Macrophage-Derived Simian Immunodeficiency Virus Exhibits Enhanced Infectivity by Comparison with T-Cell-Derived Virus

Abstract

Human immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV) infect and productively replicate in macrophages and T lymphocytes. Here, we show that SIV virions derived from macrophages have higher levels of infectivity than those derived from T cells. The lower infectivity of T-cell-derived viruses is influenced by the quantity or type of mannose residues on the virion. Our results demonstrate that the cellular origin of a virus is a major factor in viral infectivity. Cell-type-specific factors in viral infectivity, and organ-specific or disease stage-specific differences in cellular derivation of virions, can be critical in the pathogenesis of HIV and AIDS.

CD4⁺ T cells and macrophages are the major targets and sources of human immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV). In view of the essential but highly distinct roles these cell types play in the pathogenesis of HIV (7, 9, 14, 22), understanding how viral production in each of these cell types affects the subsequent activity of virions is important for understanding the development of disease. To determine if different factors to which virions are exposed during their generation can influence their infectivity, we generated SIV stocks from both macrophages and T cells.

Two molecular clones of SIV, SIVmac316 (17) and SIVmac129 (6), were used to generate viral stocks from both macrophages and T cells. These SIV stocks (herein referred to as matching viral stocks) were derived from infection of primary rhesus monocyte-derived macrophages (MDM) or primary rhesus T cells. Primary MDM were derived from freshly isolated rhesus peripheral blood mononuclear cells by immunomagnetic CD11b selection and differentiated by adherence for 6 days in the presence of macrophage colony-stimulating factor (10 ng/ml). The remaining peripheral blood mononuclear cells were then subjected to CD8 immunomagnetic depletion to yield CD4⁺ enriched cells for the primary T-cell cultures and then stimulated with phytohemagglutinin (5 μg/ml) and interleukin 2 (10 ng/ml) for 3 days. Following stimulation, T cells were cultured in the presence of interleukin 2 alone for 3 additional days.

Both MDM and T cells were inoculated with SIV after 6 days in culture, and cell-free supernatants from each cell type were collected daily for 8 days after SIV inoculation and stored at −80°C. Stocks were then pooled and either aliquoted and frozen or purified by ultracentrifugation over a 20% sucrose cushion to eliminate contaminating cell-derived factors such as cytokines and then aliquoted and stored at −80°C. Stocks from macrophages and T cells were collected, stored, and processed simultaneously to avoid preparation related differences in infectivity. Viral stocks were generated from a total of 12 different rhesus donors; only macrophage and T-cell stocks derived from cells isolated from a single donor at the same time were used for comparison. SIV stocks were quantified by both enzyme-linked immunosorbent assay (ELISA) for p27 Gag (Beckman-Coulter) and branched-DNA assay (Bayer Reference Testing Laboratory) for viral genome quantification. These assays showed no significant difference between the matching viral stocks in terms of copies of viral RNA per pg of p27 Gag. Further testing by immunoblotting (using the KK45 antibody to detect Env and FA2 to detect Gag) indicated that the ratio of Env to Gag was constant in matching viral stocks. This is consistent with previous studies showing that virions produced in T cells and macrophages have similar levels of Env (2, 3), indicating that the virus-encoded aspects of the virion are likely similar in macrophage and T-cell-derived virions.

Initial experiments used equivalent amounts of non-ultracentrifuge-purified SIVmac316 matching viral stocks, derived from primary cultures from three monkeys, to infect primary rhesus MDM from two donors. In each case, 3- to 10-fold-higher levels of infection was achieved by the macrophage-derived versus the T-cell-derived stocks, determined by both number of cells infected (by immunofluorescence) and the level of p27 (by ELISA) in the supernatant (data not shown). Since both macrophages and T cells make cytokines, growth factors, and other molecules in response to infection, ultracentrifuge-purified viral stocks were then used to remove any potential contribution of these cellular products that could alter virion infectivity from viral stocks. Equivalent amounts of purified SIVmac316 matching viral stocks were then used to infect primary rhesus MDM derived from two donors. Inoculation with SIV derived from macrophages again resulted in higher levels of infectivity than did inoculation with SIV derived from T cells (n = 4) (Fig. 1). The greater number of cells infected by macrophage-derived virus corresponds with the increase in p27 production, indicating that macrophage-derived virions are able to infect cells more effectively than T-cell-derived virions. Subsequent studies performed with both unpurified and purified viral stocks yielded similar results; the data reported below were obtained with the purified stocks.

