines of immunodeficient mice, reproduce human immunity and have been widely used in research regarding human-specific pathogens such as human immunodeficiency virus type 1 (HIV-1), human T-cell leukemia virus type 1 (HTLV-1), and Epstein-Barr virus (EBV) (16–18). Humanized mice are highly susceptible to these viruses, as infection occurs in human immune cells, and the mice show similar disorders and a human-like immune response (16–18). Infection of humanized mice with HHV-6A was also reported, but this model does not show pathogenicity (19). Although HHV-6A is genetically closely related to HHV-6B, they are classified into different species because of distinctive pathogenicities, cellular receptors, and cell tropisms. The usability of humanized mice in studies of HHV-6B infection has not been described so far.

In the present study, we established a modified humanized mouse model in which human T cells are abundantly activated with the expression of hCD134, which is a cellular receptor of HHV-6B; this expression seems to be a favorable condition for HHV-6B infection (4). With a further transfer of stimulated human cord blood mononuclear cells (CBMCs) into humanized mice after engraftment with hHSPCs, remarkable increases in the number and activation of human T cells were observed. Under these conditions, which are similar to those of patients with severe acute GVHD after HSCT, HHV-6B replication was observed in almost all main organs of the mice and at especially high levels in the spleen and lungs. Some of the mice died after infection, and differing pathological features were observed in histological analyses between infected and uninfected mice. The induction of cytokines and chemokines in the mice after HHV-6B infection was observed, as reported for clinical samples (20), indicating that they could affect mouse pathogenesis.

RESULTS

Establishment of a modified humanized mouse permissive for abundant HHV-6B proliferation.

We attempted to establish a humanized mouse with abundant human T cells for HHV-6B infection. We first used a standard protocol to prepare humanized mice. Immunodeficient 129S4-Rag2^tm1.1Flv Il2rg^tm1.1Flv Tg(SIRPA)1Flv/J (hSIRPα double-knockout [hSIRPα-DKO]) mice at 4 to 6 weeks of age were irradiated and engrafted with CD34⁺ hHSPCs isolated from individual cord blood samples to generate humanized mice (hSIRPα-DKO-hu). Human cell engraftment was evaluated by flow cytometry of peripheral blood of the mice 9 to 12 weeks after the transplantation of hHSPCs, with the detection of the panleukocyte markers human CD45 (hCD45) and mouse CD45 (mCD45) (Fig. 1A and B). Because HHV-6B preferentially replicates in activated hCD4⁺ T lymphocytes, which is consistent with the fact that its entry receptor hCD134, a member of the tumor necrosis factor (TNF) receptor superfamily, is expressed only in activated T lymphocytes (4), we further determined the detailed population of hCD45-positive cells. As described previously (21), fluorescence-activated cell sorter (FACS) analysis of the T-cell markers CD3, CD4, and CD8 along with the B-cell marker CD19 revealed that the major population of human cells in our hSIRPα-DKO-hu mice was B cells at 9 to 12 weeks posttransplantation (Fig. 1C).

FIG 1.

CBMC transfer induces the proliferation of human T cells. The changes in human cells in peripheral blood after the transfer of CBMCs in eight hSIRPα-DKO-hu mice were evaluated by FACS analysis for the expressions of human CD45, CD19, CD3, CD4, CD8, and CD134 markers along with mouse CD45. (A and B) After gating based on the size of the mononuclear cells, an increased ratio (A) and an increased number (B) of hCD45-positive cells were observed over the time course. (C and D) Ratios of human CD3 and CD19 (C) and CD4 and CD8 (D) presented as mean percentages of hCD45- and CD3-positive cells, respectively (n = 8). (E) Mean expression ratio ± standard deviation (SD) of CD134 in hCD45-positive cells after CBMC transfer (n = 5). (F) Flowchart summarizing the time scale of this research.

Since the small population of T cells might greatly affect the efficiency of HHV-6B infection, we conducted several trials to induce larger populations of T cells. We eventually achieved a larger population by administering an intraperitoneal injection of 4 × 10⁶ to 6 × 10⁶ stimulated CBMCs to hSIRPα-DKO-hu mice 9 to 12 weeks after transplantation of CD34⁺ hHSPCs. Two weeks after the CBMC injection, we observed increases in hCD45⁺ cells ranging from 58% to 98%, and the cell population consisted mainly of human CD3⁺ T cells (2nd week, mean of 96% and range of 91% to 98%) (Fig. 1A to C). Both CD4 and CD8 subsets were detected within CD3⁺ T cells (Fig. 1D). Although the rate varied among the mice, an increase in the CD8 subset was detected by calculating the CD4/CD8 ratio (1st week, mean of 3.3 and range of 1.1 to 8.5; 2nd week, mean of 0.8 and range of 0.4 to 1.8) (Fig. 1D). The expression of hCD134 was detected in hCD45⁺ cells from the 1st week; the ratio decreased, but the number of hCD134⁺ cells in peripheral blood was increased the next week as the increase in the number of total human cells (ratio, mean of 43.2% and range of 30% to 54% at week 1 and mean of 20.7% and range of 17% to 28% at week 2; numbers, mean of 1.46 × 10⁷ cells/ml at week 1 and mean of 9.87 × 10⁷ cells/ml at week 2) (Fig. 1E). A brief flowchart summarizing the time scale of this research is shown in Fig. 1F.

