Identification of Emerging Macrophage-Tropic HIV-1 R5 Variants in Brain Tissue of AIDS Patients without Severe Neurological Complications

ABSTRACT

Untreated HIV-positive (HIV-1⁺) individuals frequently suffer from HIV-associated neurocognitive disorders (HAND), with about 30% of AIDS patients suffering severe HIV-associated dementias (HADs). Antiretroviral therapy has greatly reduced the incidence of HAND and HAD. However, there is a continuing problem of milder neurocognitive impairments in treated HIV⁺ patients that may be increasing with long-term therapy. In the present study, we investigated whether envelope (env) genes could be amplified from proviral DNA or RNA derived from brain tissue of 12 individuals with normal neurology or minor neurological conditions (N/MC individuals). The tropism and characteristics of the brain-derived Envs were then investigated and compared to those of Envs derived from immune tissue. We showed that (i) macrophage-tropic R5 Envs could be detected in the brain tissue of 4/12 N/MC individuals, (ii) macrophage-tropic Envs in brain tissue formed compartmentalized clusters distinct from non-macrophage-tropic (non-mac-tropic) Envs recovered from the spleen or brain, (iii) the evidence was consistent with active viral expression by macrophage-tropic variants in the brain tissue of some individuals, and (iv) Envs from immune tissue of the N/MC individuals were nearly all tightly non-mac-tropic, contrasting with previous data for neuro-AIDS patients where immune tissue Envs mediated a range of macrophage infectivities, from background levels to modest infection, with a small number of Envs from some patients mediating high macrophage infection levels. In summary, the data presented here show that compartmentalized and active macrophage-tropic HIV-1 variants are present in the brain tissue of individuals before neurological disease becomes overt or serious.

IMPORTANCE The detection of highly compartmentalized macrophage-tropic R5 Envs in the brain tissue of HIV patients without serious neurological disease is consistent with their emergence from a viral population already established there, perhaps from early disease. The detection of active macrophage-tropic virus expression, and probably replication, indicates that antiretroviral drugs with optimal penetration through the blood-brain barrier should be considered even for patients without neurological disease (neuro-disease). Finally, our data are consistent with the brain forming a sanctuary site for latent virus and low-level viral replication in the absence of neuro-disease.

INTRODUCTION

Untreated HIV-positive (HIV-1⁺) individuals frequently suffer from HIV-associated neurocognitive disorders (HAND), which include asymptomatic neurocognitive impairment (ANI), mild neurocognitive disorder (MND), and HIV-associated dementia (HAD) (1). Overall, about 7% of untreated HIV⁺ individuals (30% of AIDS patients) suffer the more severe HADs (2, 3). In the era of highly active antiretroviral therapy (HAART), HAD and other neurocognitive issues are greatly reduced, with an overall incidence of HAD of only around 1% (3). Nevertheless, there is a continuing problem of milder neurocognitive impairments in HIV⁺ patients. These mild impairments may increase in severity with long-term therapy (3–5), while severe neurocognitive disorders, including HAD, still occur in untreated individuals and those who fail therapy (6).

Inside the brain, differentiated perivascular macrophages are the major reservoir of HIV-1 (2, 7–12). Resident microglia (also of the macrophage lineage) are also infected (7, 8, 13), while some studies support astrocyte infection (9, 14–16), particularly in pediatric cases (17–19). Neurons are rarely infected (2). HIV-1 colonizes cerebrospinal fluid (CSF) and the brain early in infection (20, 21). However, proviral DNA in brain tissue is difficult to detect during the asymptomatic phase (22–24). HAND and HAD that occur later in disease result either from the reactivation of resident HIV or by reseeding of HIV into the brain by HIV⁺ cells passing through the blood-brain barrier late in disease. HIV-1 sequences in the brain are frequently (but not always) sharply compartmentalized from those in blood and/or immune tissue (25–43). Moreover, compartmentalized sequences consistent with macrophage infection can be detected early in infection (within 4 months) (20). Together, these data are consistent with the reactivation of a compartmentalized viral population first established in the central nervous system (CNS) early in disease. However, activated monocytes in the blood were shown to migrate through the blood-brain barrier into brain tissue in increasing numbers late in disease (44). Several studies indicate that blood monocytes are infected in vivo (45–51), consistent with reseeding of the brain later on (45, 52–54). However, it is also possible that the influx of uninfected monocytes that differentiate into macrophages once in the brain supports the amplification of HIV-1 already established there.

HIV-infected T cells also have the potential to carry HIV into brain tissue. HIV⁺ T cells are frequently present in the CSF of patients with neurological complications, sometimes associated with pleocytosis (55). HIV in CSF may sometimes be compartmentalized from the virus in blood (55). It is well established that T cells circulate through brain tissue at low levels (56). The recent identification of lymphatic vessels in mice that drain CSF lymphoid cells (including T cells) from the dura mater to lymph nodes (LNs) (57, 58) highlights the importance of T cells in the immune surveillance of the brain and their potential to act as carriers of HIV into the CNS. However, whether T cells in CSF are the source of those that traffic through brain tissue is not yet clear.

Highly macrophage-tropic R5 (mac-tropic; also called R5-M-tropic) (55, 59) variants are predominant in brain tissue of AIDS patients with neurological disease (25, 55, 60–62). These mac-tropic variants can exploit low CD4 levels for infection (25, 60–62) and appear to have adapted for replication in macrophages and microglia, the predominant CD4⁺ cell types in the brain. In contrast, the vast majority of R5 Envs amplified from plasma or immune tissue (e.g., spleen or lymph node tissue) of AIDS patients with neurological diagnoses conferred weak or only modest infection of primary macrophages (25, 55, 60, 61). These late-stage, non-mac-tropic R5 viruses are likely T cell tropic (R5-T-tropic) (55, 59) and were reported to evolve an increased positive charge and fitness and reduced sensitivity to CCR5 inhibitors (25, 63–66), adaptations associated with more efficient CCR5 use and increased replication in T cells.

The genotypes and phenotypes of HIV-1 Envs present in the deep brain tissue of individuals with normal neurological function have not been studied extensively. The proviral load in brain tissue is low before neurological disease (neuro-disease) develops (22–24), making it difficult to amplify HIV env sequences. However, if env sequences were amplified, they could provide insights into Env genotypes and properties associated with HIV colonization of brain tissue before the highly mac-tropic, neurovirulent variants that are associated with HAD develop. In the present study, we examined whether env genes could be amplified from proviral DNA or RNA derived from brain tissue of 12 individuals with normal neurology or with minor neurological conditions (N/MC individuals). The tropism and characteristics of the brain-derived Envs were then investigated and compared to those of Envs derived from immune tissue of the same individuals.

