Anatomical and physiological aspects of the HIV infection pathogenesis in animal models

Full Text

Introduction

HIV is a human-specific pathogen, the infection with which leads to depletion and death of mature peripheral CD4⁺ T cells (hCD4⁺) and immature hematopoietic progenitor cells in the organs and tissues responsible for their production, including bone marrow (BM), thymus, brain and lymph nodes (LNs) [1].

HIV-1 is transmitted by percutaneous, perinatal, and sexual routes, while 80% of adult infection cases become infected through the exposure of mucosal surfaces to the virus [2]. Recent studies have demonstrated the ability of HIV to infect not only cells of the immune system, but also tissues of many organs, including the brain, the intestine, kidneys, and the prostate [3, 4].

CCR5-utilizing (non-syncytium-inducing, CCR5- tropic, or R5 isolates) HIV-1 strains are macrophage-tropic [5, 6] and are detected in blood during early or acute infection [7]. They are characterized by high rates of replication, producing large numbers of viral progeny. Some studies suggest that the initial infecting R5 virus often evolves into a dual-tropic R5/X4 virus [8]. The subsequent development of AIDS accompanied by depletion of CXCR4-expressing memory hCD4⁺ T cells in LNs, spleen, Peyer’s patches, nasal-associated lymphoid tissue, adenoids, tonsils, bronchus-associated lymphoid tissue, cryptopatches (CP), and isolated lymphoid follicles [6, 9–11] leads to emergence of a more cytopathic virus [8] utilizing the CXCR4 chemokine receptor (syncytium-inducing, CXCR4-tropic, or X4 isolate).

CCR5 is expressed at low levels on hCD4⁺ and hCD8⁺ thymocytes and on a subset of circulating memory or activated CD4⁺ T cells. Given this differential characteristic, X4 isolates, as opposed to R5 isolates, rapidly infect and destroy both immature and mature T cells, resulting in immune system collapse [12]. Nevertheless, R5 isolates are detected in approximately 50% of AIDS cases [13].

As mucosae contain higher levels of hCD4⁺ T cells with initially high activation and expression of CCR5 as compared to blood [3, 4, 14], they become the main reservoir of HIV-1 target cells in the contact route of transmission [15]. In addition, the induction of chemokines during early stages of infection initiates migration of HIV target cells to infection and inflammation sites, prompting rather than preventing the spread of the virus through its antiviral activity [16].

Regardless of the transmission route, the gastrointestinal tract (GIT) is a major site of HIV-1 replication and hCD4⁺ T cell depletion [17]. Therefore, the main consequence of acute HIV infection is massive destruction of hCD4^{+ T} cells in the GIT mucosa-associated lymphoid tissue [18]. Furthermore, the disruption of the GIT epithelium integrity leads to translocation of bacteria to the blood stream and, consequently, becomes the cause of inflammation and activation contributing to the loss of hCD4⁺ cells [19].

If mucosae remain intact, HIV can penetrate the epithelium by transcytosis and translocation of virus particles across epithelial cells [20, 21]; recent studies have demonstrated that Langerhans cells are capable of endocytic internalization of HIV-1 virions, which is followed by replication of virus particles in the cytoplasm and potential infectivity for the neighboring hCD4⁺ T cells [14]. In addition, the mucosal epithelium of the female reproductive tract (FRT) is a substantial barrier against infection; therefore, not all HIV-1 exposures result in productive infection [20].

Although hCD4⁺ T cells, brain macrophages and microglia are the main targets for HIV-1, lymphoid tissues are the main reservoir for HIV, compromising primarily latently infected, resting hCD4⁺ T cells carrying an integrated, replication-competent HIV form even in patients receiving antiretroviral therapy (ART). Biomarkers of macrophage activation and neuronal damage in the cerebrospinal fluid of HIV-infected patients with viral suppression demonstrate the induction of neurocognitive disorders through chronic inflammation of the central nervous system (CNS) tissues, presenting CNS tissues as a latent reservoir of the replication-competent virus [22]. Although resting hCD4⁺ lymphocytes are characterized by low levels of productive viral infection due to inhibition of the early reverse transcription of incoming viral genomes, the viral production, which results in formation of the fully infectious HIV particle, starts immediately after the latently infected, resting hCD4⁺ T cells have been activated [23].