In summary, the additional transfer of interleukin-2 (IL-2)- and phytohemagglutinin (PHA)-stimulated CBMCs into the mice resulted in human T-cell proliferation. We also observed that the additional transfer of unstimulated CBMCs could induce T-cell proliferation in the mice, but the appearance of this proliferation was much delayed after the CBMC transfer compared to that of the stimulated CBMCs (data not shown).

We next investigated the origin of the proliferated T cells. We used the expression of human leukocyte antigen allele 2 (HLA-A2) as the marker to distinguish the cells from the transplantation of CD34⁺ hHSPCs and the later-transferred CBMCs. The flow cytometry results revealed that the proliferated T cells after CBMC transfer were derived from the transferred CBMCs (Table 1). Although the proliferation of CBMCs might be due to the proliferation of xenoreactive T cells derived from transferred CBMCs, it may also be due to the proliferation of alloreactive T cells from transferred CBMCs that respond to hematopoietic cells originating from the hCD34⁺ hHSPCs. We thus investigated whether the proliferation of transferred CBMCs was attributable to the reaction to alloantigen derived from an HLA-mismatched HSPC donor.

TABLE 1.

Origins of the induced T cells^a

Mouse	Age (mo)	Before transfer of CBMCs			Transfer of CBMCs		2 wks after transfer
		Donor	Genotypeb of HSPCs	% hCD45+ cells	Donor	Genotypeb of CBMCs	% hCD45+ cells	Genotypeb of WBCs in blood (%)
42	4	1	−	5.68	a	+	95.4	+ (>99)
51	4	1	−	13.26	b	+	98.59	+ (>99)
53	4	1	−	17.08	c	+	31.9	+ (>99)
49	4	2	−	18.59	b	+	65.39	+ (>99)

^aHSPCs, hematopoietic stem cells; CBMCs, cord blood mononuclear cells; WBCs, white blood cells.

^bThe genotypes were distinguished by the presence (+) or absence (−) of HLA-A2.

To this end, we transferred CBMCs isolated from either the CD34-negative fraction of the same HSPC donor or that from a different donor, and we compared the proliferation of hCD45⁺ cells after the transfer. The results demonstrated that both of these sets of CBMCs stimulated an increase of T lymphocytes; no significant difference was observed (P > 0.05 detected at both week 1 and week 2) (Fig. 2A and B). We also transferred the same number of CBMCs into 10 hSIRPα-DKO mice without humanization by hHSPC transplantation. However, an explosive increase in the number of human cells was detected in only two mice as early as 3 weeks after transfer (Fig. 2C). This indicated that the first transplantation of hCD34⁺ hHSPCs played an important role in the proliferation of the engrafted human T cells, although the cells were derived from the additionally transferred CBMCs.

FIG 2.

Increase in human T cells after CBMC transfer under different conditions. The changes in human T cells in peripheral blood were monitored by FACS analysis for the expression of human CD3 along with human CD45 and mouse CD45. (A) Ratios of CD3⁺ cells in total white blood cells from six hSIRPα-DKO-hu mice after the transfer of CBMCs that were from the same donor as that for the hHSPCs used for the first engraftment. (B) Results from eight hSIRPα-DKO-hu mice after the transfer of CBMCs that were from different donors of hHSPCs. (C) Ratios of CD3⁺ cells in total white blood cells from 10 hSIRPα-DKO mice after the transfer of CBMCs.

Along with the activation and proliferation of human T cells after the transfer, symptoms of GVHD-like activity were observed. In studies of mouse models injected with PBMCs that developed xenogeneic GVHD (X-GVHD), rapid and severe weight loss (>10%) was reported to be a common sign, along with other symptoms such as hunched posture, ruffled fur, reduced mobility, and anemia (22–27). We observed similar weight loss (>10%) as the most apparent sign in mice examined here (Fig. 3A). Among the recorded 18 mice, 8 died 17 to 52 days after the transfer (Fig. 3B).

FIG 3.

GVHD-like symptoms detected after CBMC transfer. (A) Changes in relative body weights of six CBMC-transferred hSIRPα-DKO-hu mice plotted as days after transfer. (B) Survival of 18 CBMC-transferred hSIRPα-DKO-hu mice shown as days after transfer. (C) Representative immunostaining of hCD3 in the indicated organs from one hSIRPα-DKO-hu mouse 52 days after the transfer of CBMCs. V, vessel; PV, portal vein; b, biliary duct. Bars, 100 μm.

Histological and immunohistochemical analyses were performed to assess the infiltration of hCD3⁺ T cells in organs. Among the spleen, lungs, liver, and kidneys, an especially dense infiltration of lymphocytes was observed in the spleen and lungs (Fig. 3C). The distribution pattern of T cells was diffuse in the splenic parenchyma, and a perivascular infiltrative pattern was seen in the lungs and kidneys (Fig. 3C). In the liver, T cells were predominantly distributed in the portal tract (Fig. 3C). These results suggest that hCD3⁺ T cells have a certain degree of organotrophy harboring a variable aggregation pattern, with a tendency to be associated with vessels. This was also consistent with the results for PBMC-injected mice that develop GVHD (23–25, 27).