Our data confirm that N/MC patients have lower levels of HIV proviral DNA in the brain than those of individuals with neuro-AIDS. However, a substantial number of env genes were amplified from brain tissue DNA and/or RNA for eight individuals. These included a subset of mac-tropic env sequences that were amplified from either proviral DNA or RNA from the brain tissue of 4 individuals, where the sequences of mac-tropic Envs were segregated from those of non-mac-tropic Envs present in either the spleen, plasma, or brain. Although these observations are derived from a small number of individuals, they indicate that macrophage-tropic variants are established and likely to be replicating in the brain for a percentage of HIV⁺ individuals, well before serious neurological dysfunction becomes apparent.

RESULTS

PCR amplification of env sequences from proviral DNA in frontal lobe and immune tissues of N/MC individuals.

We first investigated whether env genes could be amplified and cloned from proviral DNA present in frontal lobe or spleen tissue from 12 HIV⁺ patients, the majority with no signs of neurological impairment and the others with only minor complications (N/MC individuals) (Table 1). A significant number of env clones were obtained from PCR endpoint dilutions of serially diluted frontal lobe DNAs from three N/MC individuals: CE161, CE116, and 10-12 (Table 2). These nested PCRs required significantly higher frontal lobe DNA concentrations than those for HAD patients we studied previously (Fig. 1) (60, 61). A further 4 of the 12 individuals (CE125, CB183, CE128, and CE104) yielded 1 to 3 env clones from frontal lobe tissue, while no env products were obtained from the remaining 5 individuals (CE148, 5057, 6052, 6771, and 8276) following several PCR attempts with brain tissue DNA. These observations are consistent with previous studies showing that proviral loads are lower in brain tissue of N/MC individuals than in that of individuals with neuro-AIDS (22, 23).

TABLE 1.

Characteristics of N/MC patients investigated^f

Patient no.	Yr of collection	Brain bank	Neurological disease statusb	Duration of infection (yr)	CD4 count (cells/mm3)b	Viral load		Therapy
						Plasmab	CSFa
CE161	2004	UCSD	Normalc (192)	13	11 (192)	246,000 (192)	2,850	3TC, D4T, KTA, RTV, TFV (192)
CE116	2004	UCSD	MCMD (63)	13	80 (63)	342,386 (63)	856,342	D4T, DDI, NFV (63)
CB183	2005	UCSD	Normal (191)	16	88 (21)	74,294 (21)	27,400	DDI, TFV (21)
CE148	2004	UCSD	Normal (425)	2d	37 (425)	6,026 (425)	87,517	3TC, D4T, NFV (425)
CE125	2005	UCSD	Normal (0)e	10	218 (1,230)	NT	NT	3TC, IDV, D4T
CE128	2005	UCSD	Normal (423)	21	10 (752)	40,273 (752)	587	3TC (96)
CE104	2005	UCSD	Normal (0)e	2	446 (124)	33 (124)	11	3TC, CBV, IDV (124)
5057	2007	UCLA	Normal (253)	12	663 (148)	29,282 (−1)	NT	3TC, ABC, ATV, DDI, KTA, RTV, TFV, ZDV (253)
6052	2004	UCLA	Normal (24)	14	13 (24)	75,000 (24)	NT	3TC, ABC, D4T, EFV, IDV, KTA, NVP, SQV, ZDV (24)
6771	2004	Texas	Normal (310)	10	23 (354)	75,000 (37)	2,355	ABC, APV, DDC, EFV, FTV, KTA, SQV, T20, TFV (310)
8276	2005	Texas	Normal (146)	6	3 (182)	750,000 (184)	1,101	3TC, ABC, ADV, APV, D4T, DDC, DDI, EFV, IDV, KTA, NVP, RTV, SQV, ZDV
10-12	2012	NDRI	PE	NK	NT	NT	NT	3TC, AZT
CA110	1999	UCSD	Neuro-AIDS (0)e	21	21 (60)	198,957 (60)	2,268	No use reported
7766	1999	Texas	Neuro-AIDS (0)e	12	43 (21)	1,843 (20)	NT	3TC, ABC, EFV (22)
6568	2001	Texas	Neuro-AIDS (163)	17	77 (79)	>750,000 (402)	NT	EFV (163)
10017	1999	Mt. Sinai	Neuro-AIDS (191)	9	7 (319)	389,120 (381)	NT	3TC, D4T, SQV, ZDV

^aCSF viral load measurements were taken postmortem.

^bValues in parentheses represent numbers of days before death that assessments or measurements were made.

^cNormal neurocognitive diagnoses.

^dEstimated.

^eDiagnosis was made from patient information after death.

^fAZT, zidovudine; TFV, tenofovir; LPV, lopinavir; RTV, ritonavir; KTA, lopinavir-ritonavir (Kaletra); 3TC, lamivudine; CBV, AZT-lamivudine; IDV, indinavir; ABC, abacavir; ATV, atazanavir; DDI, didanosine; NVP, nevirapine; APV, amprenavir; DDC, dideoxycytidine; EFV, efavirenz; FTV, saquinavir (Fortovase); T20, enfuvirtide; FTC, emtricitabine; MCMD, minor cognitive motor disorder; PE, psychotic episodes; NT, not tested; NK, not known; UCSD, University of California, San Diego; UCLA, University of California, Los Angeles; NDRI, National Disease Research Interchange.

TABLE 2.

Functional env clones derived from brain and immune tissues^a

Patient no.	No. of env clones from proviral DNA		No. of env clones from viral RNA
	Brain	Immune tissue	Brain	Plasma
CE161	10	23	2	10
CE116	12	18	1	NT
CB183	2	15	5	NT
CE148	0	7	NT	NT
CE125	2	10	2	NT
CE128	1	15	1	NT
CE104	2	0	0	NT
5057	0	14	0	NT
6052	0	8	NT	NT
6771	0	9	0	NT
8276	0	12	1	NT
10-12	14	16	0	NT

^aNT, not tested.

FIG 1.

Nested PCR amplification of HIV-1 env genes from proviral DNAs in the brain and spleen. (A) Concentrations of tissue DNA required to PCR amplify env genes at endpoint dilutions. (B) Larger amounts of both brain and immune tissue DNAs were required to amplify rev-env sequences from N/MC individuals than from HAD patients, indicating lower proviral loads in these tissues.

Endpoint dilutions of spleen DNA enabled multiple env clones to be produced from 11 of 12 spleen samples from N/MC individuals. These PCRs required only slightly more DNA than that in our previous study using HAD spleen or LN tissue (25), although this difference was significant (Fig. 1). env clones were also amplified from bone marrow DNA of patient 10-12 and from CE161 plasma RNA (taken 6 months before death), by use of reverse transcription-PCR (RT-PCR). Full rev-env sequences were cloned from PCR products at limiting dilution endpoints as described in Materials and Methods.

HIV-1 RNA expression in brain tissue of N/MC individuals.

