Human Herpesvirus 6 (HHV-6) Causes Severe Thymocyte Depletion in SCID-hu Thy/Liv Mice

SCID-hu Thy/Liv mice were obtained by coimplanting small pieces of human second trimester fetal liver and thymus under the kidney capsule of male homozygous CB-17 scid/scid (SCID) mice (20). Thy/Liv implants were infected 4–7 mo after implantation, as described (23). In brief, mice were anesthetized, the left kidney was surgically exposed, and the implant was inoculated with 25– 50 μl of viral stock with a titer of 10⁵ (strain GS) or 10⁴ (strain PL-1) 50% tissue culture infectious doses (TCID₅₀) per 50 μl. UV-inactivated viral stock and culture medium were used for mock inoculations. The abdominal wall incision was sutured, and the skin was closed with staples. All procedures and practices were approved by the University of California, San Francisco Committee on Human Research or by the University of California, San Francisco Committee on Animal Research.

HHV-6A_GS, a primary subgroup A isolate, was obtained from cultures of purified human cord blood CD4⁺ T cells infected with cell-free viral stock at a multiplicity of infection of 0.5. At day 6 after infection, the supernatants were collected, centrifuged at 2,500 g, and stored at −80°C. HHV-6B_PL-1, a primary subgroup B isolate, was obtained from in vitro–activated peripheral blood cells of a healthy adult blood donor and propagated for a single passage in activated cord blood CD4⁺ T cells. No HHV-7 contamination was detected by nested PCR assay, and the stocks were also negative for Mycoplasma. Virus titration was performed by infecting triplicate cultures of phytohemagglutinin-activated human peripheral blood mononuclear cells with serial 10-fold dilutions of the viral stock.

SCID-hu mice were killed by CO₂ inhalation, and implants were removed at different times after inoculation. Implants were processed for histopathology and electron microscopy as detailed below. For cytofluorimetric analysis, cell sorting, and quantitative PCR, single-cell suspensions were obtained by grinding the implants in a nylon bag. The cells were counted in a Coulter counter and divided into aliquots for the different assays.

Thy/Liv implants were fixed in 10% buffered formalin, embedded in paraffin, cut into 4-μm-thick sections, and stained with hematoxylin and eosin (H&E). For immunohistochemistry, unstained sections were pretreated with trypsin and pronase. The sections were then stained with a polyclonal antibody specific for CD3 (Dako) or an mAb specific for CD68 (Dako), using a standard peroxidase-labeled streptavidin-biotin (LSAB-peroxidase) method (LSAB2; Dako), followed by visualization with diaminobenzidine (DAB) and CuSO₄ enhancement (24). Before the second phase of staining, the samples were denatured in 2 N HCl for 10 min. The sections were then stained with 9A5D12, an mAb specific for the p41 early protein of HHV-6 (25). This antibody was revealed by a standard LSAB-alkaline phosphatase method (LSAB2; Dako) followed by visualization with Vector Red (Vector Laboratories). Negative controls included purified nonimmune mouse or rabbit serum (Dako), followed by LSAB-peroxidase staining. The sections were then denatured in HCl, and LSAB-alkaline phosphatase staining was carried out after incubation with either mouse serum (no stain) or 9A5D12 (red staining only).

Small pieces of Thy/Liv implants were fixed in 2.5% glutaraldehyde, and samples were postfixed in osmium tetroxide, dehydrated in 2,2-dimethoxypropane, and embedded in Poly/Bed 812 resin (Polysciences, Inc.). Thin sections (600–900 Å) were cut with a diamond knife, mounted on copper grids, stained with uranyl acetate and lead citrate, and examined on a JEOL 100 SX electron microscope.

Single-cell suspensions were stained with an mAb cocktail containing CD4-FITC (Becton Dickinson), CD8-PE (Becton Dickinson), and CD3-tricolor (CD3-TC) (Caltag). Cells were washed, resuspended in PBS containing 1% paraformaldehyde, and analyzed on a FACScan^® (Becton Dickinson) for CD4, CD8, and CD3 expression after gating on a live-cell lymphoid population identified by forward- and side-scatter characteristics. For cell sorting, unfixed cells were stained as for FACS^® analysis. A FACS Vantage™ (Becton Dickinson) was used to purify four populations of thymocytes: CD3⁺ CD4⁺CD8⁻ (SP4), CD3⁺CD4⁻CD8⁺ (SP8), CD3⁺CD4⁺CD8⁺ (DP), and CD3⁻CD4⁺CD8⁻ (ITTP). Sorted cells were frozen as dry pellets and stored at −80°C for DNA extraction.