Despite the symptoms and mortality caused by rejection and the risk of progression to GVHD, we suspected that efficient HHV-6B replication occurred in this model due to the abundance of activated T cells (Fig. 1A and E).

Infection of CBMC-transferred hSIRPα-DKO-hu mice with HHV-6B.

We therefore attempted to infect the humanized mice established here with HHV-6B. As the first trial, three CBMC-transferred hSIRPα-DKO-hu mice (which had enriched human T cells) were challenged with HHV-6B. These mice died at 7, 11, and 13 days postinfection (dpi). To further analyze the relationship between these deaths and HHV-6B infection, we first monitored the changes in human cells. Nine CBMC-transferred mice were inoculated with cell-free HHV-6B via intravenous injection 3 to 4 days later when the ratio of hCD45⁺ cells to total human and mouse CD45⁺ cells was >10% and the ratio of CD3⁺ T lymphocytes in hCD45⁺ cells was >90%.

We evaluated the changes of human cells in peripheral blood by conducting FACS analyses at 1 and 8 dpi for three mice and at 4 and 11 dpi for six mice. Decreases in both the ratio and the number of hCD45⁺ cells were detected from 4 dpi (which was earlier than for the control group of seven uninfected mice), and the decrease in the ratio of hCD45⁺ cells was significant at 8 dpi (Fig. 4A and B). In light of a report indicating that HHV-6B had a greater effect on the CD4⁺ subset in both in vitro and clinical research (12), we also checked the ratios of the hCD4⁺ subset in hCD3⁺ cells against the hCD8⁺ subset. The results revealed a decreased ratio of the CD4⁺ subset after infection, while no significance was detected compared with the uninfected group (Fig. 4C).

FIG 4.

Changes in cell populations and body weights after HHV-6B infection. After HHV-6B infection, nine CBMC-transferred hSIRPα-DKO-hu mice were divided into two groups. Peripheral blood was taken at 1 and 8 dpi from group 1 (n = 3) and at 4 and 11 dpi from group 2 (n = 6) and analyzed by FACS analysis. Seven CBMC-transferred hSIRPα-DKO-hu mice were used as the control group. (A and B) Mean ratios (A) and numbers (B) of hCD45⁺ cells per the time course ± SD. Student’s t tests were used to determine significance, which is shown above each data set (n.s., not significant). (C) Ratios of the percentages of human CD4⁺ cells to human CD8⁺ cells in CD3⁺ cells from the same mice used in panels A and B presented as means and SD. (D) Changes in relative body weights of five infected CBMC-transferred hSIRPα-DKO-hu mice. (E) Survival of 21 infected CBMC-transferred hSIRPα-DKO-hu mice shown as days after transfer. The infections were performed on the 10th to 14th days.

We also routinely measured the changes in the body weights of five hSIRPα-DKO-hu mice that were challenged by HHV-6B infection from the transfer of CBMCs until the end of infection (Fig. 4D). One mouse died on the 18th day (4 dpi), and two mice died on the 22nd day (8 dpi), with weight loss. One mouse exhibited anemia, showing a light-colored blood sample taken for the check of human cells, and it was thus euthanized on the 21st day (7 dpi). The fifth mouse was alive until the 28th day (14 dpi) and was euthanized for analysis of infection (Fig. 4D). The survival of all 21 mice used in this research is summarized in Fig. 4E. The time points of the deaths ranged from 3 to 14 dpi.

HHV-6B detection in organs.

For our analysis of HHV-6B infection in vivo, 14 CBMC-transferred hSIRPα-DKO-hu mice and 3 hSIRPα-DKO-hu mice were infected with HHV-6B and sacrificed at various time points (Table 2). Because efficient virus infection in the spleen was observed in a preliminary experiment, we measured the viral DNA amount in the spleen as a marker of infection in these mice. In all 14 mice that were sacrificed at different time points because of their distinct body conditions, the peak of viral replication seemed to be at around 7 to 10 dpi, and it decreased after 14 dpi (Fig. 5A). The decrease in genome DNA was consistent with the decrease in the number of human T cells (Fig. 4A and B) and was reasonable considering that the virus could no longer proliferate when the human cells were exhausted by the infection.

TABLE 2.

Past history and characteristics of mice used for HHV-6B detection and multiplex chemokine assays^a