FIG. 1.

Macrophage-derived virus has greater infectivity than T-cell-derived virus. Monocyte derived macrophages plated on chamber slides were inoculated in triplicate with SIVmac316 derived from macrophages and T cells. Slides were stained to determine percentages of cells infected, and supernatants were analyzed for p27 levels. (A and B) Examples of immunofluorescence (red, anti-Gag visualized with rhodamine; blue, cell nuclei visualized with DAPI [4′,6′-diamidino-2-phenylindole]) of cells infected with matched stocks of macrophage-derived virions (A) and T-cell-derived virions (B). (C) Bars show percentages of macrophages infected with macrophage and T-cell-derived virus (inoculated using the p27 Gag levels indicated); symbols and lines show levels of virus in supernatants (measured by p27 Gag). The percentage of macrophages infected was determined by counting all the cells on a slide; thus, no error is given.

Additional studies were performed via infection of a number of different cell types with matching viral stocks derived from the different SIV strains. Cell types infected included two indicator lines, CCR5-expressing (Hi-5) GHOST cells (HOS derivative, which produce green fluorescent protein in response to infection), LuSIV cells (174xCEM derivative, which produce luciferase in response to infection), and primary rhesus MDM and primary rhesus T cells to better examine the potential in vivo relevance of this effect. Infection of primary cells was performed in MDM derived from six macaques and in T cells derived from three macaques. These infections were performed via spinoculation (2 h at 1,200 × g at 25°C), to minimize the effects of virion diffusion and attachment on infection and to synchronize the infections by facilitating virion attachment to the cell surface (18, 19). Infections were analyzed by ELISA quantification of supernatant p27 levels.

All infections in both the indicator and primary cells with matching viral stocks (Fig. 2A to E) resulted in a greater level of infection by macrophage-derived viral stocks than by T-cell-derived viral stocks. The increased infectivity of macrophage-derived virus over T-cell-derived virus in multiple cell types using different viral strains demonstrates that the enhanced infectivity of macrophage-derived virions is not dependent on viral strain or on target cell type but rather is dependent only on the cell type in which the virions were produced.

FIG. 2.

Cellular origin determines infectivity independently of viral strain and target cell. Hi-5 GHOST cells were spinoculated in triplicate with different concentrations of SIVmac 316 (A, B) or SIVmac129 (B) derived from either macrophages or T cells, fixed at 48 h, and analyzed by flow cytometry for green fluorescent protein expression to determine the percentage of cells infected. (C) LuSIV cells were spinoculated (six replicates) with SIVmac316 derived from primary macrophages or T cells and then analyzed by luminescence 40 h postinfection to determine the level of infectivity. (D) Primary macrophages were spinoculated in triplicate with the indicated amounts of SIVmac316 generated in primary T cells and primary macrophages, and then supernatants were analyzed for 6 days postinfection (P.I.) for p27 production. (E) Primary CD4⁺ T cells were infected with SIVmac316 derived from primary T cells and macrophages, and then supernatants were analyzed for 6 days postinfection for p27 production. (F) Doubly passaged infection. SIVmac316 generated in primary macrophages or T cells and then passaged through the other cell type was used for infection of macrophages by spinoculation. Supernatants were collected for 6 days and analyzed for p27 production. (G) Matched SIVmac316 virus stocks, produced in cells derived from two rhesus macaques (516 and 521), were added at 4°C to Hi-5 GHOST cells for 30 min, and then cells were washed and total RNA was isolated. Relative amounts of SIV adhering to the cells were determined by real-time quantitative PCR using primers for SIV and 18S RNA by the method. (H) Standard inoculation versus spinoculation. SIVmac316 generated in primary T cells and macrophages was used to infect Hi-5 GHOST cells by standard inoculation and spinoculation. Cells were analyzed at 48 h by FACS.