DNA was extracted from frozen pellets by proteinase K digestion and purified by phenol-chloroform extraction by standard techniques. Viral load was quantitated with the TaqMan^® fluorogenic detection system on an ABI Prism 7700 Sequence Detector^® (Perkin-Elmer Applied Biosystems [26]). PCR was performed with primers 5′-CAAAGCCAAATTATCCAGAGCG-3′ and 5′-CGCTAGGTTGAGGATGATCGA-3′ and a probe located in the highly conserved open reading frame U67 of HHV-6 (27): 5′-CACCAGACGTCACACCCGAAGGAAT-3′. PCR amplification consisted of denaturation at 95°C for 15 min, followed by 40 cycles of denaturation at 95°C for 15 s and annealing/extension at 58°C for 60 s (Locatelli, G., F. Santoro, A. Gobbi, F. Veglia, P. Lusso, and M. Malnati, manuscript in preparation).

Nonparametric statistical analyses were performed using the Mann-Whitney U test (StatView 5.0; Abacus Concepts).

58 SCID-hu Thy/Liv mice from 6 different cohorts, each made with tissues from a single fetal donor, were inoculated with HHV-6 by direct intrathymic injection of 10⁵ TCID₅₀ of strain GS or mock infected with UV-inactivated viral stock or with cell culture medium. Mice were killed 4, 7, 11, 14, and 27 d after inoculation, and the implants were removed. Viral replication was monitored by quantitative PCR on total thymocyte suspensions with the TaqMan^® real time detection system. HHV-6 DNA was detected as early as 4 d after inoculation in each of the three implants examined (Fig. 1). The viral load peaked on day 14 with a mean value of ∼6,000 copies/10³ cells. In the implants harvested 27 d after inoculation, a drastic reduction in the number of genomic copies, with a mean of ∼400 copies/ 10³ cells, was observed (Fig. 1). No HHV-6 DNA was detected in the mock-infected control implants. The productive nature of HHV-6A_GS infection in Thy/Liv implants was demonstrated by transmission electron microscopy showing viral particles at different stages of maturation (Fig. 2). In particular, naked viral nucleocapsids were observed in the nucleus, tegument-coated particles in the cytoplasm (not shown), and mature, enveloped virions in cytoplasmic vacuoles. This pattern is consistent with the documented ultrastructural development of HHV-6 in cells infected in vitro (28).

HHV-6A_GS infection of Thy/Liv implants was associated with a marked, progressive depletion of thymocytes (Fig. 1). The number of thymocytes harvested from the HHV-6A_GS–infected implants was significantly reduced at day 14 (P < 0.01) compared with mock-infected implants and was reduced by >98% by day 27 (P < 0.001). A representative HHV-6A_GS–infected Thy/Liv implant is shown in Fig. 3 A. Compared with the mock-infected implant shown in Fig. 3 B, the HHV-6A_GS–infected thymus had a marked expansion of the medullary zone, apparent destruction or compression of the cortical zone, and destruction of the corticomedullary junction (Fig. 3 A). The medullary expansion appeared to be driven by endothelial proliferation or by vascular ectasia and fibrocyte proliferation. The endothelial cells lining medullary vessels appeared to be activated, in a fashion resembling vascular endothelium in the setting of inflammation. The expanded medullary zone contained atypical cells with prominent nuclear inclusions (Cowdry type A; Fig. 3 C). Immunohistochemical staining with an mAb specific for the p41 early protein of HHV-6 demonstrated that most of these atypical cells were infected (Fig. 3 D). Double-staining analysis showed that virtually all the infected cells were CD3⁺, albeit with varying degrees of intensity (Fig. 3 E). By contrast, double-label staining with CD68 did not indicate a tropism of HHV-6A_GS for macrophages in this model (Fig. 3 F).