Mouse	Mouse condition				Conditions of serum for cytokine/chemokine assay			Tissue samples for detection of virus			Outcome for mouse
	Age (mo)	Expt performed			No. of days after CBMC transfer	Day postinfection	% hCD45+ cells	Day postinfection	Quantitative PCR result	Histological analysis result
		HSPC	CBMC	HHV-6B
51	3	+	+	+	32	9	4.4	9	P	−	Survival
53	3	+	+	+	32	9	4.1	9	P	P	Debility
55	3	+	+	+	NA			9	P	−	Survival
68	3	+	+	+	27	14	39.9	NA			Survival
98	4	+	+	+	21	1	77.4	11	P	−	Death
					28	8	52.3
b3	4	+	+	+	21	1	55.2	NA			Survival
b6	4	+	+	+	21	1	29.1	NA			Survival
b29	3	+	+	+	14	3	45.6	10	P	P	Debility
b32	4	+	+	+	14	4	62.5	21	P	−	Survival
b36	4	+	+	+	14	4	45.8	18	P	−	Survival
					21	11	30.7
b39	4	+	+	+	14	3	41.2	13	P	P	Death
b64	4	+	+	+	21	3	38.8	4	P	−	Death
b66	4	+	+	+	NA			14	P	−	Debility
b72	4	+	+	+	NA			7	P	−	Death
b81	4	+	+	+	21	7	41.9	7	P	P	Debility
b95	4	+	+	+	31	8	17.1	NA			Survival
r28	5	+	+	+	17	2	99.0	8	P	−	Survival
r30	5	+	+	+	17	2	97.4	8	P	−	Survival
517	11	+	+	+	27	13	25.5	NA			Debility
44	4	+	+	−	32		10.1		−	N	Survival
49	3	+	+	−	21		25.3		−	−	Survival
74	4	+	+	−	52		97.5		−	N	Death
b45	4	+	+	−	28		22.4		−	−	Debility
b55	4	+	+	−	21		44.5		−	−	Debility
37	5	−	−	−					−	−	Survival
40	4	−	−	−					−	−	Survival
43	6	−	−	−					−	−	Survival
r16	6	+	−	+	NA			7	P	−	Survival
r19	6	+	−	+	NA			7	P	−	Survival
r52	5	+	−	+	NA			7	P	−	Survival

^aHSPC, hematopoietic stem cell; CBMC, cord blood mononuclear cell; NA, not available; +, experiments performed; −, experiments not performed; P, positive; N, negative.

FIG 5.

Viral genome copies in organs. (A) HHV-6B genome copies detected by quantitative PCR for the total DNAs extracted from tissues of spleen shown for the time course after infection (n = 14). (B) HHV-6B genome copies detected in the spleen, liver, brain, gut, kidney, heart, and lung after infection (n = 7) (mean, 8.2 dpi; median, 8 dpi; range, 7 to 11 dpi).

We next analyzed the viral DNA amount in each organ for 7 of the 14 mice that were sacrificed at the peak of viral replication (median, 8 dpi; range, 7 to 11 dpi). DNA samples were extracted from the spleen, liver, brain, gut, kidney, heart, and lung for viral genome detection by quantitative PCR. The highest copy numbers were detected in the spleen (mean of 17,158 and range of 1,064 to 43,481 copies per ng of total DNA) and lungs (mean of 9,367 and range of 305 to 34,506 copies per ng of total DNA), and the viral genome was also detected in all of the other organs (except one sample of the gut) from the infected group (Fig. 5B). In all of the infected hSIRPα-DKO-hu mice without an additional CBMC transfer, less of the viral genome was detected (<1 copy per ng of total DNA) (Table 3).

TABLE 3.

Viral genome copies in infected hSIRPα-DKO-hu mice without transfer of CBMCs

Mouse	% hCD45+ cells before infection	No. of copies of the HHV-6B genome/ng DNAa
		Spleen	Liver	Brain	Gut	Kidney	Heart	Lung
r16	5.7	0.01	UD	UD	UD	UD	0.01	UD
r19	4.3	UD	0.02	UD	UD	UD	UD	0.07
r52	16.3	0.45	0.02	UD	0.05	UD	0.03	1.00

^aUD, undetectable.

To further evaluate the impact of HHV-6B infection on various organs, we performed histological and immunohistochemical analyses. In representative photographs of two infected mice that showed debility and were euthanized for viral detection at 7 dpi (<84% of body weight and reduced mobility) (Fig. 6), marked viral infection, as shown by a cytomegaly-like appearance, was observed in the spleen and lung (Fig. 6A, yellow arrows), which was consistent with the high viral copy numbers in these organs. Infiltrations of hCD3⁺ T cells were detected in the spleen and lung and also in the liver and kidney, but less so (Fig. 6B), as in the additional CBMC-transferred noninfected mice (Fig. 3C). Infected cells were detected by rabbit antiserum against immediate early protein 1 (IE1) (Fig. 6C). Infiltrations of hCD3⁺ T cells and viral antigen were scarcely detected in other organs, including the intestinal tract, pancreas, brain, and myocardium (data not shown), but the significance of these minor histological findings remains undetermined.

FIG 6.

Histological analyses and immunostaining of infected mice. (A) Representative histological findings for the spleen of an infected mouse. (Left) Low-power view of the tissue showing infiltrations of atypical large cells. (Right) High-power view indicating infected giant cells (megakaryocytes) (arrows). Bars, 50 μm. (B and C) Immunostaining of hCD3 (B) and HHV-6B IE1 (C). V, vessel; PV, portal vein; b, biliary duct. Bars, 100 μm.

Measurement of human cytokines and chemokines after infection.

