Limited dissemination and shedding of the UL128 complex-intact, UL/b'-defective rhesus cytomegalovirus strain 180.92

Abstract

The UL128 complex of human cytomegalovirus (CMV) is a major determinant of viral entry into epithelial and endothelial cells and a target for vaccine development. The UL/b' region of rhesus CMV contains several open reading frames, including orthologs of the UL128 complex. We recently showed that the coding content of the rhesus CMV (RhCMV) UL/b' region predicts acute endothelial tropism and long-term shedding in vivo in the rhesus macaque model of CMV infection. The laboratory-passaged RhCMV 180.92 strain has a truncated UL/b' region but an intact UL128 complex. To investigate whether the presence of the UL128 complex alone was sufficient to confer endothelial and epithelial tropism in vivo, we investigated tissue dissemination and viral excretion following experimental RhCMV 180.92 inoculation of RhCMV-seronegative rhesus macaques. We show the presence of at least two virus variants in the RhCMV 180.92 infectious virus stock. A rare variant noted for a nontruncated wild-type-virus-like UL/b' region, rapidly emerged during in vivo replication and showed high-level replication in blood and tissues and excretion in urine and saliva, features similar to those previously reported in naturally occurring wild-type RhCMV infection. In contrast, the predominant truncated version of RhCMV 180.92 showed significantly lower plasma DNAemia and limited tissue dissemination and viral shedding. These data demonstrate that the truncated RhCMV 180.92 variant is attenuated in vivo and suggest that additional UL/b' genes, besides the UL128 complex, are required for optimal in vivo CMV replication and dissemination.

Importance: An effective vaccine against human CMV infection will need to target genes that are essential for virus propagation and transmission. The human CMV UL128 complex represents one such candidate antigen since it is essential for endothelial and epithelial cell tropism, and is a target for neutralizing antibodies in CMV-infected individuals. In this study, we used the rhesus macaque animal model of CMV infection to investigate the in vivo function of the UL128 complex. Using experimental infection of rhesus macaques with a rhesus CMV virus variant that contained an intact UL128 complex but was missing several other genes, we show that the presence of the UL128 complex alone is not sufficient for widespread tissue dissemination and virus excretion. These data highlight the importance of in vivo studies in evaluating human CMV gene function and suggest that additional UL/b' genes are required for optimal CMV dissemination and transmission.

FIG 1

Molecular characterization of RhCMV 180.92 by variant-specific conventional and quantitative real-time PCR. Schematic representation of the UL/b′ region ORF and alignment of fibroblast-adapted RhCMV 180.92 (accession number DQ120516) to wild-type (WT) RhCMV (EF990255). Conventional PCR primer pair names (Table 1), along with positions of their respective amplicons, are shown above the UL/b′ region alignments. Control pGEM-T Easy plasmid for quantitative real-time PCR and location of the RhCMV 180.92 and B-region primer-probe sets are shown in the bottom. Forward and reverse primers for the RhCMV 180.92 amplicon are located within UL148 and rh167, respectively. Forward and reverse primers for the B-region amplicon are located within UL132 and the intron spanning UL132 and UL148, respectively.

FIG 2

Conventional and real-time PCR amplification of RhCMV-specific segments within the UL/b′ region. DNA extracted from RhCMV UCD59 virus stock, RhCMV 180.92 virus stock, and frozen mesenteric lymph node (MLN) tissue collected at necropsy from the SIV-infected rhesus macaque Mm180.92, the monkey from whom RhCMV 180.92 was originally recovered, were used. (A) Conventional PCR runs with UL/b′ region primers on RhCMV UCD59 (left panel), RhCMV 180.92 (middle panel), and MLN from Mm180.92 (right panel). Segments within the UL/b′ region corresponding to the low-passage-number WT-like isolate RhCMV UCD59 (ULb′-1 to -4) are present in RhCMV strain 180.92 (middle panel). Amplicon specific to the truncated sequence of RhCMV 180.92 (ΔCED) is minimal to completely absent from DNA isolated from RhCMV UCD59 (left panel) and animal Mm180.92 (right panel), respectively. (B) Quantitative real-time PCR showing total and 180.92 truncated variant-specific RhCMV copy numbers per μg of DNA extracted from RhCMV UCD59, RhCMV 180.92, and MLN of Mm180.92. Median values of three replicate wells shown.

