Baseline Susceptibility of Primary Human Immunodeficiency Virus Type 1 to Entry Inhibitors

Abstract

Human immunodeficiency virus type 1 plasma viruses from 29 entry inhibitor-naive patients were characterized for their susceptibilities to T-20, AMD3100, and RANTES. A strikingly wide range of susceptibilities to T-20 was observed that was influenced by coreceptor usage but not by the susceptibilities of the viruses to inhibitors that target the chemokine receptors or by polymorphisms in the gp41 N helix.

The human immunodeficiency virus type 1 (HIV-1) entry process (2, 19, 27) can be inhibited by several drugs (3, 14, 22, 23), which belong to three groups according to the step they inhibit: (i) inhibitors of the interaction between the viral surface glycoprotein (gp120) and CD4, which target the CD4-binding site on gp120; (ii) inhibitors of the interaction between gp120 and CCR5 or CXCR4 (e.g., chemokines and their derivatives or small organic molecules that antagonize chemokine receptor activity); and (iii) fusion inhibitors, which are peptides derived from the sequence of the viral transmembrane glycoprotein (gp41) that prevent the formation of a hairpin structure required for membrane fusion. One of these peptides, T-20 (enfuvirtide), is currently being evaluated in phase III clinical trials (15, 25, 26). The optimization of treatment strategies that include entry inhibitors will rely on the availability of methods capable of determining baseline viral susceptibility and acquired resistance to these drugs. In addition, the characterization of the determinants of baseline susceptibility and of acquired resistance to entry inhibitors may provide valuable information on the viral entry process and on the precise mechanism of action of these drugs.

We have developed a recombinant virus assay that permits the assessment of viral susceptibility to entry inhibitors. We modified a pNL4-3 molecular clone by deleting the region of the envelope gene encoding gp120 and the ectodomain of gp41 (positions 6480 to 8263) and replacing it with a linker that contains a unique MluI restriction site (vector 43-Δenv). Recombinant virus was produced by cotransfection of 293-T cells with the MluI-linearized 43-Δenv vector and a reverse transcription-PCR product, amplified from patient plasma samples, which encompasses the deleted region and carries short overlaps that allow homologous recombination. Virus-containing supernatants were used to infect subconfluent U373MG-CD4 cells expressing either CCR5 or CXCR4 (17), in the absence or in the presence of increasing concentrations of entry inhibitors. These target cells carry an HIV-1 long terminal repeat-lacZ cassette, which allows the quantification of single cycle infectivity by a colorimetric assay based on HIV-1 Tat-induced expression of β-galactosidase (24). The concentrations inhibiting 50% of virus infectivity (IC₅₀s) were calculated by using the median-effect equation (6).

The recombinant virus assay was first used to determine the baseline susceptibilities of subtype B primary viruses to the fusion inhibitor T-20 (American Peptide Company, Inc., Sunnyvale, Calif.). Plasma samples selected for the study were obtained from 29 entry inhibitor-naive patients; two patients (codes 12 and 17) were treatment naive, and the remainder had been treated with multiple reverse transcriptase and/or protease inhibitors. Samples from 23 patients harbored R5 exclusive viruses and were characterized by plasma viral loads ranging from 1,400 to 227,000 (median, 39,650) copies/ml. Patients are numbered in the order of decreasing susceptibility to T-20 (Table 1). The baseline susceptibility to T-20 for these viruses ranged from 3 to 1,002 ng/ml, with a median IC₅₀ of 159 (±55) ng/ml. The variability in the range of susceptibilities to T-20 was much wider than that measured for other antiretroviral agents (13). IC₅₀s for most patients were in the range of previously reported data for T-20-naive patients, obtained by virus culture on peripheral blood mononuclear cells (8, 9).

TABLE 1.

Baseline susceptibilities to T-20 of primary viruses

Viral tropism	Patient code or virus	CD4+ CCR5+ target cells		CD4+ CXCR4+ target cells
		IC50a (ng/ml)	SDb	IC50a (ng/ml)	SDb
R5	1	3	2.47
	2	23	2.41
	3	28	1.52
	4	30	2.41
	5	31	1.16
	6	32	1.32
	7	33	1.56
	8	35	2.25
	9	45	2.96
	10	45	1.25
	11	50	1.15
	12	95	1.24
	13	101	1.22
	14	103	2.84
	15	115	1.37
	16	122	1.64
	17	134	1.46
	18	193	1.4
	19	220	1.53
	20	286	1.77
	21	309	1.38
	22	616	1.26
	23	1,002	1.21
R5X4	24	2	1.84	90	1.62
	25	3	1.31	32	1.83
	26	28	1.11	356	1.39
	27	136	1.59	61	1.57
X4	NL4-3c			30	1.46
	HxB2c			2	1.17
	28			228	1.28
	29			408	2.24

^aIC₅₀s represent the geometric means of at least three independent experiments.