Several specific cytokines and chemokines induced by primary HHV-6B infection have been reported in clinical research (20, 28). To investigate the cytokines/chemokines induced or reduced in our animal model, we analyzed 40 cytokines/chemokines using the Bio-Plex multiplex immunoassay system (Fig. 7 and Table 4). The sera used in this assay are listed in Table 2, including those from infected CBMC-transferred hSIRPα-DKO-hu mice (n = 16), uninfected CBMC-transferred hSIRPα-DKO-hu mice (n = 5), and uninfected hSIRPα-DKO mice (n = 3). All of the measured data are summarized in Fig. 7.

FIG 7.

Heat map of data from cytokine/chemokine assays. The heat map indicates the results for the 40 cytokines/chemokines detected in serum from infected CBMC-transferred hSIRPα-DKO-hu mice (18 samples) versus uninfected CBMC-transferred hSIRPα-DKO-hu mice (5 samples) and uninfected hSIRPα-DKO mice (3 samples). The line of numbers above the graph shows the days postinfection. Groups were defined as follows: group 1, cytokines/chemokines detected only in infected mice; group 2, cytokines/chemokines detected at significantly higher levels in infected mice than in uninfected mice (P < 0.05); group 3, chemokines detected at significantly lower levels in infected mice than in uninfected mice (P < 0.05); group 4, cytokines/chemokines detected at the same levels in infected mice and uninfected mice; group 5, cytokines/chemokines not detected.

TABLE 4.

The 40 cytokines/chemokines analyzed in this study

Cytokine/chemokinea
6Ckine/CCL21
BCA-1/CXCL13
CTACK/CCL27
ENA-78/CXCL5
Eotaxin/CCL11
Eotaxin-2/CCL24
Eotaxin-3/CCL26
Fractalkine/CX3CL1
GCP-2/CXCL6
GM-CSF
Gro-α/CXCL1
Gro-β/CXCL2
I-309/CCL1
IFN-γ
IL-1β
IL-2
IL-4
IL-6
IL-8/CXCL8
IL-10
IL-16
IP-10/CXCL10
I-TAC/CXCL11
MCP-1/CCL2
MCP-2/CCL8
MCP-3/CCL7
MCP-4/CCL13
MDC/CCL22
MIF
MIG/CXCL9
MIP-1α/CCL3
MIP-1δ/CCL15
MIP-3α/CCL20
MIP-3β/CCL19
MPIF-1/CCL23
SCYB16/CXCL16
SDF-1α+β/CXCL12
TARC/CCL17
TECK/CCL25
TNF-α

^aBCA-1, B-cell-attracting chemokine 1; CCL21, chemokine (C-C motif) ligand 21; CTACK, cutaneous T-cell-attracting chemokine; CXCL13, chemokine (C-X-C motif) ligand 13; CX3CL1, chemokine (C-X3-C motif) ligand 1; ENA-78, epithelial cell-derived neutrophil-activating peptide 78; GCP-2, granulocyte chemotactic protein 2; GM-CSF, granulocyte-macrophage colony-stimulating factor; Gro-α, growth-related oncogene α; IFN-γ, gamma interferon; IL-1β, interleukin-1β; IP-10, gamma interferon-induced protein 10; I-TAC, interferon-inducible T-cell alpha chemoattractant; MCP-1, monocyte chemotactic and activating factor 1; MDC, macrophage-derived chemokine; MIF, macrophage migration inhibitory factor; MIG, monokine induced by gamma interferon; MIP-1α, macrophage inflammatory protein 1α; MPIF-1, myeloid progenitor inhibitory factor 1; SCYB16, small inducible cytokine B 16; SDF-1α+β, stromal cell-derived factor 1α and -β; TARC, thymus- and activation-regulated chemokine; TECK, thymus-expressed chemokine; TNF-α, tumor necrosis factor alpha.

Among the 40 cytokines and chemokines assessed, 2 cytokines, i.e., IL-6 and IL-10, and 5 chemokines, chemokine (C-C motif) ligand 24 (CCL24) (or eotaxin-2), CCL15 (or macrophage inflammatory protein 1δ [MIP-1δ]), chemokine (C-X-C motif) ligand 8 (CXCL8) (or IL-8), CXCL11 (or interferon-inducible T-cell alpha chemoattractant [I-TAC]), and CXCL16 (or small inducible cytokine B 16 [SCYB16]), were detected only in the infected group (group 1). Comparison of the levels between the infected and uninfected groups revealed that the production of six chemokines, including CCL2 (or monocyte chemotactic and activating factor 1 [MCP-1]), CCL8 (or MCP-2), CCL17 (or thymus- and activation-regulated chemokine [TARC]), CCL19 (or MIP-3β), CCL22 (or macrophage-derived chemokine [MDC]), and CXCL9 (or monokine induced by gamma interferon [MIG]), were significantly induced in the infected group (group 2), while only one chemokine, CXCL13 (or B-cell-attracting chemokine 1 [BCA-1]) was detected as being significantly downregulated after infection (group 3).

Among the remaining 26 cytokines and chemokines, 20 were detected as being present at the same level between the infected and uninfected groups (group 4), whereas the other 6 were not detected in any group (group 5). Among these six cytokines and chemokines, chemokine (C-X3-C motif) ligand 1 (CX3CL1) (or fractalkine) was detected in all of the samples (even in the hSIRPα-DKO mice) at similar values near the lower range of the detection limit, and the 10-fold-diluted samples showed results similar to those for their origins. Thus, we decided to move this chemokine to the undetected group.