The nested PCR approach used to amplify env sequences from proviral DNA will yield sequences from latent proviruses as well as actively replicating HIV. It is therefore possible that env sequences amplified from proviral DNA in brain tissue of N/MC individuals are derived from archived HIV that is latent. To evaluate more specifically whether brain tissue of N/MC individuals contains actively expressing proviruses, we used RT-PCR to amplify env sequences from HIV RNA present in brain tissue. This approach will amplify HIV sequences from either virion RNA or mRNA expressed in cells and will represent current proviral expression.

Using RT-PCR, products were obtained from frontal lobe RNA by use of several different primer pairs. However, amplification was inefficient and required high concentrations of tissue RNA, indicating that expression was at a low level. Nevertheless, several full-length rev-env sequences were amplified from RNAs from individuals CE161, CE125, CE116, CB183, CE128, and 8276, while smaller V1-V5, V3, or gp41 DNA fragments were also obtained from these individuals as well as from patients 10-12 and 6771 (Table 3). No products were obtained from individuals CE104 and 5057. There was insufficient tissue from the remaining individuals to undertake RT-PCR. Full rev-env sequences from RNAs from individuals CE161, CE125, CE116, CB183, and 8276 were cloned as described above.

TABLE 3.

RT-PCR results for env sequences from frontal lobe tissue RNA

Patient status	Patient no.	RT-PCR result
		gp160	V1-V5	gp41	V3
N/MC	CE161	+	+	+	+
	10-12	−	+	+	+
	CE116	+	+	+	+
	CE125	+	−	+	+
	CB183	+	+	+	+
	8276	+	+	+	+
	6771	−	+	+	−
	5057	−	−	−	−
	CE104	−	−	−	−
Neuro-AIDS	7766	+	+	+	+
	6568	+	+	+	+

Phylogenetic analyses of env genes amplified from both brain and immune tissues of N/MC individuals.

Maximum likelihood phylogenetic trees for full-length env genes were calculated using MEGA, version 7. Phylogenetic trees for individuals CE161, CE116, 10-12, CE125, CE128, and CB183 are shown in Fig. 2. These individuals gave rise to env sequences derived from brain tissue as well as spleen tissue. For patient CE161, several brain env sequences formed a distinct cluster segregated from spleen-derived env sequences. A similar cluster of brain-derived env sequences was also observed for patient CE116, as well as a single divergent brain env sequence. An additional, clearly compartmentalized group of env sequences from spleen tissue was also observed for patient CE116. However, these corresponded to CXCR4-using variants (see below). The env sequences from patient 10-12 formed two predominant groups, although both contained env sequences derived from spleen, bone marrow, and brain tissue. The small number of env sequences amplified from proviral DNAs from brain tissue of patients CE125, CE128, and CB183 mainly segregated with spleen-derived env sequences (Fig. 2).

FIG 2.

Phylogenetic analysis (MEGA 7) of HIV-1 envelope nucleotide sequences amplified from brain and immune tissues. env sequences were amplified by nested PCR from endpoint-diluted DNAs extracted from brain tissue (green spots) or immune tissue (red spots) and from RNAs extracted from brain tissue (yellow spots). Mac-tropic Envs are boxed in yellow. Numbers at branch points represent bootstrap values. env sequences that contained stop codons were omitted. Sequences for CXCR4-using Envs are marked with an asterisk.

Several full-length env sequences recovered by RT-PCR from frontal lobe RNAs from patients CE161, CB128, and CB183 segregated with the majority of proviral env sequences from the spleen and (for CE161) with plasma-derived env sequences. A single env sequence from CE116 frontal lobe RNA clustered with several env sequences from brain proviral DNA, while a single env sequence from CB183 segregated alone. Finally, two env sequences from CE125 brain RNA clustered with a single env sequence from spleen DNA. Together, these data highlight viral RNA expression in the brain tissue of several N/MC individuals consistent with low-level, ongoing replication. The close relationship of most RNA-derived env sequences to those derived from proviral DNA suggests a close relationship of proviruses with currently active viruses.

Coreceptor use.

Coreceptor use was first evaluated using the Web-based PSSM program (http://indra.mullins.microbiol.washington.edu/webpssm/). The majority of Envs were predicted to be R5 by this method, including all Envs from brain tissue. Envs predicted to use CXCR4 included compartmentalized groups of spleen Envs from CE116 (discussed above), CE125, 8276, and CE148 (Fig. 2; see Fig. 4A). Pseudovirions carrying Envs predicted to use CXCR4 mediated infection of HeLa HIJ cells, which express CD4 and CXCR4 but not CCR5 (67), consistent with CXCR4 use (data not shown). Coreceptor use for each Env is incorporated into the phylogenetic trees in Fig. 2 and 4A, with CXCR4-using Envs each marked with an asterisk.

FIG 4.

Evolutionary relationships and tropisms of env sequences amplified from immune tissue of N/MC individuals who did not yield env sequences from the brain. Phylogenetic trees (maximum likelihood) (A) and macrophage infectivities of Envs (B) are shown for N/MC individuals who yielded few or no env sequences from the brain and for patient CE104, who yielded just two env sequences, only from brain tissue.

Infectivity of Envs for primary macrophages and HeLa TZM-bl cells.

We next tested whether Envs from N/MC brain tissue were macrophage tropic by preparing Env⁺ pseudoviruses, measuring the infectivity of Env⁺ pseudoviruses for primary macrophages, and comparing it with the infectivity for HeLa TZM-bl cells (Fig. 3 and 4B). HeLa TZM-bl cells express high levels of CD4 and CCR5 and are universally sensitive to CCR5-using HIV-1.

FIG 3.

Macrophage-tropic R5 Envs were detected in N/MC brain tissue. A subset of Envs were derived from brain tissue-infected primary macrophage cultures. In contrast, envelopes from immune tissue were mainly non-mac-tropic. Envs boxed in green were derived from brain RNA. Data shown are for individuals whose brain tissue yielded env sequences. Infectivity is presented as the number of FFU per milliliter and was derived from titers averaged from at least two assays using macrophages prepared from different donors. Macrophage infectivity data for other individuals are presented in Fig. 4.

Envs derived from proviral DNAs from both spleen and brain tissue yielded pseudoviruses that mediated a range of infectivities for HeLa TZM-bl cells, with many having similar or higher titers than those of control Envs, which included the non-mac-tropic R5 Envs JR-CSF and LN8 and the highly mac-tropic Envs JR-FL and B59 (60, 61, 68) (Fig. 3).

Envs derived from the spleen mainly conferred undetectable infection on macrophages, although there were several exceptions, including a single Env from the spleen of patient CE125 and a minority of Envs from patients CB183 and 6772 that mediated low to modest macrophage infection.