To evaluate the effects of viral replication on different thymocyte subpopulations, we performed quantitative flow cytometry on total thymocyte suspensions obtained from six different cohorts of SCID-hu Thy/Liv mice in three independent experiments. HHV-6A_GS–infected implants were harvested 4, 7, 11, 14, and 27 d after inoculation and examined for expression of CD3, CD4, and CD8. 4 and 7 d after inoculation, the percentages of the different thymocyte subsets were unchanged compared with mock-infected implants (Table I). The earliest detectable phenotypic change was a reduced percentage of ITTPs at day 11 (P < 0.05), a time at which the percentages of the other thymocyte subpopulations were not yet significantly changed (Table I). By day 14, a reduction in the percentage of live thymocytes, identified by forward- and side-scatter profiles, was observed (Table I). At this time point, ITTPs remained significantly depleted (P < 0.01), and there was also a reduction in the percentage of DP thymocytes (P < 0.001) and an increase in the percentage of SP4 (P < 0.05) and SP8 (P < 0.01) thymocytes (Table I). By day 27, most of the thymocytes were dead or dying, and it was no longer possible to analyze the different subpopulations (Table I).

To study the cellular tropism of HHV-6A_GS in Thy/Liv implants, the distribution of viral DNA was quantitated in the different subpopulations of thymocytes as a function of time after inoculation. Total thymocytes were harvested at 4, 7, 11, and 14 d after inoculation and sorted into DP, SP4, SP8, and ITTP subpopulations. The viral load in each subpopulation was measured by quantitative PCR (Table II). At day 4 after inoculation, viral DNA was detectable in ITTPs but was less abundant or undetectable in the other subpopulations. By day 7, HHV-6 DNA was detectable in all the thymocyte subsets, but was significantly more abundant in ITTPs (P < 0.05) and remained so at 11 and 14 d after inoculation. No HHV-6 DNA was detected in any of the thymocyte subpopulations from Thy/Liv implants infected with UV-inactivated viral stock (data not shown).

To study the effects of HHV-6 subgroup B on Thy/Liv implants, six SCID-hu Thy/Liv mice were inoculated with 10⁴ TCID₅₀ of strain PL-1 (a primary isolate passaged in vitro only twice), or mock infected with cell culture medium. Mice were killed 10 d after inoculation, and the implants were removed. Quantitative PCR analysis revealed efficient replication of HHV-6B_PL-1 (mean viral load of ∼2,200 copies/10³ cells) similar to that found in HHV-6A_GS–infected implants (Table III). The effects of HHV-6B_PL-1 replication on thymocyte subpopulations were assessed by flow cytometry. As observed for strain GS at a similar time point (day 11; Table I), the main effect of HHV-6B_PL-1 on thymocyte subpopulations was a reduction in the percentage of ITTPs (Table III). At this relatively early time point, the percentage of live thymocytes was also decreased (Table III).

This study provides the first evidence that HHV-6 can replicate in the human thymus in vivo, leading to a progressive destruction of the organ. The severity of the lesions caused by HHV-6 in Thy/Liv implants supports the hypothesis that this virus may be immunosuppressive, as previously suggested by its tropism for CD4⁺ T cells and by the severe cytopathic effects it exerts on infected cells in vitro. Interestingly, the histological lesions we observed in HHV-6–infected Thy/Liv implants are similar to those recently described in the thymus of an HIV-1–negative infant with immunosuppression associated with widespread HHV-6 infection (5), suggesting that SCID-hu Thy/ Liv mice may reproduce the pathological mechanisms of HHV-6 infection in humans. However, despite the high prevalence of HHV-6 infection in the human population, only rare cases of putative HHV-6–related immunosuppression have been reported to date. It remains to be determined whether thymic infection by HHV-6 is an uncommon occurrence or, alternatively, whether its pathological consequences in vivo may be attenuated by host immune responses.