DISCUSSION

Our experiments revealed that additional CBMC-transferred hSIRPα-DKO-hu mice were susceptible to infection with HHV-6B, which eventually showed T-cell predominance. This is the first model to show HHV-6B pathogenesis described as follows. The mice infected with HHV-6B became weaker day by day and sometimes died at approximately 7 to 14 days postinfection. High copy numbers of the viral genome (∼10,000 copies/ng total DNA at 8 dpi) were detected in the mouse spleen and lung, and lower levels of the viral genome were also detected in the liver, gut, kidney, heart, and brain (Fig. 5B). The viral protein was also detected in the spleen, liver, kidney, and lungs (Fig. 6C). These results were consistent with the immunostaining results and with the results of a study in which infiltrations of human T cells were detected in these organs after the transfer of human cells (27).

Because the spleen is one of the sites of the proliferation and maturation of immune cells, our detection of a high level of HHV-6B genome DNA reflected the abundance of human T cells in this organ (Fig. 6B). Our observation of a high level of HHV-6B viral genome DNA in the lungs of the mice that was comparable to that in the spleen is an unpredicted result. The proliferation of HHV-6B in the lungs demonstrated organ tropism in vivo, indicating that this might be evidence of a relationship between HHV-6B and idiopathic pneumonia syndrome (IPS), a disease that was recently reported to be caused by an infectious agent after allo-HSCT (29, 30). The most frequent pathogen detected from the bronchoalveolar lavage fluid was HHV-6 (29% of all samples) (30), and the relationship between HHV-6B and lower respiratory tract disease, including IPS, was further discussed in another study (31). All these results indicated the pathogenic role of HHV-6B. Thus, our mouse model could be helpful to further investigate the role of HHV-6B in IPS.

Several research groups have reported that the intravenous or intraperitoneal transfer of human T lymphocytes (such as PBMCs) into SCID mice led to the consistent development of X-GVHD (22–24, 26). In these studies, relatively high numbers of cells, i.e., 3 × 10⁷ to 5 × 10⁷ (up to 10⁸) PBMCs, were commonly used, and relatively high mortality rates (>80%) were observed (22–24, 26, 27). In our present investigation using hSIRPα-DKO-hu mice, we used a lower number of cells (2 × 10⁶ to 5 × 10⁶ CBMCs) for each mouse, and high human T-cell chimerism (>90%; up to 99%) was achieved in >90% of the mice, with relatively high survivability (Fig. 1C and Fig. 3B). In a previous experiment with DKO mice, an additional depletion of mouse macrophages promoted the proliferation of engrafted cells with lower doses of PBMCs (27).

The relatively low cell numbers required for human T-cell engraftment in the present study may be related to the expression of hSIRPα on host macrophages, which inhibits the phagocytosis of transferred human cells by the interaction of hSIRPα on macrophages with hCD47 on human T cells (21). This may be similar to the effect of the depletion of macrophages in previous studies (21, 27). We also observed that the first humanization of the mice with hHSPCs played an important role in the latter explosive induction of T cells by CBMC transfer (Fig. 2). Although the detailed mechanisms remain to be elucidated, the proliferated and activated T cells expressing hCD134 and the entry receptor of HHV-6B provided suitable conditions for HHV-6B infection and replication.

In our murine infection model, human cell transfer occurred before HHV-6B infection, and this might imitate a clinical HHV-6B infection by the reactivation of latent virus in allo-HSCT patients (7–10, 32–35). A relationship between GVHD and HHV-6B reactivation has been proposed (32, 33), although the cause-effect relationship and underlying mechanism remain unclear. As we note above, the additional transfer of CBMCs caused T-cell proliferation and activation similar to those of the previously reported X-GVHD murine model, and thereby, the mice became much more susceptible to HHV-6B. Thus, HHV-6B infection in the CBMC-transferred hSIRPα-DKO-hu mice in our present study may support the idea that GVHD promotes HHV-6B reactivation/replication, as suggested in our previous study (35).

As the induction of hCD134 is thought to be key for HHV-6B reactivation/replication, this induction may be involved in the mechanism of the pathogenesis of HHV-6B (4, 35). It was reported previously that GVHD was associated with the induction of CD134 in a rat bone marrow transplantation model (36, 37) and in an HSCT recipient who developed acute and chronic GVHD (38–40). Pritchett et al. described a relationship between the expression of CD134 after umbilical cord blood transplantation and the level of HHV-6B (41). Our clinical study revealed a clear association between the expression of CD134 and HHV-6B reactivation/replication after allo-HSCT, indicating their tight correlation (35). The HHV-6B infection observed here in the mice (which manifested as a GVHD-like cell population change and the expression of CD134) implies a causative role of GVHD in HHV-6B reactivation/replication; there are several reports that HHV-6B reactivation may trigger GVHD (32, 34, 42). Thus, GVHD and HHV-6B reactivation/replication have been found to be recursively correlated with each other. The important point about HHV-6B replication is that it requires hCD134 expression. This issue should be addressed by establishing an HHV-6B latency infection model in future studies.