FIG 3

Histopathologic analysis of CMV localization in tissues of a SIV-infected CMV-seronegative rhesus macaque experimentally inoculated with RhCMV 180.92. Photomicrographs of H&E-stained and immunohistochemically stained tissue sections from a rhesus macaque with simian AIDS and disseminated CMV after SIV and RhCMV 180.92 coinfection. (A and B) RhCMV-induced inflammation in the heart (A) and colon (B) characterized by mild edema and tissue infiltration by lymphocytes and histiocytes admixed with cytomegalic and karyomegalic cells containing multiple large round to oval magenta-colored CMV intranuclear inclusion bodies surrounded by clear halo and chromatin margination (magnification, ×200; inset magnification, ×600). (C) Immunohistochemical analysis of kidney tissue showing distribution of RhCMV 180.92 infection in a wide range of cell types identified by double immunostaining of CMV-IE1 (magenta) with CD31 for endothelial cell identification (brown [a]), cytokeratin for epithelial cell identification (brown [b]), vimentin for fibroblast identification (brown [c]), and CD68 for macrophage identification (brown [d]). Magnification, ×600. (D) RhCMV-induced retinitis characterized by the presence of numerous immunohistochemically positive cells for CMV-IE1 (brown) associated with focal retinal degeneration and disruption of retinal internal and external nuclear layers shown at low (×200 [a]) and high (×400 [b]) magnifications.

FIG 4

Viral load and tissue distribution of RhCMV in a SIV-infected CMV-seronegative rhesus macaque experimentally inoculated with RhCMV 180.92. Real-time and conventional PCR data shown. (A) Median DNA copy numbers of total RhCMV plasma virus burden (solid line) and RhCMV 180.92 variant (dashed line). (B) Percentage of RhCMV 180.92 variant copy number (hashed column) compared to percentage of WT-like RhCMV variant (RhCMV-WT) copy numbers (solid column), as measured by variant-specific real-time PCR at 2, 3, 4, and 6 weeks after RhCMV 180.92 experimental inoculation. (C) Genomic DNA extracted from various tissues was assayed to differentiate total RhCMV (IE1), RhCMV-WT variant (UL146), and RhCMV 180.92 variant (ΔCED) using a variant-specific conventional PCR assay. Lane 18 is nontemplate control (water [NTC]) serving as a negative control, and DNA extracted from RhCMV 180.92 virus stock (Inoculum, lane 17) and β-actin amplifications (Actin) served as positive controls. (D) Quantitative analysis of RhCMV 180.92 variant DNA copy number/μg of DNA as measured by real-time PCR of 100 ng of tissue genomic DNA. Median values determined for three replicate wells are shown. ILN, inguinal lymph node; MLN, mesenteric lymph node; TLN, tracheobronchial lymph node; BM, bone marrow; FC, frontal cortex; SSC, sacral spinal cord.

FIG 5

Kinetics of plasma CMV DNA load and patterns of shedding of RhCMV 180.92 by variant-specific conventional and real-time PCR in five SIV-negative CMV-seronegative rhesus macaques experimentally inoculated with RhCMV 180.92. (A) DNA copy number of total RhCMV plasma virus burden (solid column) and RhCMV 180.92 variant (hashed column) measured by variant-specific real-time PCR at 1 and 4 weeks after RhCMV inoculation. Individual monkey data represent the median values obtained for three replicates. The graph at the far right shows a comparison of the mean total and variant RhCMV plasma copy number + the standard errors of the mean for animals with detectable CMV loads at 1 and 4 weeks after RhCMV inoculation. (B) Shedding of RhCMV 180.92 variant (hashed column) compared to total RhCMV (solid column) in urine and saliva samples collected at 8 weeks (Mm231.05, Mm320.06, and Mm332.06) or 20 weeks (Mm78.05 and Mm152.05) after RhCMV inoculation and measured by variant-specific real-time PCR. Individual monkey data represent the median values obtained for three replicates. The graph at the far right compares the mean total and variant copy numbers + the standard errors of the mean in urine and saliva of animals with detectable CMV load. A significant difference in CMV load was determined by using a paired Student t test. *, P < 0.05; **, P < 0.01. (C and D) RhCMV shedding patterns in urine (C) and saliva (D) samples were assayed to differentiate total RhCMV (IE1), RhCMV-WT variant (UL146), and RhCMV 180.92 variant (ΔCED) contributions by using a variant-specific conventional PCR assay. DNA extracted from RhCMV 180.92 virus stock (Inoculum) served as a positive control.