All Envs from 10-12 brain tissue DNA were also tightly non-mac-tropic, as were most from CE161 and CE116 DNAs. However, two of the CE161 brain tissue Envs mediated high levels of macrophage infection, while two of the CE116 brain tissue Envs mediated modest to high macrophage infection levels. Finally, the small number of Envs derived from brain tissue proviral DNAs from a further 5 individuals were non-mac-tropic.

The tropism of Envs from brain tissue RNA was intriguing. Two functional Envs from CE125 brain tissue were mac tropic (Fig. 3), and these clustered with a sole mac-tropic Env from the spleen (Fig. 2; note that mac-tropic Envs are boxed and shaded yellow), forming a compartmentalized group. A single full-length env sequence was recovered from CE116 brain RNA and clustered with a group of env sequences from brain proviral DNA. This group mediated different levels of macrophage infectivity. In contrast, non-mac-tropic env sequences recovered from brain RNAs from CE161, CB128, and CB183 clustered phylogenetically with other non-mac-tropic env sequences from brain and spleen tissue.

Although they are limited, these observations suggest that mac-tropic variants are present as actively expressing viruses in brain tissue of at least some N/MC individuals. There is also evidence of active, non-mac-tropic variants in brain tissue of individuals. Overall, these data contrast greatly with our experience with neuro-AIDS patients, in whom highly mac-tropic Envs are consistently detected and form the majority of Envs present in brain tissue (25, 60, 61). Nevertheless, mac-tropic Envs detected in four N/MC patients here were segregated into compartmentalized groups (Fig. 2) similar to those in most HAD/neuro-AIDS patients, as previously reported (25).

env sequences amplified from immune but not brain tissue of N/MC individuals.

Five N/MC individuals readily yielded env sequences from immune tissue DNAs (Fig. 4A), while no env sequences could be amplified from frontal lobe DNAs despite several attempts, although a single CXCR4-using env sequence was amplified from patient 8276 RNA. Nearly all immune tissue Envs from these patients conferred background or low macrophage infection (Fig. 4B). Finally, patient CE104 yielded only two env sequences from brain tissue and none from immune tissue. These two brain Envs did not confer macrophage infection (Fig. 4B).

sCD4 inhibition of mac-tropic and non-mac-tropic N/MC Env⁺ pseudoviruses.

In previous studies on HAD Envs, a strong correlation between R5 Env sensitivity to soluble CD4 (sCD4) inhibition and macrophage tropism was observed (69, 70). Thus, highly mac-tropic Envs from neuro-AIDS brain tissue were also highly sensitive to sCD4 inhibition. This is consistent with a high Env-CD4 affinity that is required for efficient entry into macrophages that express low levels of CD4 (71–73). In contrast, non-mac-tropic Envs from immune tissue are generally resistant. We investigated whether the mac-tropic Envs derived from N/MC brain tissue were sensitive to sCD4 compared to non-mac-tropic Envs from the same individuals. We wanted to establish whether sCD4 sensitivity to mac-tropic R5 Envs applied to mac-tropic brain Envs from individuals without severe neuro-AIDS. In addition, confirmation of enhanced sCD4 sensitivity would be additional support for a mac-tropic phenotype, while reduced sensitivity compared to that of Envs from HAD could indicate a less fulminant mac-tropic phenotype.

We tested Envs from the four patients who carried mac-tropic Envs in brain tissue: CE161, CE116, CE125, and CB183 (Fig. 5). Mac-tropic Envs (thick-bordered symbols) were compared to non-mac-tropic Envs (thin-bordered symbols) derived from both brain and immune tissues. The two mac-tropic CE161 Envs from the frontal lobe were highly sensitivity to sCD4, while all the non-mac-tropic Envs from brain or immune tissue were more resistant. For CE116, the most mac-tropic frontal lobe Env was sensitive to sCD4, while less mac-tropic and non-mac-tropic Envs from the brain and spleen were more resistant. Similarly, for patient CE125, the three mac-tropic Envs (two from the brain and one from the spleen) were sensitive to sCD4, while non-mac-tropic Envs from both the brain and spleen were resistant. Only one partially mac-tropic Env was amplified from patient CB183. This Env was the most sensitive to sCD4 compared to other Envs from brain and spleen tissue, though it was only slightly more sensitive than the other CB183 Envs. These observations indicate that mac-tropic Envs from N/MC individuals show enhanced sensitivity to sCD4, as previously noted for those from neuro-AIDS brain tissue.

FIG 5.

Soluble CD4 inhibition of brain and immune tissue Envs derived from N/MC patients CE161, CE116, CE125, and CB183. Macrophage-tropic Envs derived from brain tissue were generally more sensitive to sCD4 than non-mac-tropic R5 Envs from the same patients. Thick-bordered symbols represent mac-tropic Envs, and thin-bordered symbols represent non-mac-tropic Envs.

Envs from brain or spleen tissue of patient 10-12 were also investigated. A subset of 10-12 Envs from the brain, bone marrow, and spleen were sensitive to sCD4 even though they were non-mac-tropic (Fig. 6A). Phylogenetically, these sensitive Envs segregated separately (cluster 1) from Envs that were sCD4 resistant (cluster 2) (Fig. 6B). An enhanced sCD4 sensitivity for this cluster of 10-12 Envs probably indicates an increased Env-CD4 affinity (as it does for mac-tropic Envs). However, further investigation is required to understand why this does not result in increased macrophage infection via the low cell surface CD4 on these cells.

FIG 6.

Soluble CD4 sensitivity of 10-12 Envs. Soluble CD4-sensitive and -resistant Envs (A) clustered into two groups phylogenetically (B). BM, bone marrow.

Env determinants previously associated with macrophage tropism, brain infection, or neuro-disease.

Several Env determinants have been associated with macrophage tropism, brain infection, and/or HAD. These include an asparagine at residue 283, a contact residue for CD4 (74, 75), as well as the loss of a glycan at N386 (at the N terminus of V4) (74, 76). A substitution in a conserved acidic residue in the V1 loop was also reported to confer macrophage tropism for an Env derived from the plasma of a pediatric HAD patient (77). Others have reported the presence of determinants in the V1-V2 loops (74, 78). However, many highly mac-tropic Envs reported for the brain (25) or CSF (55) do not carry any of the reported determinants, and it is clear that other, unknown determinants are involved. Here we focused on individuals that yielded several mac-tropic brain Envs (CE161, CE116, and CE125) as well as the two clusters of Envs from patient 10-12. None of the mac-tropic Envs amplified from brain tissue or Envs from other tissues carried N283, and most carried a glutamic acid at V1 residue 153 (Table 4). A proportion of mac-tropic Envs from patient CE116 brain tissue had lost the glycan at residue 386. However, brain and mac-tropic Envs from other individuals retained N386 (Table 4). Elucidating the determinants of macrophage tropism for the Envs from the N/MC individuals studied here will thus require further investigation.

TABLE 4.