The experiments reported here were focused mainly on HHV-6A_GS, but limited studies with a subgroup B virus (PL-1) indicate that viruses of both subgroups can replicate in the Thy/Liv organ and induce cytopathicity. Clinical and epidemiological studies have shown an association of both subgroup A and B viruses with immunodeficiency (2), although subgroup B viruses have also been frequently isolated from the general population of nonimmunocompromised subjects. Coinfection by A and B strains was also documented in the infant with HHV-6–associated immunosuppression mentioned above (5). Further work will be required, both in humans and in SCID-hu Thy/Liv mice, to determine whether there are any intrinsic differences between the two HHV-6 subgroups in their propensity to induce immunosuppression in vivo.

The immunohistochemical detection of the CD3 antigen on the majority of the infected cells suggests that mature thymocytes can be infected by HHV-6 in vivo. However, the early and persistent depletion of ITTPs indicates that HHV-6 has a particular tropism for these immature cells. Indeed, quantitative PCR on sorted thymocyte subpopulations showed a markedly higher HHV-6 DNA load in ITTPs than in the other subpopulations. ITTPs are a relatively rare but rapidly dividing subpopulation that represents an early step of thymocyte maturation and gives rise to DP thymocytes and their mature SP progeny (21). Destruction of ITTPs may result in the generalized suppression of thymopoiesis and ultimately in the depletion of the thymus.

The thymus is the primary site of T lymphocyte development in utero and in early life. Although its activity diminishes progressively with age, this organ may still be functional in adults and, during conditions of severe T cell depletion (including HIV-1 infection), it may play an important role in maintaining and reconstituting the peripheral immune system (29–32). Accordingly, abundant thymic tissue has been detected by chest computed tomography in HIV-1–seropositive adults, suggesting that thymic function may be enhanced, by a compensatory mechanism, in some HIV-1–infected patients (32). However, persistent thymic function in an adult could be quickly abolished by thymic infection with HIV or other T cell–tropic viruses such as HHV-6.

Because of its tropism for CD4⁺ T lymphocytes and its positive interactions in vitro with HIV-1, HHV-6 has been suggested as a possible cofactor in AIDS (6). This hypothesis is supported by clinical evidence of active and widespread HHV-6 infection in symptomatic AIDS patients (7–9). HHV-6 infects human thymocytes in vitro (3) and has been detected in the thymus in vivo (5, 33). Thus, HHV-6 infection of the thymus may play a role in the immunodeficiency associated with HIV-1 infection. However, it is technically difficult to study the thymus in humans, and no reports are currently available on the presence of thymic HHV-6 infection in patients with HIV disease. The SCID-hu Thy/Liv mouse model, carrying a human graft that is morphologically and functionally equivalent to the human thymus, provides a unique system to study the pathogenesis of HHV-6 and HIV-1 infections in vivo (10–14). Interestingly, some chemokine receptor CXCR4-using strains of HIV-1, like NL4-3, suppress thymopoiesis in SCID-hu Thy/Liv mice by direct infection and destruction of ITTPs (11). The tropism of HHV-6 and HIV-1 for the same thymocyte subset raises the possibility of synergistic interactions in the thymic environment. Future studies of coinfection with HHV-6 and HIV-1 in SCID-hu Thy/Liv mice may provide important insights into the role of HHV-6 in HIV-1 disease.

We thank Morgan Jenkins for critical review of this manuscript, Stephen Ordway and Gary Howard for editing the manuscript, and Lisa Gibson and Eric Wieder for assistance with cell sorting.

This work was supported by European Union Biomed 2 grant PL951301 and Istituto Superiore di Sanità-Rome grant 40A059 (to P. Lusso), and by National Institute of Allergy and Infectious Diseases grants AI40312, AI65309 (to J.M. McCune), and CA73534 (to B.G. Herndier). J.M. McCune is an Elizabeth Glaser Scientist supported by the Elizabeth Glaser Pediatric AIDS Foundation.

DAB

diaminobenzidine

DP

CD3⁻ CD4⁺CD8⁺ and CD3⁺CD4⁺CD8⁺ double-positive

H&E

hematoxylin and eosin

HHV-6

human herpesvirus 6

ITTP

intrathymic T progenitor cell

SCID

severe combined immunodeficiency

SP4

mature CD3⁺ CD4⁺CD8⁻ single-positive 4

SP8

mature CD3⁺CD4⁻CD8⁺ single-positive 8

TCID₅₀

50% tissue culture infectious dose