The enhanced infectivity of macrophage-derived virus could result from stable changes to the virus, such as alterations in its genome, or from transient changes, such as addition of cellular proteins to the viral surface. We generated two different twice-passaged viral stocks of SIVmac316 by generating viral stocks in either macrophages or T cells and subsequently passaging those stocks through the other cell type, i.e., macrophage-derived stocks were secondarily passaged in T cells and T-cell-derived stocks were secondarily passaged in macrophages. Identical amounts of both types of twice-passaged stocks were then used to infect primary MDM derived from two rhesus macaques. The macrophage-derived stocks secondarily passaged in T cells generated much lower levels of infection than the T-cell-derived stocks secondarily passaged in macrophages (Fig. 2F). Thus, the viral stocks functioned like virus derived from the cells in which they were last passaged, rather than the cells from which they were initially derived. Sequence analysis of three independent molecular clones of full-length gp120 derived from viral genomic RNA from each twice-passaged stock revealed no differences (data not shown), further supporting the idea that the enhanced infectivity results from transient changes to the virions.

To determine if generation in either macrophages or T cells altered virion infectivity through changes in viral attachment capacity, an attachment assay (modified from reference 20) was performed using Hi-5 GHOST cells, revealing no significant difference in attachment for matched stocks derived from two different monkeys (Fig. 2G). Additionally, Hi-5 GHOST cells were infected with identical amounts of matching SIVmac316 (Fig. 2H) and SIVmac129 (data not shown) stocks by both spinoculation and standard inoculation. In both types of inoculation, macrophage-derived virions produced higher levels of infection than did those derived from T cells, indicating that the increased infectivity of macrophage-derived viral particles is not due to increased or decreased attachment capacity.

As cell-type-specific variability in the glycosylation of HIV and SIV envelope proteins has been reported (10-12) and virion glycosylation has been shown to affect HIV and SIV infectivity (8, 15), we examined differences in virion glycosylation pathways as another possible mechanism for the increased infectivity of macrophage-derived virus. Microarray analysis was used to examine differences in the expression of genes involved in glycosylation processes between uninfected primary macrophages and T cells from three donors grown under the conditions used to produce our viral stocks. A large number (344) of gene transcripts differed significantly (P < 0.01) in expression between macrophages and T cells. A total of 42 genes, 24 involved in glycan degradation and 18 involved in glycan transfer, were at least twofold higher in macrophages than in T cells, and another 12 genes whose products are involved in glycan transferase activity were at least twofold higher in T cells than in macrophages (Table 1). Differential expression of two representative genes was corroborated by quantitative real-time PCR analyses, demonstrating increased expression of the mannosidase MAN2B1 gene in macrophages (15.0-fold ± 3.0-fold in the three donors, compared to their respective T cells) and increased expression of the sialyltransferase ST6GAL1 gene in T cells (6.1-fold ± 1.5-fold, compared to their respective macrophages).

TABLE 1.

Probe sets identifying genes involved in glycan degradation and glycan transferase that show significant changes between macrophages and T cells