^bStandard deviations of the geometric means were calculated as the antilogarithms of the standard deviations of the logarithms of the IC₅₀s.

^cNL4-3 and HxB2 are reference T-cell-line-adapted viruses.

IC₅₀s were also calculated for plasma samples from four patients with dual-tropic (R5X4) virus populations and two patients who harbored X4 exclusive viruses (Table 1). Viral load for these patients ranged from 2,000 to 75,000 copies/ml. Overall, IC₅₀s for T-20 measured on CXCR4⁺ cells were within the range observed with CCR5⁺ cells, varying from 32 to 408 ng/ml. It should be emphasized, however, that susceptibility could be affected by different numbers of chemokine receptor molecules expressed on target cells. Because only a small number of X4 viruses were tested, these results do not resolve the currently debated issue of the differential susceptibilities of viruses with different tropisms to T-20 (8, 9). It is interesting, however, that viruses from patients 24, 25, and 26, which displayed a mixed R5 and X4 phenotype, were 10 times more susceptible to T-20 when tested on CCR5⁺ cells than when tested on CXCR4⁺ cells, whereas the T-20 susceptibility of virus from patient 27 was not significantly different when measured on these cell lines. These results indicate that, for a given virus population, the usage of alternative chemokine receptors can affect the efficacy of T-20 inhibition.

The N helix of the ectodomain of gp41 is the molecular target of T-20 and other peptides based on the sequence of the C helix of gp41 (4, 16). Accordingly, the development of resistance to T-20 both in vitro and in vivo is associated with selection of mutants carrying substitutions in the ₃₆GIV₃₈ tripeptide sequence and in other positions of the N helix, between residues 36 and 45 (20, 25; P. Sista, T. Melby, M. L. Greenberg, D. Davison, L. Jin, S. Mosier, M. Mink, E. Nelson, L. Fang, N. Cammack, M. Salgo, and T. J. Matthews, abstract from the XI International HIV Drug Resistance Workshop 2002, Antivir. Ther. 7[Suppl. 1]:S23, 2002). These mutations appear to reduce virus replication capacity and are generally not found in viruses in the absence of T-20 selective pressure (20, 25, 28; J. Lu, P. Sista, N. Cammack, and D. Kuritzkes, abstract from the XI International HIV Drug Resistance Workshop 2002, Antivir. Ther. 7[Suppl. 1]:S74, 2002; Sista et al., Antivir. Ther. 7:S23, 2002). To determine whether the wide range of susceptibilities to T-20 observed for the primary viruses analyzed here depended on polymorphism in this region, we sequenced the N-helix region from 16 plasma viruses characterized by different tropisms and by different levels of baseline susceptibility to T-20, including the most and the least sensitive viruses (Table 2). All samples carried the wild-type ₃₆GIV₃₈ tripeptide sequence. In several samples, polymorphisms were detected at positions of the N helix that have previously been shown to be subject to natural variation in T-20-naive patients (18). None of these polymorphisms was repeatedly found in samples characterized by low or high susceptibility to T-20. Our data suggest that the determinants for the natural susceptibility to T-20 are different from currently described determinants of acquired resistance and reside outside the gp41 N helix.

TABLE 2.

Sequence comparison of the N helix of the gp41 ectodomain (residues 30 to 79) from primary isolates characterized by different susceptibilities to T-20

Patient code or virus	Viral tropism	N-helix sequence
		3040506070
HxB2a	X4	ARQLLSGIVQ QQNNLLRAIE AQQHLLQLTV WGIKQLQARI LAVERYLKDQ
NL4-3a	X4	------D-------------------------------------------
1	R5	--L-------------K---------------------------------
4	R5	-----------------------------------------------Q--
6	R5	--S--------------------------------------------Q--
7	R5	-----------------------------------------------R--
8	R5	--S-----------------------------------------------
9	R5	------------S---K---------K-----------------------
13	R5	-----------------------------------------------Q--
15	R5	----------------K--------------------------L---Q--
19	R5	------------------------M-------------------------
21	R5	--------------------------------------------------
22	R5	------------S---M-------M--------------L----------
23	R5	----------------Q---------------------------------
25	R5X4	-----------------------------------------------R--
26	R5X4	--L---------S----------------------------------R--
27	R5X4	--L---------S-----------------------------L----R--
29	X4	--L--------------------------------------------R--

^aNL4-3 and HxB2 are reference T-cell-line-adapted viruses.

