Experimental Adaptive Evolution of Simian Immunodeficiency Virus SIVcpz to Pandemic Human Immunodeficiency Virus Type 1 by Using a Humanized Mouse Model

Abstract

Human immunodeficiency virus type 1 (HIV-1), the causative agent of AIDS, originated from simian immunodeficiency virus from chimpanzees (SIVcpz), the precursor of the human virus, approximately 100 years ago. This indicates that HIV-1 has emerged through the cross-species transmission of SIVcpz from chimpanzees to humans. However, it remains unclear how SIVcpz has evolved into pandemic HIV-1 in humans. To address this question, we inoculated three SIVcpz strains (MB897, EK505, and MT145), four pandemic HIV-1 strains (NL4-3, NLCSFV3, JRCSF, and AD8), and two nonpandemic HIV-1 strains (YBF30 and DJO0131). Humanized mice infected with SIVcpz strain MB897, a virus phylogenetically similar to pandemic HIV-1, exhibited a peak viral load comparable to that of mice infected with pandemic HIV-1, while peak viral loads of mice infected with SIVcpz strain EK505 or MT145 as well as nonpandemic HIV-1 strains were significantly lower. These results suggest that SIVcpz strain MB897 is preadapted to humans, unlike the other SIVcpz strains. Moreover, viral RNA sequencing of MB897-infected humanized mice identified a nonsynonymous mutation in env, a G413R substitution in gp120. The infectivity of the gp120 G413R mutant of MB897 was significantly higher than that of parental MB897. Furthermore, we demonstrated that the gp120 G413R mutant of MB897 augments the capacity for viral replication in both in vitro cell cultures and humanized mice. Taken together, this is the first experimental investigation to use an animal model to demonstrate a gain-of-function evolution of SIVcpz into pandemic HIV-1.IMPORTANCE From the mid-20th century, humans have been exposed to the menace of infectious viral diseases, such as severe acute respiratory syndrome coronavirus, Ebola virus, and Zika virus. These outbreaks of emerging/reemerging viruses can be triggered by cross-species viral transmission from wild animals to humans, or zoonoses. HIV-1, the causative agent of AIDS, emerged by the cross-species transmission of SIVcpz, the HIV-1 precursor in chimpanzees, around 100 years ago. However, the process by which SIVcpz evolved to become HIV-1 in humans remains unclear. Here, by using a hematopoietic stem cell-transplanted humanized-mouse model, we experimentally recapitulate the evolutionary process of SIVcpz to become HIV-1. We provide evidence suggesting that a strain of SIVcpz, MB897, preadapted to infect humans over other SIVcpz strains. We further demonstrate a gain-of-function evolution of SIVcpz in infected humanized mice. Our study reveals that pandemic HIV-1 has emerged through at least two steps: preadaptation and subsequent gain-of-function mutations.

Keywords: HIV-1; SIVcpz; cross-species transmission; humanized mouse; viral evolution.

FIG 1

Evolutionary relationship of HIV-1 and SIVcpz. (A) Phylogenetic tree of HIV-1 and SIVcpz. The amino acid sequences of nine viral proteins (Gag, Pol, Vif, Vpr, Tat, Rev, Vpu, Env, and Nef) were concatenated and aligned as described in Materials and Methods. The GenBank accession numbers of the HIV/SIVcpz sequences used in this analysis are listed in Table 3. The tree was constructed using the maximum-likelihood method. Bootstrap values are represented on each node. The virus strains used in this study are shown in bold in individual colors. The scale bar indicates 0.2 amino acid substitutions per site. Ptt, Pan troglodytes troglodytes; Pts, Pan troglodytes schweinfurthii. (B) HIV-1 and SIVcpz used in this study. NA, not applicable (71). The Env V3 region of NL4-3 was swapped with that of JRCSF. (C) TZM-bl cell assay. One microgram of each IMC, except for YBF30, was transfected into HEK293T cells. Forty-eight hours posttransfection, the culture supernatant was harvested as virus supernatant (for more detail, see Materials and Methods). The amount of viral particles was quantified by p24 ELISA, and the infectivity of each virus was determined using the TZM-bl cell assay. The assay was performed in triplicate, and averages with SDs are shown. The virus supernatant of YBF30 was prepared as described in Materials and Methods. (D) The percentages of similarity between the viruses are summarized.