Function and probe set	Accession no.	Gene symbol	Gene product name	GMIa		Increase (fold)
				Macrophages	T cells
Glycan degradation (increased in macrophages)
204443_at	NM_000487.3	ARSA	Arylsulfatase A	182.7	86.9	2.1
206129_at	NM_000046.1	ARSB	Arylsulfatase B	90.6	16.2	5.6
NM_004315_s_at	NM_004315	ASAH1	N-Acylsphingosine amidohydrolase (acid ceramidase) 1	1,154.8	58.1	19.9
202838_at	NM_000147.1	FUCA1	Fucosidase, α-l-1, tissue	7,910.3	115.1	68.7
202812_at	NM_000152.2	GAA	Glucosidase, alpha; acid	399.6	91.5	4.4
204417_at	NM_000153.1	GALC	Galactosylceramidase	703.7	196.7	3.6
206335_at	NM_000512.2	GALNS	Galactosamine (N-acetyl)-6-sulfate sulfatase	440.1	121.1	3.6
209093_s_at	K02920.1	GBA	Glucosidase, beta; acid (includes glucosylceramidase)	541.5	146.1	3.7
209093_at	K02920.1	GBA	Glucosidase, beta; acid (includes glucosylceramidase)	650.9	119.6	5.4
214430_at	NM_000169.1	GLA	Galactosidase, alpha	2437.7	391.9	6.2
201576_at	NM_000404.1	GLB1	Galactosidase, beta 1	964.1	254.6	3.8
209727_at	M76477.1	GM2A	GM2 ganglioside activator	1186.5	40.9	29.0
203676_at	NM_002076.1	GNS	Glucosamine (N-acetyl)-6-sulfatase	297.3	39.7	7.5
202605_at	NM_000181.1	GUSB	Glucuronidase, beta	1,369.9	387.9	3.5
201765_at	AL523158	HEXA	Hexosaminidase A (alpha polypeptide)	1,035.3	208.6	5.0
201944_at	NM_000521.2	HEXB	Hexosaminidase B (beta polypeptide)	4,679.3	522.8	9.0
209166_at	U68567.1	MAN2B1	Mannosidase, alpha, class 2B, member 1	934.4	297.3	3.1
203778_at	NM_005908.1	MANBA	Mannosidase, beta A, lysosomal	403.0	96.0	4.2
202943_at	M38083.1	NAGA	α-N-Acetylgalactosaminidase	496.5	208.0	2.4
204360_s_at	NM_000263.1	NAGLU	α-N-Acetylglucosaminidase	263.7	86.7	3.0
BC000722_x_at	BC000722	NEU1	Sialidase 1 (lysosomal sialidase)	818.7	213.1	3.8
208926_s_at	U84246.1	NEU1	Sialidase 1 (lysosomal sialidase)	643.0	140.2	4.6
AF338436_at	AF338436	NPL	N-Acetylneuraminate pyruvate lyase	2,027.9	26.3	77.1
200661_at	NM_000308.1	PPGB	Protective protein for β-galactosidase	2,307.8	357.6	6.5
203767_at	AI122754	STS	Steroid sulfatase (microsomal), arylsulfatase C, isozyme S	61.6	27.9	2.2
233555_s_at	NM_018837	SULF2	Sulfatase 2	45.4	19.2	2.4
224724_at	NM_018837	SULF2	Sulfatase 2	93.9	26.7	3.5
Glycan transferase (increased in macrophages)
AB049584_s_at	AB049584	B3GNT1	UDP-GlcNAc:β-Gal β-1,3-N-acetylglucosaminyltransferase 1	575.5	199.2	2.9
225612_at	NM_032047	B3GNT5	UDP-GlcNAc:β-Gal β-1,3-N-acetylglucosaminyltransferase 5	229.6	18.8	12.2
1555962_x_at	NM_145236	B3GNT7	UDP-GlcNAc:β-Gal β-1,3-N-acetylglucosaminyltransferase 7	36.7	13.0	2.8
221484_at	NM_004776.2	B4GALT5	UDP-GlcNAc:β-Gal β-1,4- galactosyltransferase, polypeptide 5	324.9	95.1	3.4
239647_at	NM_152889	CHST13	Carbohydrate (chondroitin 4) sulfotransferase 13	166.5	33.3	5.0
206756_at	NM_019886.1	CHST7	Carbohydrate (N-acetylglucosamine 6-O) sulfotransferase 7	249.8	115.5	2.2