As mentioned above, we observed rapid weight loss in almost all of the infected mice, and some of them died. However, we have not achieved significant differences in weight changes and survival rates between infected and uninfected mice so far (Fig. 3A and B and Fig. 4D and E). In these infected mice, the death of the HHV-6B-infected T cells resulted in a relatively faster disappearance of the human cells, which targeted the bodies of the mice (Fig. 4A and B). This may have the effect of relieving the progression of GVHD, although further investigation should be required to mention it. Even under these conditions, weight loss and death still occurred in the infected mice, indicating that the infection resulted in a somehow different cause of death in these mice (Fig. 3 and Fig. 4D and E). The mutual relationship between HHV-6B and GVHD or other factors and their positive and negative effects are thought to result in a complicated set of symptoms in our mouse model, and this is also true for clinical patients with HHV-6B infection.

We analyzed 40 cytokines and chemokines in serum from both infected and uninfected hSIRPα-DKO-hu mice transferred with CBMCs (Fig. 7). The secretion of 13 cytokines/chemokines specific for our infected mice was found (Fig. 7). The relatively high levels of secretion of cytokines/chemokines related to inflammatory host responses, such as IL-6, IL-8, CCL2, and CCL8, in the early phase of infection (<4 dpi) might be due to the immune response against HHV-6B, whereas the reduction of these cytokines/chemokines might be related to the reduction in the number of human cells (Fig. 4A and B). Interestingly, CCL24, which is known to be strongly chemotactic for resting T lymphocytes, was detected at almost the same level during infection. The only downregulated chemokine, CXCL13, was reported to be an indicator of acute gastrointestinal GVHD in a murine model (43). The smaller amount of this chemokine might support the hypothesis that a relief of GVHD symptoms might occur after HHV-6B infection, and thus, the reasons for the death of the mice in the present study could be different between the infected and uninfected mice. Regarding GVHD, the secretions of inflammatory cytokines and chemokines indicated the subtypes of human T cells involved in the development of disease (26, 44–47), although this could be altered by HHV-6B infection.

Previously, we reported that CXCL11, CXCL16, CCL2, and CCL8 were upregulated in serum from febrile children with primary HHV-6B infection (20), while CCL2 was also reported by another group (28). Here, we observed that similar ones were upregulated by HHV-6B infection compared to uninfected groups. In addition, IL-8, IL-10, and CCL2 have also been detected in HHV-6B infection in vitro (48, 49). We thus conclude that the cytokines and chemokines secreted in the mice infected with HHV-6B could have affected host pathogenesis.

In conclusion, we developed a small-animal model that shows HHV-6B infection and pathogenesis. By transferring additional CBMCs to hSIRPα-DKO-hu mice, high ratios of activated human T cells required for HHV-6B infection were achieved in these mice. Efficient infection was detected, with high copy numbers of the viral genome and the expression of viral proteins. Changes of cytokines and chemokines were detected in this mouse model that were similar to those in primary HHV-6B infections of humans and in vitro HHV-6B infections. We expect the further use of this model for the elucidation of virus pathogeneses and the development of HHV-6B-specific antivirals.

MATERIALS AND METHODS

Antibodies.

A violetFluor 450-conjugated monoclonal antibody (mAb) against human CD45 (clone HI30), a phycoerythrin (PE)-conjugated mAb against human CD4 (clone OKT4), a PE-cyanine 7 (Cy7)-conjugated mAb against human CD3ε (clone UCHT1), an allophycocyanin (APC)-conjugated mAb against human CD8 alpha (clone SK1), and an APC-conjugated mAb against human CD19 (clone SJ25C1) used for flow cytometry were purchased from Tonbo Biosciences (San Diego, CA). A peridinin chlorophyll protein (PerCP)-cyanine 5.5 (Cy5.5)-conjugated mAb against mouse CD45 (clone 30-F11), a PE-Cy7-conjugated mAb against human CD34 (clone 581), and a fluorescein isothiocyanate (FITC)-conjugated mAb against human leukocyte antigen allele A2 (HLA-A2) (clone BB7.2) were purchased from BioLegend (San Diego, CA). A monoclonal antibody against human CD3ε (clone F7.2.38) was purchased from Dako (Glostrup, Denmark) and used for immunostaining. The mouse mAb against human CD134 was introduced in previous work (4) and was conjugated with FITC using the SureLink FITC labeling kit (Funakoshi, Tokyo, Japan) according to standard protocols. Rabbit antisera against HHV-6B IE1 were described in previous work and were used for immunohistological analyses (4).

Cells and viruses.