Env determinants of macrophage tropism^a

Patient no.	Tissue	Virus phenotype	Presence of residue (%) (other residue present)
			N283	N386	N362	G153
CE161	Immune tissue	Non-mac-tropic Env	0	100	100	0 (M)
	Brain	Mac-tropic Env	0	100	100	0 (M)
CE116	Immune tissue	Non-mac-tropic Env	0	100	100	0 (E)
	Brain	Mac-tropic Env	0	20	100	0 (E)
CE125	Immune tissue	Non-mac-tropic Env	0	100	0	0 (E)
	Brain	Mac-tropic Env	0	100	0	0 (E)
10-12	Cluster 1	sCD4 resistant	0	100	100	0 (E)
	Cluster 2	sCD4 sensitive	0	0	100	0 (E)

^aV1-V5 sequences selected for analysis from patients CE161, CE116, and CE125 were compartmentalized, mac-tropic Envs in the brain and compartmentalized, non-mac-tropic Envs in immune tissue (spleen). For patient 10-12, compartmentalized sCD4-resistant and sCD4-sensitive clusters were investigated.

Finally, the two clusters of Envs from patient 10-12 were also evaluated. The sCD4-sensitive cluster (cluster 2) (Fig. 6) had significantly fewer N-linked glycosylation sites (NLGSs) and slightly fewer hydrophobic residues than those of the sCD4-resistant Envs.

Do Envs from brain tissue of N/MC individuals carry distinct characteristics?

Previously, we reported that brain Envs from 4 of 5 HAD patients studied carried a lower overall positive charge (V1-V5 region) than that of Envs from immune tissue (25). Note that there were insufficient Envs derived from the fifth patient in that study to evaluate the charge differences rigorously. Here we examined the 3 N/MC patients (CE161, CE116, and 10-12) from whom a significant number of env sequences were amplified from both the brain and spleen. We compared these three patients with three neuro-AIDS patients we studied previously, for whom there was a clear compartmentalization of genotypes and macrophage tropism between brain and immune tissues. We investigated the overall positive charge, number of potential NLGSs (PNGSs), amino acid length, hydrophilicity, and hydrophobicity of the V1-V5 region to try to identify properties specific for Envs in brain tissue (Fig. 7).

FIG 7.

Characteristics of the V1-V5 region of gp120 for patients CE161, CE116, and 10-12. Overall positive charge, length, number of N-linked glycosylation sites (PNGSs), hydrophilicity, and hydrophobicity were estimated for the V1-V5 region of compartmentalized mac-tropic and non-mac-tropic Envs in brain and immune tissues, respectively, for patients CE161 and CE116. The same parameters were also estimated for sCD4-resistant (cluster 1 [C1]) and sCD4-sensitive (cluster 2 [C2]) Envs derived from patient 10-12. These characteristics were also estimated for mac-tropic and non-mac-tropic R5 Envs from brain and immune tissues, respectively, of three neuro-AIDS patients that we described previously (25).

Of the two N/MC patients with compartmentalized brain Envs, CE161 (but not CE116) carried brain V1-V5 sequences that had a highly significantly lower overall positive charge than that of Envs from the spleen. In contrast, all three neuro-AIDS patients carried compartmentalized Envs in the brain with highly significantly lower V1-V5 positive charges, as we reported previously (25). For CE116 and CE161, there were also highly significant differences in V1-V5 hydrophilicity, and for CE116, fewer NLGSs for mac-tropic brain Envs. Less significant differences were also detected between brain and immune tissue Envs for CE161 and CE116, for V1-V5 length, number of NLGSs, and hydrophobicity.

Significant differences in number of NLGSs, length, hydrophilicity, and hydrophobicity were also detected between brain and immune tissue V1-V5 sequences for the three neuro-AIDS patients. However, unlike the V1-V5 charge, these differences did not form unified patterns for the three patients. The 10-12 Env cluster 2 that was sensitive to sCD4 had significantly fewer NLGSs, shorter V1-V5 regions, a lower charge, and a lower hydrophobicity but did not exhibit differences in hydrophilicity compared to the sCD4-resistant cluster 1 (Fig. 7).

Together, these data indicate that the compartmentalization of Envs observed in both N/MC and neuro-AIDS patients result in Envs with distinct characteristics. The implications of these observations for founder effects or the selection of distinct Env phenotypes in the brain or other tissues are addressed in Discussion.

DISCUSSION

Highly mac-tropic R5 HIV variants predominate in brain tissue (25, 60–62) and CSF (55) in individuals with neuro-AIDS, including HAD. However, it is unclear whether individuals without neuro-AIDS or with only minor symptoms carry HIV within brain tissue and whether any virus present is macrophage tropic or actively replicating.

In the present study, we investigated the sequences of HIV-1 Envs derived from immune and brain tissues of patients with normal neurology or only minor complications (N/MC individuals) before expressing Envs on pseudoviruses and evaluating their properties. We showed here that (i) env sequences were difficult to amplify from proviral DNA in brain tissue for most N/MC individuals, consistent with previous reports (22–24); (ii) fewer mac-tropic Envs were present in brain tissue of N/MC individuals than in that of patients with neuro-AIDS or HAD; (iii) mac-tropic R5 Envs in brain tissue formed phylogenetic clusters distinct from those of most non-mac-tropic Envs from the spleen or brain; (iv) there was evidence of active proviral expression in brain tissue for some individuals; and (v) Envs from immune tissue of the N/MC individuals were nearly all tightly non-mac-tropic, contrasting with our previous data for neuro-AIDS patients, for whom Envs derived from immune tissue mediated a range of macrophage infectivities, from background levels to modest infection, with a small number of Envs from some patients mediating high macrophage infection levels (25, 60, 61).

Overall, our data demonstrate the presence of mac-tropic variants in brain tissue well before serious neurological disease develops, consistent with the observations of Sturdevant et al., who reported the presence of compartmentalized, mac-tropic Envs in the CSF early in infection (20). The origin of the mac-tropic variants in brain tissue demonstrated here is unclear. They were divergent in sequence from non-mac-tropic Envs derived from either immune or brain tissue. Note that the mac-tropic brain Envs had genetic distances from the closest non-mac-tropic immune tissue Envs similar to those for Envs present in the neuro-AIDS brain (Fig. 8). The latter observation is not consistent with recent evolution of these mac-tropic Envs from non-mac-tropic Envs. It seems more likely that they are derived from a resident population of variants that are either in a latent form or replicating at a low level in the brain from an earlier stage of infection. Alternatively, they may be derived from a separate macrophage-tropic population of variants present elsewhere in the body. However, without an analysis of variants amplified from brain and immune tissues at different times, it is not possible to obtain an accurate assessment of evolutionary time scales or to evaluate these different possibilities with any confidence.

FIG 8.