219956_x_at	NM_007210.2	GALNT6	UDP-N-acetyl-α-d-galactosamine:polypeptide N-acetylgalactosaminyltransferase 6	358.3	154.8	2.3
219956_at	NM_007210.2	GALNT6	UDP-N-acetyl-α-d-galactosamine:polypeptide N-acetylgalactosaminyltransferase 6	678.4	195.5	3.5
231780_at	NM_021996.3	GBGT1	Globoside α-1,3-N-acetylgalactosaminyltransferase 1	203.1	100.3	2.0
230788_s_at	NM_145649	GCNT2	Glucosaminyl (N-acetyl) transferase 2, I-branching enzyme	143.2	70.5	2.0
213552_at	NM_015554	GLCE	Glucuronyl C5-epimerase	32.2	12.1	2.7
205466_s_at	NM_005114.1	HS3ST1	Heparan sulfate (glucosamine) 3-O-sulfotransferase 1	149.8	9.8	15.3
219697_at	NM_006043.1	HS3ST2	Heparan sulfate (glucosamine) 3-O-sulfotransferase 2	1,059.6	11.7	90.6
201126_s_at	NM_002406.2	MGAT1	Mannosyl (α-1,3)-glycoprotein β-1,2-N-acetylglucosaminyltransferase	1,014.6	303.2	3.3
224598_at	NM_054013	MGAT4B	Mannosyl (α-1,3)-glycoprotein β-1,4-N-acetylglucosaminyltransferase, isozyme B	934.3	461.7	2.0
220189_s_at	NM_014275.1	MGAT4B	Mannosyl (α-1,3)-glycoprotein β-1,4-N-acetylglucosaminyltransferase, isozyme B	391.1	175.2	2.2
224598_s_at	NM_054013	MGAT4B	Mannosyl (α-1,3)-glycoprotein β-1,4-	375.6	158.1	2.4
202607_at	NM_001543.3		N-acetylglucosaminyltransferase,	141.2	41.9	3.4
210942_s_at	NM_006100.2		isozyme B	155.7	4.8	32.4
204542_at	NM_006456.1	ST6GALNAC2	ST6 (α-N-acetyl-neuraminyl-2,3-β-galactosyl-1,3)-N-acetylgalactosaminide α-2,6-sialyltransferase 2	115.4	25.1	4.6
218801_at	NM_020121.2	UGCGL2	UDP-glucose ceramide glucosyltransferase-like 2	39.4	7.4	5.3
Glycan transferase (increased in T cells)
219649_at	NM_013339	ALG6	Asparagine-linked glycosylation 6 homolog (yeast, α-1,3-glucosyltransferase)	131.0	308.0	2.4
211812_s_at	NM_003781.2	B3GALNT1	β-1,3-N-Acetylgalactosaminyltransferase 1	8.0	18.7	2.3
211379_s_at	NM_033168	B3GALNT1	β-1,3-N-Acetylgalactosaminyltransferase 1	4.9	12.3	2.5
223374_at	NM_033169	B3GALNT1	β-1,3-N-	14.8	40.3	2.7
219439_at	NM_020156.1		Acetylgalactosaminyltransferase 1	23.3	47.0	2.0
219049_at	NM_018371			10.2	84.7	8.3
203921_at	NM_004267.1	CHST2	Carbohydrate (N-acetylglucosamine-6-O) sulfotransferase 2	58.4	119.4	2.0
203044_at	NM_014918	CHSY1	Carbohydrate (chondroitin) synthase 1	112.1	240.0	2.1
201995_at	NM_000127.1	EXT1	Exostoses (multiple) 1	31.2	84.4	2.7
210506_at	NM_004479.2	FUT7	Fucosyltransferase 7 (α-1,3-fucosyltransferase)	36.5	76.5	2.1
1554930_s_at	NM_178157	FUT8	Fucosyltransferase 8 (α-1,6-fucosyltransferase)	34.5	265.6	7.7
203988_s_at	NM_004480.1	FUT8	Fucosyltransferase 8 (α-1,6-fucosyltransferase)	28.1	292.0	10.4
204152_s_at	NM_002405	MFNG	Manic fringe homolog (Drosophila)	63.7	238.0	3.7
204153_s_at	NM_002405.1	MFNG	Manic fringe homolog (Drosophila)	133.3	541.0	4.1
207563_s_at	NM_181673			96.8	466.5	4.8
201998_at	NM_173216.1	ST6GAL1	ST6 β-galactosamide α-2,6-sialyltranferase 1	116.9	1,155.1	9.9