Although HIV can replicate in chimpanzees (Pan troglodytes), it cannot infect mice, rats, rabbits, or macaques [24, 25]. The species-specificity of HIV is not limited to its interactions with CXCR4 or CCR5 on the surface of hCD4⁺ T cells. Multiple barriers to HIV replication, which have been found in mammals, limit the availability of adequate animal models to study fundamental aspects of HIV biology and its interactions with the host [16]. For example, rat or mouse cells that have been developed to express hCD4 and CCR5 or CXCR4 on their surface do not support HIV replication due to the absence of additional viral restriction factors such as TRIM5α and APOBEC3. The Vif protein interacts directly with APOBEC3, disrupting its activity and preventing its incorporation into viral particles [26, 27]. In its turn, TRIM5α regulates the ability of some retroviruses to infect human cells and the ability of HIV to infect primate cells. This protein interacts with the viral capsid and by blocking the process prevents its integrity disruption during the reverse transcription [28].

Chronic immune activation is a hallmark of the HIV pathogenesis [29]. Hyperimmune responses with the release of pro-inflammatory cytokines such as tumor necrosis factor-alpha (TNF-α) exhibiting cytotoxic properties [30] and HIV proteins [23] can increase the HIV replication rate and hCD4⁺ and hCD8⁺ T cell apoptosis [31]. In most cases, apoptosis is the result of direct viral infection or an indirect effect of immune activation. The direct cytopathic effect of HIV is caused by disturbance of metabolic processes and disruption of integrity of cell membranes of hCD4⁺ T lymphocytes and precursors, leading to apoptosis. Autophagy is pointed out as a possible cause of accidental death of hCD4⁺ cells [30, 32]. Another causative factor can be the activity of cytotoxic hCD8⁺ T cells against normal hCD4⁺ cells [20, 33]. The reduction of peripheral blood hCD4⁺ lymphocytes over time is caused by destruction of BM, lymphoid tissue, and by inhibition of cell reproduction, for example, in the thymus [30, 34].

Model organisms keeping natural anatomical and physiological characteristics

Great apes

Previous genetic studies have shown that HIV-1 groups M and N evolved from the simian immunodeficiency virus (SIV) infecting chimpanzees (SIVcpz), while groups O and P originated from SIV of gorillas (SIVgor) circulating in the gorilla population (Gorilla gorilla spp.) [2, 35]. SIVgor resulted from the SIVcpz transmission 100–200 years ago [36]. The analysis of highly conserved regions of the viral genome, which changed when the human species barrier has been crossed by monkey precursors of HIV-1, detected one site in the Gag-30 viral matrix protein, which encoded methionine and switched to arginine in the assumed ancestors of HIV-1 groups M, N, and O, thus resulting in arginine or lysine presence as a basic amino acid in most of the HIV-1 strains with than the isogenic virus with lysine at the same position, while the opposite process is observed in human cells [37]. The reversibility of the process is confirmed by experimental infection of chimpanzees with HIV-1 [38], which resulted in reversion of the basic residue from Gag-30 to methionine [39]. The progressive loss of CD4⁺ T cells, severe thrombocytopenia, and AIDS-specific signs result in fatal outcome in SIVcpz-infected chimpanzees [40].

Ethical issues significantly limit studies involving great apes. Although the kinetics of SIVcpz replication in vitro in hCD4⁺ T cells is similar to that of HIV, cell cultures fail to capture all the conditions of virus replication and transmission in vivo [41].