Macrophage-tropic N/MC and neuro-AIDS envelopes from the brain have similar genetic distances from the nearest non-mac-tropic Envs in immune tissue. (A) Distances from mac-tropic Envs to the nearest non-mac-tropic immune tissue Env. (B) Distances from the nearest mac-tropic brain Env to the nearest non-mac-tropic immune tissue Env.

For patients CE116 and CE161, compartmentalized groups that contained mac-tropic R5 Envs also included other Envs which were less macrophage tropic as well as several that were nonfunctional. It is tempting to speculate that the mac-tropic variants evolved from the closely related but less mac-tropic ones. However, it is also possible that less functional and less mac-tropic Envs arose in response to evolutionary pressures in the brain, perhaps exerted by the low level of immunoglobulin present there (79–82).

The majority of N/MC patients were on ART in the last year of their lives. However, most patients also had significant viral loads in plasma and (where tested) CSF (Table 1), indicating that these patients may represent virological failures or that they had drug adherence issues. Three of the four patients who carried compartmentalized mac-tropic variants in the brain had significant viral loads in both plasma and CSF. Viral replication at one or both of these sites may therefore represent an increased risk for mac-tropic variants emerging in brain tissue.

It is worth considering that none of the brain tissue samples taken were derived from patients who had been perfused after death. The identification of mac-tropic Env variants in brain tissue that were genetically distinct from non-mac-tropic Envs present in immune tissue (and, where tested, plasma) strongly supports brain macrophages as their origin. However, it is likely that non-mac-tropic Envs from the brain that cluster genetically with those from immune tissue are derived from HIV in infected CD4⁺ T cells present in blood capillaries rather than from T cells or other cell types present in the brain parenchyma.

In a previous study of neuro-AIDS patients, we reported that the V1-V5 amino acid sequences of mac-tropic brain Envs carried a lower positive charge than that of Envs from immune tissue. This lower Env charge may carry an advantage for infection of brain tissue. “Low-charge Envs” are present on viruses replicating early in infection, before Envs with a higher charge (associated with enhanced Env-CCR5 interactions) evolve later in disease and target T cells with lower CCR5 levels (63–66, 83). It is possible that variants with low-charge Envs colonized the brain early in infection, persisting there and evolving enhanced macrophage tropism and neurovirulence. However, only N/MC patient CE161 carried brain Envs with a significantly lower charge than that for Envs from immune tissue, and it is possible that low Env charge is associated with the development of neurovirulence. Differences in V1-V5 length, number of NLGSs, hydrophilicity, and hydrophobicity were detected between brain and immune tissue Envs for both N/MC and neuro-AIDS individuals. The extent that these differences in Env characteristics reflect different selective pressures in immune tissue compared to the brain is unclear. However, the inconsistent spread of the characteristics studied across both N/MC and HAD patients may be more consistent with the influence of founder effects. Nevertheless, these data support the presence of viruses with very distinct characteristics and properties (including macrophage tropism) in the brain and immune tissues well before neuro-AIDS develops.

The substantial number of non-mac-tropic Envs amplified from brain tissue of three patients (and several for others) is intriguing. Many segregate phylogenetically with non-mac-tropic Envs present in the spleen, bone marrow, or plasma. As mentioned above, it is therefore possible that they are derived from HIV in infected CD4⁺ T cells present in blood capillaries rather than from T cells that have penetrated the brain parenchyma. However, our observations may support an intriguing possibility in which HIV is carried into brain tissue by T cells, which have been shown to circulate through brain tissue at a low level (56). These T cells may exit the brain via lymphatic vessels which serve the meninges (57, 58). It is also worth noting that non-mac-tropic R5 Envs have been detected in CSF of individuals early in infection (20) as well as in late disease, following the onset of neuro-AIDS (55). These T-tropic R5 Envs in CSF were frequently clonally expanded and were associated with pleocytosis of T cells into the CSF. However, it has not yet been demonstrated whether T cells present in CSF then traffic into brain tissue itself or can deliver HIV for colonization of macrophages and microglial cells present there.

While it seems likely that non-mac-tropic R5 viruses amplified from brain tissue are derived from infected T cells, this was not definitively shown. For patient 10-12, we enriched CD11b⁺ macrophage/microglial cells from disaggregated brain tissue by using magnetic microbeads (Miltenyi Biotec Inc.). The enriched macrophage/microglial cells and flowthrough cells were then tested for the presence of Envs by PCR. Only the macrophage/microglia-depleted cell fraction was positive (not shown). However, whether T cells or other cell types in this fraction were infected requires further study using fresh tissue from new individuals.

In situ hybridization for detection of HIV RNA and immunohistochemistry were also used to try to identify HIV⁺ cells in brain tissue sections. For non-mac-tropic variants, resident astrocytes could be targets, as well as T cells that are restricted to blood capillaries or that are present in brain tissue itself. For the N/MC individuals studied, HIV⁺ cells were not observed (data not shown). However, this approach (which targeted the gag gene) may not be sensitive enough to reliably identify productively infected cells if they are rare, while it will not detect latently infected cells or infected astrocytes, which produce mRNAs mainly for early HIV proteins (84, 85). Regardless, this approach did not reveal any unusual increase in the numbers of T cells apparent in the brain parenchyma or in blood capillaries, although it can be considered only a very limited survey of infected brain tissue from each individual. The identification of the target cells for non-mac-tropic variants in brain tissue and their specific locations requires further investigation.

Previously, we reported a significant correlation between macrophage tropism and sensitivity to sCD4 inhibition. Highly mac-tropic R5 Envs from brain tissue were uniformly sensitive to sCD4 inhibition, whereas non-mac-tropic Envs were more resistant (69, 70). The increased sensitivity to sCD4 may reflect a more exposed CD4 binding site and/or an increased capacity to undergo sCD4-induced conformational changes that inactivate Envs (69). In the present study, all the mac-tropic Envs from brain tissue of patients CE116, CE161, CE125, and CB183 were more sensitive to sCD4, contrasting with the non-mac-tropic Envs from the same individuals, which were more resistant. Surprisingly, several Envs from patient 10-12 were sensitive to sCD4 despite being non-mac-tropic. All the sCD4-sensitive Envs clustered together phylogenetically (Fig. 6B) and included Envs derived from the brain as well as from both the spleen and bone marrow. The origin of these Envs and the selective pressures in vivo that resulted in their emergence are unclear. Arrildt et al. also recently described R5 Envs that were not macrophage tropic yet were sensitive to sCD4 inhibition (86). They argued that these Envs were intermediates in the evolution of more mac-tropic forms. It is unclear whether the same is true for the sCD4-sensitive 10-12 Envs, which were present inside and outside the brain but in the absence of any detectable mac-tropic Envs.