FIG 2

Dynamics of SIVcpzPtt infection in humanized mice. (A and B) The virus supernatants containing 5 ng of p24 antigen (from HIV-1M JRCSF [n = 5], HIV-1M AD8 [n = 6], HIV-1N YBF30 [n = 6], HIV-1N DJO0131 [n = 12], SIVcpzPtt MB897 [n = 10], SIVcpzPtt EK505 [n = 9], and SIVcpzPtt MT145 [n = 7]) or RPMI 1640 (n = 8; for mock infection) were inoculated into humanized mice, and the amount of vRNA in the plasma (A) and the level of peripheral human CD4⁺ T cells (CD45⁺ CD3⁺ CD4⁺ cells) (B) were analyzed at 0, 1, 2, 3, 5, 7, 9, 12, and 15 wpi. Each average with SEM is shown as a dot with an error bar, and the values from each mouse are shown by lines. x axes, numbers of weeks postinfection. (A) The detection limit of HIV-1 RNA is 800 copies/ml plasma. (C to F) Virological features of infected mice at 15 wpi. (C and F) Maximum vRNA by 15 wpi (C) and acute vRNA (i.e., vRNA at 2 wpi) (F) are shown. The values from each infected mouse are shown by dots. The horizontal bars represent averages with SEMs. (C) Filled dots represent the results newly obtained in this study, while open dots in the HIV-1M section (strains NL4-3, NLCSFV3, JRCSF, and AD8) represent the results from our previous studies (34, 38–41). (D and E) Dynamics of SIVcpzPtt infection in the spleens of humanized mice. (D) The levels of virus-infected cells (i.e., Gag-positive cells) in the spleens of mice infected with HIV-1M JRCSF (n = 5), HIV-1M AD8 (n = 5), HIV-1N YBF30 (n = 4), HIV-1N DJO0131 (n = 7), SIVcpzPtt MB897 (n = 7), SIVcpzPtt EK505 (n = 8), and SIVcpzPtt MT145 (n = 8) and of mock-infected mice (n = 11) were analyzed by flow cytometry. (E) The numbers of human CD4⁺ T cells in the spleens of mice infected with HIV-1M JRCSF (n = 5), HIV-1M AD8 (n = 6), HIV-1N YBF30 (n = 6), HIV-1N DJO0131 (n = 10), SIVcpzPtt MB897 (n = 8), SIVcpzPtt EK505 (n = 9), and SIVcpzPtt MT145 (n = 8) and of mock-infected mice (n = 11) were analyzed by flow cytometry. (A to C and F) Statistically significant differences were determined with the Mann-Whitney U test and are represented as follows: double dagger, P was <0.05 versus mock-infected mice; blue asterisk, P was <0.05 versus HIV-1M; and red asterisk, P was <0.05 versus SIVcpzPtt MB897. (D and E) Asterisks indicate statistically significant differences (P < 0.05 by the Mann-Whitney U test) from mock-infected mice. NS, no statistical significance.

FIG 3

Correlation between acute vRNA and CD4⁺ T cell decrease. (A) Correlation between virological parameters. The statistical values (P values and the absolute values of Spearman's rank correlation coefficient [r_s]) between the two virological parameters shown in Fig. 1C and 2 are summarized. (B and C) Correlation between acute vRNA level and CD4⁺ T cell decrease. The minimum CD4⁺ T cell decrease by 15 wpi (y axes) and acute vRNA level (i.e., plasma vRNA at 2 wpi) (x axes) are shown. The averages with SEMs (B) and the result from each mouse infected with HIV-1M AD8 (n = 6), HIV-1N YBF30 (n = 6), HIV-1N DJO0131 (n = 12), SIVcpzPtt MB897 (n = 10), SIVcpzPtt EK505 (n = 9), and SIVcpzPtt MT145 (n = 7) (C) are shown. The lines represent exponential approximation. Spearman's rank correlation coefficient (r_s) was adopted to determine a statistically significant correlation between each value.

FIG 4

Emergence of SIVcpzPtt derivatives in infected humanized mice. (a) Full-length vRNA in the plasma of mice infected with SIVcpzPtt MB897 (n = 3), SIVcpzPtt EK505 (n = 2), and SIVcpzPtt MT145 (n = 2) at 15 wpi was sequenced. Briefly, vRNA was isolated from the plasma of infected mice at 15 wpi, and cDNA was generated. Then, the 11 DNA fragments represented in the figure (fragments A to K) were obtained by RT-PCR using the primers listed in Table 2. Mouse identification numbers are shown on the left of each sequence and are identical to those in Table 1. The mutation g7032a (gp120 G413R) is indicated by a red asterisk. (B) Activity of SIVcpzPtt Vif to counteract human APOBEC3 proteins. The expression plasmids for SIVcpzPtt Vif or HIV-1M NL4-3 Vif (as a positive control) were cotransfected with pNL4-3Δvif and either APOBEC3F or APOBEC3G expression plasmids into HEK293T cells. (Top) TZM-bl cell assay. The infectivity of released virus was determined using TZM-bl cells. The assay was performed in triplicate. The data represent averages with SDs. (Bottom) Western blot. The input of cell lysate was standardized to α-tubulin (TUBA), and representative results are shown. (C) Structural homology model of SIVcpzPtt MB897 gp120 and its G413R derivative. The structure model of gp120 of SIVcpzPtt MB897 (top) and its point mutant, MB897 gp120 G413R (bottom), were generated by homology modeling via the SWISS-Model Server Program based on the crystal structure of the gp120 of HIV-1 clade C strain ZM176.66 (PDB accession number 4JST) as a template. The surface model (left) and cartoon model (middle) of parental MB897 gp120 and its G413R mutant are shown. The residue at position 413 is indicted in red (G413, top) or blue (R413, bottom). The boxed regions in the middle panels are enlarged in the right panels. (D) TZM-bl cell assay. One microgram of the IMCs of SIVcpzPtt MB897 (parental MB897) and its derivative (MB897 gp120 G413R) was individually transfected into HEK293T cells. Forty-eight hours posttransfection, the culture supernatant was harvested as virus preparation (for more detail, see Materials and Methods). The amount of viral particles was quantified by p24 ELISA, and the infectivity of each virus was determined with a TZM-bl cell assay. This assay was performed in triplicate, and averages with SDs are shown. *, P < 0.05 by Student's t test.