Non-human primates

It has been found that HIV-2 groups A through H originated from SIV of sooty mangabeys (Cercocebus atys) after it was transmitted to humans [2, 35, 42]. In the meantime, mangabeys, even when infected with SIV of sooty mangabeys, do not tend to develop AIDS [43], demonstrating their relevance as a natural model of HIV pathogenesis in "elite controllers" having low viremia levels due to the absent chronic immune activation [44, 45]. In its turn, HIV-1 infecting humans does not replicate in most of the non-human primates; therefore, related monkey viruses with their inherent limitations are used to model HIV infection [46].

Early in vivo studies of HIV infection pathogenesis in macrophages were conducted by infecting southern pig-tailed macaques (Macaca nemestrina) with SIV [47]. However, the replication of HIV and SIV in macrophages was difficult to study at that time; therefore, the presence of viral nucleic acids in these cells was attributed to the phagocytosis of infected T cells and cell debris [48].

SIV infection of southern pig-tailed macaques triggers virus production after activation of latently infected, resting CD4⁺ T cells in tissues and peripheral blood [23]. Although SIV was absent in peripheral blood of 85.7% of the animals undergoing ART, latently infected macrophages containing the replication-competent virus were detected in basal ganglia and parietal cortex ex vivo. However, this model employs viruses having tropism both for CD4⁺ lymphocytes and for macrophages; therefore, this study cannot be directly compared with studies of other phenotypically limited animal models or HIV-infected patients [22]. Infection of the related species, Sundaland pig-tailed macaques (Macaca leonina), with chimeric HIV demonstrates the relationship between higher intracellular APOBEC3 expression and lower virus replication rates during the acute stage [49, 50].

The inability to infect rhesus macaque (Macaca mulatta) cells with HIV-1 in vitro is in part due to the HIV-1 capsids being blocked by the TRIM5α host cell restriction factor and the failure of Vif to bind and induce APOBEC3 degradation. To bypass these limitations, sequences of the capsid and Vif of SIV of macaques (SIVmac) are incorporated in HIV-1 molecular clones. After its in vitro passage in human CEMx174 cells, the chimeric virus acquires 88% of the HIV-1 genome, including the capability of replicating steadily and inducing massive cytopathic effects. Thus, the accidental or adaptive bypass of capsid and Vif-associated limitations can guarantee the cross-species transmission of primate lentiviruses [46].

The HIV tropism for hCCR5⁺/hCD4⁺ lymphocytes in the intestinal lamina propria (LP) [51, 52] positively correlates with the data obtained from the SIVmac₂₅₁ infected animals [53–55], when the virus affects 30–60% of all CD4⁺ T cells during the acute stage. On the 3^rd day after they were infected intravenously (i/v), almost all the animals demonstrate an increase in the number of CD4⁺ T cells, which reaches the peak on the 10^th day to be followed by a sharp decrease in their number in blood and LNs, and a few days later – in the intestinal LP having highest levels of CD4⁺ lymphocytes. However, after having disappeared from blood, CCR5⁺/CD8⁺ T cells continue to be detected in other tissues [56].

Within two days of the intravaginal (i/vag) inoculation of SIV, rhesus macaques developed infection of the cervix. The virus spreads further through migration of CD4⁺ and dendritic cells to regional LNs and then to the blood stream. The infection is propagated not only in activated and proliferating T cells corresponding to the short-lived population that produces the bulk of HIV-1 virions in humans, but also in resting T cells [57].

To imitate in vivo development of CNS disorders associated with AIDS, rhesus macaques infected with SIVmac₂₅₁ received i/v injections of CD8 antibodies to promote migration and deposition of monocytes and macrophages to brain tissues, thus resulting in development of SIV-induced encephalopathy [58]. In its turn, the preliminary depletion of CD4⁺ lymphocytes by recombinant antibodies causes persistent infection of macrophages and the CD4-independent membrane developed by the virus to promote its entry to the CCR5-expressing cells in the absence of CD4 [59].