HHV-6B strain HST was prepared as described previously (50). For injection and virus propagation, we obtained umbilical cord blood mononuclear cells (CBMCs) from RIKEN (The Institute of Physical and Chemical Research, Ibaraki, Japan) and the Hyogo Cord Blood Bank (Hyogo, Japan). Human CD34⁺ cells were isolated from fresh cord blood cells by density gradient centrifugation (Lymphoprep; Alere Technologies, Norway), followed by positive immunomagnetic selection with anti-human CD34 microbeads (Miltenyi Biotec, Bergisch-Gladbach, Germany). After isolation, the expression of hCD3, hCD19, hCD34, and HLA-A2 was measured by flow cytometry. Cells with high CD34⁺ purity (≥95%) and with low T- and B-cell contamination (<1% and <5%, respectively) were used for the study. Isolated CD34⁺ cells were frozen at −80°C in fetal bovine serum (FBS) containing 10% dimethyl sulfoxide (DMSO) until use. The CD34-negative fraction was also frozen at −80°C and used for the adaptive-transfer experiment. Studies using human samples were approved by the ethical committee of the Kobe University Graduate School of Medicine. The human T-cell line MT4 was cultured at 37°C with 5% CO₂ in RPMI 1640 medium with 8% fetal bovine serum containing 20 μg/ml gentamicin and used for virus titration as described previously (51).

Generation of humanized mice.

Immunodeficient 129S4-Rag2^tm1.1Flv Il2rg^tm1.1Flv Tg(SIRPA)1Flv/J (hSIRPα-DKO) mice and 129S4-Rag2^tm1.1Flv Il2rg^tm1.1Flv/J (DKO) mice were obtained from Jackson Laboratory (Bar Harbor, ME, USA). The mice were maintained under specific-pathogen-free conditions at the Institute for Experimental Animals at the Kobe University Graduate School of Medicine. All animal experiments were approved by the Institutional Animal Care and Use Committee and were handled in accordance with Kobe University animal experimentation regulations (permit no. 1209 and 160162).

To generate the hSIRPα-DKO-hCD34 mice, isolated hCD34⁺ hHSPCs were thawed in Iscove’s modified Dulbecco’s medium (IMDM) containing 10% FBS and 50 μg/ml DNase I (Sigma). Cells were washed twice with phosphate-buffered saline (PBS). A total of 50,000 to 100,000 hCD34⁺ hHSPCs per mouse were transplanted intravenously into 4- to 6-week-old hSIRPα-DKO mice after the mice were irradiated twice with a total of 3.75 Gy (52). Blood samples were collected at 9 to 12 weeks posttransplantation, and their peripheral lymphocytes were analyzed by flow cytometry.

Adaptive transfer of stimulated CBMCs and in vivo HHV-6B infection.

To expand the numbers of human T cells in the humanized mice, a CD34-negative fraction of the same or different cord blood donor cells, or CBMCs without CD34-positive selection, was stimulated for 72 h with recombinant human interleukin-2 (IL-2) (2 ng/ml) and phytohemagglutinin (PHA) (5 μg/ml). After stimulation, the cells were harvested, washed twice with additive-free RPMI 1640, and injected intraperitoneally into hSIRPα-DKO-hCD34 mice. The portion of human CD3-positive cells was monitored by flow cytometry every week after the transfer of CBMCs. At the time when the percentage of hCD3⁺ cells among human CD45⁺ cells was >10% and the ratio of CD3⁺ T lymphocytes in hCD45⁺ cells was >90%, 200 μl of prepared cell-free HHV-6B (6.1 × 10⁴ 50% tissue culture infective doses [TCID₅₀]/ml) was injected intravenously into each mouse. The body weights of the mice were routinely measured 3 times per week, and the mice were euthanized when the weights had fallen to <85% compared with those before the transfer of CBMCs or at 2 weeks postinfection.

Flow cytometry.

Flow cytometry was performed with an SA3800 spectral cell analyzer (Sony Biotechnology, San Jose, CA). Peripheral blood from the mice was first treated with ACK lysis buffer (150 mM NH₄Cl, 10 mM KHCO₃, 0.1 mM EDTA) to remove the red blood cells. The remaining cells were washed twice with PBS containing 0.5% (wt/vol) bovine serum albumin (BSA) and 2 mM EDTA and then stained with fluorophore-conjugated antibodies on ice for 30 min. The cells were washed once after staining, and analysis was then performed. The proportion of human cells was calculated as follows: % hCD45⁺ = [hCD45⁺/(hCD45⁺ + mCD45⁺)] × 100%.

Real-time PCR.

Genomic DNAs were extracted from the tissues with the use of a Qiagen DNeasy blood and tissue kit (Qiagen, Hilden, Germany). The primers and probe for HHV-6B U67 and the PCR conditions were described previously (35).

Histological analyses and immunostaining.

Formalin-fixed and paraffin-embedded (FFPE) tissue blocks were sectioned at 3 μm and stained with hematoxylin and eosin (HE). For immunostaining, blocks were sectioned at 4 μm and stained with each antibody against human CD3 (1:50) and the antiserum against IE1 (1:2,000). The immunostaining process was performed using a Bond Max autostainer (Leica Microsystems, Wetzlar, Germany). Dual staining of human CD3 and IE1 was performed with a PermaBlue/AP kit (Diagnostic BioSystems, Pleasanton, CA) according to the manufacturer’s instructions.

Multiplex chemokine assays.

The serum levels of 40 human cytokines and chemokines (Table 3) in HHV-6B-infected and control mice were analyzed using the Bio-Plex human chemokine 40-plex panel (Bio-Rad Laboratories, Hercules, CA) on a Bio-Plex 200 system according to the manufacturer’s protocol.

Statistical analyses.

Statistical differences were determined by Student’s t test, and a P value of <0.05 was considered statistically significant.