We reasoned that a virus characterized by efficient interaction with chemokine receptors could be expected to display reduced unmasking of the gp41 ectodomain and/or accelerated kinetics of the entry process, thereby reducing the time that the molecular target of T-20 is exposed. In the absence of a method to accurately measure the avidity of interactions between virus and chemokine receptors expressed on the cellular membrane, we used the susceptibilities to AMD3100 (a generous gift from E. De Clercq) and RANTES (R & D Systems, Minneapolis, Minn.) to estimate the efficiency of the interaction with the chemokine receptors. We measured the susceptibilities of the four R5X4 viruses and the two X4 viruses to AMD3100, which inhibits the interaction with CXCR4 (10, 21) (Table 3). Their baseline susceptibilities, expressed as IC₅₀s, ranged from 0.08 to 3.57 ng/ml. No correlation was found between susceptibility to AMD3100 and that to T-20 (R² = 0.2906, P = 0.2697). Similarly, we measured the susceptibilities of 11 of the 23 R5 viruses to RANTES, a natural ligand of CCR5 (1, 5, 7, 11, 12). Inhibition of HIV infection by RANTES is due both to competition for CCR5 and to down-modulation of CCR5 surface expression by endocytosis. To reduce the influence of CCR5 down-modulation, which is dependent on the concentration of RANTES, we compared the susceptibilities of different R5 viruses at a fixed concentration of RANTES. The percentage of inhibition of infection at 250 ng/ml ranged from 20 to 60.2% (Table 4). Again, there was no correlation between viral susceptibility to T-20 and that to RANTES (R² = 0.0154, P = 0.7164). These data suggest that the efficiency of interaction with chemokine receptors was not a major determinant of the natural susceptibility of viruses to T-20.

TABLE 3.

Baseline susceptibilities to AMD3100 of primary viruses on CD4⁺ CXCR4⁺ target cells

Patient code	Viral tropism	IC50a (ng/ml)	SDb
26	R5X4	0.08	1.83
25	R5X4	0.28	1.37
28	X4	0.32	1.42
27	R5X4	0.48	2.01
24	R5X4	0.96	1.84
29	X4	3.57	1.49

^aIC₅₀s represent the geometric means of at least three independent experiments.

^bStandard deviations of the geometric means were calculated as described in footnote b, Table 1.

TABLE 4.

Inhibition of infection by RANTES (250 ng/ml) of primary viruses on CD4⁺ CCR5⁺ target cells

Patient code	Viral tropism	% Inhibition of infection	SDa
3	R5	20.0	1.00
18	R5	20.7	1.51
23	R5	31.4	1.08
13	R5	36.2	1.27
8	R5	37.0	1.06
7	R5	37.3	1.47
1	R5	37.9	1.44
21	R5	40.3	1.19
9	R5	46.9	1.19
4	R5	49.2	1.11
22	R5	60.2	1.29

^aStandard deviations of the geometric means were calculated as described in footnote b, Table 1.

In conclusion, our approach permits the accurate measurement of the susceptibility of plasma virus to entry inhibitors targeting virus adsorption and membrane fusion (AMD3100, RANTES, and T-20). Our system may provide a valuable tool for the optimization of treatment strategies that include entry inhibitors and for detection of the appearance of phenotypic resistance in treated patients. We measured a wide range of baseline susceptibilities to T-20, which were not associated with polymorphism in the N helix of gp41. Thus, the domains that determine baseline susceptibility to T-20 are, at least in part, distinct from those implicated in the development of acquired resistance to this drug. We showed that, for dual-tropic viruses, the alternative usage of CCR5 or CXCR4 could affect their susceptibility to T-20. We did not find a correlation between the susceptibilities of plasma viruses to entry inhibitors that target different steps in the entry process. Sequence analysis of the entire envelope cassette from a large number of plasma viruses, selected on the basis of their different susceptibilities to fusion inhibitors, will help to identify the determinants of this characteristic and may contribute to further understanding of the fusion process.