In summary, our data provide new insights into HIV replication and the evolution of mac-tropic Envs in brain tissue of patients before the onset of serious neurological complications. The detection of mac-tropic R5 variants in the brain is a strong indication of the presence of an infected population of brain macrophages in the absence of neuro-disease that might represent an increased risk for future development of neurological disease. It will be important in future studies to compare the presence and relationship of mac-tropic Envs in brain tissue to those present in CSF, a site accessible to monitoring during the life of patients. Finally, the presence of an HIV-infected macrophage population in brain tissue will be a challenge for the development of strategies to completely eradicate HIV from the body and to effect a cure.

MATERIALS AND METHODS

HIV-1⁺ patients.

Tissue samples from all individuals except for patient 10-12 were provided by the National NeuroAIDS Tissue Consortium (NNTC) (Table 1). The majority of patients were classified as having AIDS, with the exception of CE104, 5057, and CE125. The NNTC individuals were diagnosed for neurological issues by use of a battery of neuropsychological tests. Normal neurology was designated if the patient had no significant cognitive complaints, no evidence of impairment on neuropsychological testing, and/or no loss of functional capacity. Patient CE116 was diagnosed as having minor cognitive disorder (MCMD) with evidence of minor cognitive decline. Brain frontal lobe and spleen samples were obtained postmortem and kept frozen at −80°C. Unfrozen tissue from patient 10-12 was provided on ice within 48 h of death by the National Disease Research Interchange (NDRI). This individual had no diagnosed neurocognitive issues but had not undergone extensive neuropsychological testing. Table 1 includes information (where available) on ART, noting the last record of therapy before death. The patient tissues used in this study were provided without identifiers. The study of these tissues (described here) was therefore not considered Human Subjects Research by the University of Massachusetts Medical School Institutional Review Board.

Nucleic acid extraction and PCR amplification of env sequences from single-molecule templates.

Total DNA was extracted and purified from the brain and spleen (and, for patient 10-12, bone marrow) by use of a QIAamp DNA minikit (Qiagen Inc.). Total RNA was purified from plasma (patient CE161) by use of a High Pure viral RNA kit (Roche) and from brain tissue by use of TRizol and a Qiagen RNeasy kit. RNase-free DNase (Qiagen Inc.) was also included in order to eliminate any DNA contamination. DNA and RNA were eluted in nuclease-free water and stored at −80°C until analysis.

Limiting dilution PCRs were carried out on tissue DNA or RNA (RT-PCR) to amplify a rev-env product, which includes the entire env gene. rev-env products were cloned from PCR dilutions for which one-third or less of PCRs were positive, i.e., with a >80% chance of being derived from single genomes (87). RT-PCRs were performed using the SuperScript III One-Step RT-PCR system with PlatinumTaq High Fidelity polymerase (Invitrogen Inc.), and Phusion DNA polymerase (Finnzymes Inc.) was used for PCRs. RT-PCRs and PCRs were set up to amplify an approximately 3-kb rev-env fragment by using the following primers. Outer primers were SG3-up (5′-TACAGTGCAGGGGAAAGAATAATAGACATAATA-3′) (88), SG3-lo (5′-AGACCCAGTACAGGCRARAAGC-3′) (88), RevenvA (5′-TAGAGCCCTGGAAGCATCCAGGAAG-3′), and EnvN (5′-CTGCCAATCAGGGAAGTAGCCTTGTGT-3′) for RT-PCRs and RevenvA and EnvN for PCRs. Inner primers were RevenvBTOPO (5′-CACCTAGGCATCTCCTATGGCAGGAAGAAG-3′) and Env-lo (5′-GTTTCTTCCAGTCCCCCCTTTTCTTTTAAAAAG-3′) (88). Primers to amplify smaller env fragments (V1-V5) included ED5 (5′-ATGGGATCAAAGCCTAAAGCCATGTG-3′) and 578 (5′-TGGTCCCTCATATCTCCTCCTCCA-3′). PCR products for cloning and sequencing were purified from a 0.8% crystal violet-stained agarose gel by use of a QIAquick gel extraction kit (Qiagen Inc.).

Primers to amplify smaller env fragments from RNA by RT-PCR were as follows: for V3, JA9 (5′-CACAGTACAATGTACACATG-3′) and JA12 (5′-ACAGTAGAAAAATTCCCCTC-3′) (89); and for GP41, GP40F1 (5′-TCTTAGGAGCAGCAGGAAGCACTATGGG-3′) and GP41R1 (5′-AACGACAAAGGTGAGTATCCCTGCCTAA-3′) (90).

Envelope cloning, sequencing, and analyses.

Purified PCR products were cloned into pcDNA 3.1D/V5-His-TOPO (pcDNATM3.1 directional TOPO expression kit; Invitrogen Inc.) before transformation into competent Escherichia coli cells (TOP10; Invitrogen Inc.). Colonies were screened for correct rev-env insertions by PCR, using a universal T7 promoter primer (5′-TAATACGACTCACTATAGGG-3′), primer M5-R (5′-CCAGCTGGGGCACAATAATGTATGGGAATTGG-3′, a reverse primer that hybridizes within the env insert) (91), and Go Taq Green master mix (Promega Inc.). Plasmid DNA was purified using a QIAprep miniprep kit (Qiagen Inc.).

The rev-env clones were sequenced by Genewiz Inc. and analyzed phylogenetically to establish the population diversity. Envelope sequences were analyzed for the presence of N283 and other determinants associated with macrophage tropism or brain infection (61, 74, 75, 77), as well as length, charge, hydrophobicity, hydrophilicity, and potential N-linked glycosylation sites (NLGSs) of the V1-V5 sequence. Sequences were also assessed for mutations that may inactivate Envs, including deletions, premature stop codons, loss of conserved cysteines involved in disulfide bonding, etc. Nonfunctional envelopes with stop codons or deletions were not included in the analyses, unless stated otherwise.

Phylogenetic analyses.

gp160 env gene nucleotide sequences were assembled and aligned using Clustal X (92), with manual adjustment. All positions with an alignment gap of one or more nucleotides were excluded from the analysis.

Phylogenetic and molecular evolutionary analyses were conducted using MEGA, version 7 (93). Maximum likelihood phylogenetic trees were generated using the most appropriate evolutionary model for each analysis (94). Bootstrap analyses on 1,000 replicates assessed the robustness of the tree. Significant (≥70%) bootstrap values were assigned to internal tree nodes. Reference sequences representing three HIV-1 group M subtype B (http://www.hiv.lanl.gov/) envelopes (FR.83.HXBc2.K03455, TH.90.BK132.AY173951, US.98.1058_11.AY331295, and NL.00.671_00T36.AY423387) were used as outgroups.

A phylogenetic tree was plotted that included representative sequences from all individuals studied. env sequences from each individual segregated separately, consistent with different origins and not with the possibility of laboratory contamination (not shown). In addition, each env sequence was tested with HIV BLAST (HIV databases) to check for relationships with existing env sequences. No matches were observed, again consistent with the lack of contamination with existing env clones.