FIG 5

Functional analysis of SIVcpzPtt MB897 derivatives. (A) Scheme of the SIVcpzPtt MB897 derivatives used in this assay. The nucleotide numbers, which are based on the database (accession no. JN835461), are represented in parentheses in gray. (B and C) Activity of SIVcpzPtt Vpu to down-modulate human CD4 and tetherin proteins. Vesicular stomatitis virus Gag (VSVg)-pseudotyped HIV-1M and SIVcpzPtt solutions were inoculated into MT-4 cells at an MOI of 0.1. Forty-eight hours postinfection, the levels of surface CD4 and tetherin on Gag-positive cells were analyzed by flow cytometry. Representative dot plots (B) and a summary of mean fluorescent intensities (MFI) (C) are shown. (D) Activity of MB897 NLvpuTM to downregulate human tetherin protein. One microgram of the IMCs of SIVcpzPtt MB897 (pMB897) and MB897 NLvpuTM (pMB897 NLvpuTM) was individually transfected into HeLa cells. Forty-eight hours posttransfection, the level of surface tetherin on Gag-positive cells and the percentage of Gag-positive cells were analyzed by flow cytometry. Representative dot plots (D) and a summary (E) are shown. The amount of released viral particles in the culture supernatant was quantified by p24 ELISA, and the results are shown in the right portion of panel D. (F) TZM-bl cell assay. One microgram each of the IMCs of SIVcpzPtt MB897 and its derivatives were individually transfected into 293T cells. Forty-eight hours posttransfection, the culture supernatant was harvested as the virus preparation. The amount of viral particles was quantified by p24 ELISA, and the infectivity of each virus was determined using the TZM-bl cell assay. The value of parental MB897 was set to 100 to facilitate comparison. This assay was performed in triplicate, and averages with SDs are shown. (G and H) Viral replication assay in activated human CD4⁺ T cell culture. (G) PHA-activated human PBMCs were infected with each virus at an MOI of 0.1, and the amount of viral Gag in the culture supernatant was quantified. Dpi, days postinfection; NS, no statistical significance. (H) Cumulative amount of viral Gag. The assay was performed in triplicate, and averages with SDs are shown. (E, G, and H) *, P < 0.05 versus the value in parental MB897 (by Student's t test).

FIG 6

In vivo competition assay. (A) Scheme of the in vivo competition assay. In this assay, the 4 types of virus preparations containing 1.25 ng of p24 antigen each were coinoculated into humanized mice. After 6 weeks, the infected mice were euthanized and the vRNA in plasma was sequenced. (B and C) Proportion of each virus at 6 wpi. (B) The chromatograms of the sequences of input virus (left) and vRNA in the plasma of 5 infected mice (mouse identification numbers are 74 to 78 and are identical to those in Table 1) at 6 wpi are shown. The chromatograms of the direct sequence of the regions of MA (top), Vpu (middle), and gp120 (bottom) are represented, and the percentage of each sequence is indicated under the chromatogram. (C) The percentage of each mutant at 6 wpi (left) and the fold change of each mutant from the parental virus (right) are shown. The values from each infected mouse are shown by dots. The horizontal bars represent averages with SEMs. In the right panel, the numbers represent the fold change with respect to the parental virus. The values 25% (left) and 1 (right) are indicated by a broken horizontal line.

FIG 7

Investigation of the adaptive mutation of SIVcpz in humanized mice. (A and B) The virus supernatants containing 5 ng of p24 antigen (SIVcpzPtt parental MB897 [n = 10] or MB897 3mut derivative [n = 8]) were inoculated into humanized mice, and the amounts of vRNA in the plasma (A) were analyzed at 0, 1, 2, 3, 5, 7, and 9 wpi. Each average with SEM is shown as a dot with an error bar, and the values from each mouse are shown by lines. x axes, numbers of weeks postinfection. (A) The detection limit of HIV-1 RNA is 800 copies/ml plasma, and the results of parental MB897 are identical to those in Fig. 2A. (B) Area under the curve of the plasma vRNA. (C) TZM-bl cell assay. One microgram of each IMC was individually transfected into HEK293T cells. Forty-eight hours posttransfection, the culture supernatant was harvested as the virus preparation. The amount of viral particles was quantified by p24 ELISA, and the infectivity of each virus was determined using the TZM-bl cell assay. The value of parental virus was set to 100 to facilitate comparison. This assay was performed in triplicate, and averages with SDs are shown. (A and B) Asterisks indicate statistically significant differences (P < 0.05 by the Mann-Whitney U test) from parental MB897. (C) Asterisks indicate statistically significant differences (P < 0.05 by Student's t test) from the respective parental virus.