^aGMI, geometric mean intensity from three independent samples.

To examine the functional effects of differential glycosylation of viral particles, we removed glycans from virion surfaces using different glycosidases and then examined the infectivity of glycosidase-treated matching stocks in LuSIV indicator cells. Infection of these cells with α-2,3,6,8,9-neuraminidase-treated matching stocks of SIVmac316 led to similar, small changes in infectivity in both macrophage and T-cell-derived viral stocks (Fig. 3A and B). In contrast, treatment of matching stocks with α-1,2,3-mannosidase showed no significant enhancement in infectivity of macrophage-derived virions (Fig. 3C) but a 3.5-fold increase in infectivity of T-cell-derived virions (P < 0.01, n = 6, analysis of variance [ANOVA] with Tukey's post hoc test) (Fig. 3D).

FIG. 3.

Removal of mannose from the virion surface affects infectivity. SIVmac316 derived from primary macrophages (A, C, E) or T cells (B, D, F) was digested with the indicated concentrations of glycosidases and used to infect LuSIV cells by spinoculation. Cells were then analyzed by luminescence 40 h postinfection. All statistical analyses performed by ANOVA with Tukey's post hoc test; n = 6 for each experiment.

Additional infections were performed with matching viral stocks treated with a distinct mannosidase enzyme, α-1,2,3,6-mannosidase, to examine more complete removal of mannose from the virion. Macrophage-derived stocks of SIVmac316 treated with α-1,2,3,6-mannosidase demonstrated a significant but slight increase in infectivity of less than twofold (Fig. 3E; P < 0.01, n = 6, ANOVA with Tukey's post hoc test). As with α-1,2,3-mannosidase treatment, T-cell-derived stocks treated with α-1,2,3,6-mannosidase demonstrated a large increase in infectivity, up to ninefold (Fig. 3F; P < 0.01, n = 6, ANOVA with Tukey's post hoc test).

Changes in gp120 and gp41 glycosylation patterns can strongly alter the infectivity of different strains of HIV (10-12). Several studies show that desialylation of viral particles via neuraminidase digestion enhanced the HIV/SIV infectivity (8, 13, 15, 16, 21), in general agreement with the findings in this report. The discovery that mannosidase treatment increases the infectivity of virions derived from T cells differs from previous reports (8, 15). However, the cellular origin of a virion is crucial to examination of glycosylation-mediated effects on infectivity due to differential exposure to a variety of different glycosylation related pathways and enzymes during virion production (10, 11, 23), and unlike in previous studies, the virus in this report was generated in primary macrophages and T cells.

The data discussed above demonstrate a significant cell-type-specific variation in the infectivity of macrophage- and T-cell-derived SIV particles and suggest that the greater infectivity of macrophage-derived virions over T-cell-derived virions is due to a lower number of mannose residues on the surface of viral particles derived from macrophages. Support for this idea comes from data revealing that the glycosylation patterns of the heavily glycosylated, mannose-rich gp120 (4, 24) differ in a host cell-specific fashion (5, 10-12). We also find that several genes whose products degrade glycans are increased in macrophages. The fact that the majority of the glycoconjugates on gp120 contain high-mannose glycans (4, 24) further strengthens the possibility that mannose residues on gp120 affect infectivity. On the other hand, differential glycosylation of numerous cellular proteins incorporated into virions in a cell type-dependent manner (1), as opposed to effects on viral proteins, could also change viral infectivity.

These data show that the viral infectivity of SIV is influenced by the cellular origin of the virion and that macrophages generate virions with greater levels of infectivity than T cells. The data suggest that these differences may be due to differences in the number or arrangement of mannose residues on the surfaces of macrophage- and T-cell-derived virions. The implications of this finding are important in the context of HIV infection and disease progression, particularly in the study of transmission, the dynamics of viral spread, and organ-specific viral evolution and pathogenesis. As macrophages and T cells each play distinct roles in disease, the finding that viruses derived from the different cell types may behave differently has important implications for our understanding of HIV pathogenesis.