Coreceptor use for Envs was calculated using Web PSSM (https://indra.mullins.microbiol.washington.edu/webpssm/).

Cell cultures.

Env⁺ pseudovirions were prepared in 293T cells by transfection. HeLa TZM-bl cells (95) were used to estimate Env⁺ pseudovirion infectivity titers. Env⁺ pseudovirion infectivity was also evaluated on CD4⁺ CXCR4⁺ CCR5⁻ HeLa HIJ cells to monitor CXCR4 use. HeLa TZM-bl cells express high levels of CD4, CCR5, and CXCR4 and contain HIV-inducible β-galactosidase and luciferase reporter genes. 293T cells, TZM-bl cells, and HIJ cells (67) were maintained in Dulbecco's modified Eagle's medium (DMEM; Gibco-Invitrogen, Carlsbad, CA) supplemented with 5% fetal bovine serum (FBS) (60, 74, 96).

Macrophage cultures were prepared from blood monocytes by adherence, using buffy coats provided by New York Biologics Inc., as described previously (60, 74, 96). Monocytes were cultured for 5 days in DMEM containing 10% AB⁺ human serum (HS) for differentiation before being set up for infection. On the day prior to infection, the macrophages were washed and resuspended in DMEM containing 10% HS and cultured in 48-well tissue culture plates (1.25 × 10⁵ cells/0.5 ml/well).

Authentication of cell lines.

HeLa TZM-bl cells were obtained from the NIH AIDS Reagent Program and stored frozen at an early passage. These cells are susceptible to R5 and X4 HIV-1 strains and express β-galactosidase and luciferase when infected by HIV. The early-passage cells used here exhibited these phenotypes, confirming their authenticity. 293T/17 cells were obtained from the ATCC and were kept frozen at an early passage. They were used to produce HIV-1 Env⁺ pseudovirus stocks following transfection of appropriate plasmid DNA constructs. The consistently high-infectivity-titer Env⁺ pseudoviruses produced from them exhibited the predicted Env tropisms and properties and verified their authenticity. Finally, HeLa-CD4 HIJ cells were obtained from Navid Madani in 2003 and stored frozen at an early passage. These HeLa cells express CD4 and CXCR4 but not CCR5. Their sensitivity to HIV-1 X4 viruses but resistance to R5 viruses was routinely tested to confirm their authenticity.

Production of Env⁺ pseudovirions and infectivity assays.

Env⁺ pseudovirions were prepared by cotransfection of the Env⁺ pTOPOenv vector with an Env⁻ pNL4.3Δenv construct that carried a premature stop codon in env (60) into 293T cells by use of a calcium phosphate kit (Profection; Promega Inc.). Cell-free supernatants were harvested after 48 h of culture, and aliquots were frozen at −152°C prior to analysis.

Env⁺ pseudovirions were titrated on HeLa TZM-bl cells, HeLa HIJ cells, and macrophages. For HeLa TZM-bl and HeLa HIJ cells, 2 × 10⁴ cells/0.5 ml were added to each well of 48-well plates the day prior to virus titration. Virus titers were determined as described previously (60). Briefly, 100-μl aliquots of serially diluted viral supernatants in DMEM (with 5% FBS) were added to cells and incubated for 3 h. Aliquots of 0.4 ml of DMEM (with 5% FBS) were added, and cultures were incubated for 48 h (HeLa TZM-bl cells) or 72 h (HIJ cells). TZM-bl cells were then fixed in 0.5% glutaraldehyde in phosphate-buffered saline (PBS) and stained for β-galactosidase expression by use of X-Gal substrate (0.5 mg/ml X-Gal [5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside], 3 mM potassium ferricyanide, 3 mM potassium ferrocyanide, 1 mM magnesium chloride). HeLa HIJ cells were fixed in cold 1:1 methanol:acetone, washed, and immunostained for p24 by use of monoclonal antibodies 38:96K and EF7 (UK Centre for AIDS Research), followed by an anti-mouse IgG–β-galactosidase conjugate and X-Gal substrate (60).

Macrophages seeded in 48-well plates were pretreated with 0.1 ml DEAE-dextran (10 μg/ml) in DMEM containing 10% human plasma for 30 min at 37°C before virus supernatants were added and the plates inoculated by spinoculation for 45 min in a benchtop centrifuge (97). Infected macrophages were incubated for a further 3 h at 37°C before the addition of 0.4 ml of DMEM (10% AB⁺ male human serum) and incubation at 37°C for 7 days. Macrophages were then fixed and immunostained for p24 as described for HIJ cells. DEAE-dextran and spinoculation enhance virus infectivity by up to approximately 20-fold by increasing attachment (97) and entry (98). Infection following this procedure helps to maximize macrophage infection and to distinguish the most mac-tropic Env⁺ pseudoviruses. It does not bypass the requirement of CD4 and CCR5 for infection, which remains sensitive to entry inhibitors, including maraviroc (not shown). Env⁺ pseudovirions are capable of only a single round of replication, so the number of focus-forming units (FFU) was estimated by counting individual or small groups of blue-stained, infected cells by light microscopy. Average numbers of FFU per milliliter were then calculated. All values represent averages for at least two independent experiments, each done in duplicate and using macrophages from different donors. Error bars in figures were calculated for replicate wells from both experiments.

Each set of macrophage, TZM-bl, and HIJ infections included HIV-1 pseudovirions carrying the following control Envs: the mac-tropic R5 Envs JR-FL and NA20 B59 along with the non-mac-tropic Envs JR-CSF and LN8.

Inhibition and neutralization assays.

Inhibition assays with sCD4 were carried out as described previously, using HeLa TZM-bl cells as target cells (69, 70). Fifty-microliter samples of serially diluted sCD4 were mixed with 50 μl Env⁺ pseudovirions carrying 200 FFU at 37°C for 1 h and added to HeLa TZM-bl cells. Residual infectivity was evaluated by measuring the activity of the Tat-dependent β-galactosidase reporter gene stably integrated into TZM-bl cells (95). The medium was removed, and 100 μl of medium without phenol red was added. Cells were then fixed and solubilized by adding 100 μl of Beta-Glo (Promega Inc.). Luminescence was then read in a BioTek Clarity luminometer.

Statistical methods.

Significant differences between values for different Env groups were evaluated using two-tailed, nonparametric Mann-Whitney tests via Prism 7.

sCD4 concentrations that caused 50% inhibition (IC₅₀s) were estimated using Prism 6 for Mac OS X. When inhibition failed to reach 100%, IC₅₀ estimates were winsorized by defining them manually from graphs plotted in Prism. Two-tailed, nonparametric Mann-Whitney tests were used to evaluate whether statistically significant differences existed between distributions of sCD4 IC₅₀s for Env⁺ pseudovirions from the brain and spleen.

Accession number(s).

The nucleotide sequences of new envelopes reported here have been assigned GenBank accession numbers KX156365 to KX156580.