Ethical issues limiting in vivo studies using human hematopoietic stem cells (HSCs) in primates [60] prompted further studies of HIV infection in small animal models that partially have a human immune system.

Table 1 summarizes the data on model organisms having natural anatomical and physiological characteristics.

 

Table 1. Models maintaining natural anatomical and physiological features

Model	Method of infection; pathogen	Pathophysiological features
Chimpanzee	Natural; SIV of schimpanzee	CD4+ T-cell depletion, severe thrombocytopenia, and signs of AIDS leading to death [40]
	Intravenously (i/v); HIV-1	Mild thrombocytopenia and leukopenia [38]
Sooty mangabey	I/v; SIV of sooty mangabey-infected homologous plasma	Absence of chronic immune activation that preclude high levels of viremia [44]
Rhesus macaque	I/v; HIV-1 chimeric strain	Possibility of interspecific transmission between primates [46]
	I/v; SIVmac251	30–60% of all CD4+ T cells are affected. Almost complete loss of all CCR5+/CD4+ T cells in the jejunum [53–56]. The spread of the virus through the migration of dendritic and CD4+ T cells, including dormant ones, into the regional LN and bloodstream [57]. Acceleration of migration and deposition of monocytes and macrophages in brain tissues stimulate the development of SIV-induced disorders in the central nervous system against the background of AIDS [58].Persistent infection of macrophages and the appearance of CD4-independent envelope in the virus after the previous infection depletion of CD4+ cells [59]
Northern pig-tailed macaque	I/v; HIV-1-infected peripheral blood autologous mononuclear cells	Stable replication of HIV-1 in a non-target organism [50]
Sundaland pig-tailed macaque	I/v; macrophage-tropic SIV	Virus replication after activation of only latently infected resting CD4+ T cells [23]
	I/v; immunosuppressive and neurovirulent SIV	Persistent infection of macrophages [47]
	I/v; SIV	Retention of latently infected macrophages in the basal ganglia and parietal cortex in the absence of SIV after antiretroviral therapy [22]

 

Chimeric model organisms

The hu-Thy/Liv model

The chimeric SCID-hu Thy/Liv model constructed by transplanting human fetal thymus and liver tissues under the kidney capsule to homozygous C.B-17_scid/scid mice with severe combined immunodeficiency (SCID) [61, 62] has a reduced graft-versus-host disease (GVHD) and can support functions of grafts for up to 15 months [63]. The transplanted thymus tissues demonstrate morphological and functional similarity to tissues of the human normal fetal thymus by the composition of thymocyte subpopulations and expression of hCXCR4 and hCCR5 [12].

In contrast to identical kinetics of HIV infection during progression to AIDS, which is observed both in R5-isolate carriers and in X4-isolate carries, R5 isolates go through a two-stage infection cycle in the SCID-hu Thy/Liv model, first slowly replicating in medullary stromal cells, without causing any apparent pathology, and then infecting cortical CD4⁺/CD8⁺ thymocytes, thus subsequently causing their moderate depletion [6, 12]. The absolute levels of replication X4 isolates in cortical thymocytes, which are observed after 17 days of the infection [6, 12], decrease after 20 days of the infection due to HIV-1-mediated depletion of these thymocytes, resulting in their complete loss during the next 15–20 days [61]. At the same time, the p24 levels in mice infected with X4 isolates are 14 times as high as those in mice infected with R5 isolates due to the effect of Nef on replication and enhancement of the cytopathic component of the pathogenesis of X4 isolates during infection in vivo [64]. At the same time, X4 isolates can acquire an R5-like phenotype and infect medullary stromal thymocytes that poorly express CCR5, slowly replicating in them and occasionally causing their depletion [1].

In the SCID-hu Thy/Liv model, transplanted thymocytes can be infected only by the mode different from natural routes of pathogen transmission. The absence of systemic viremia prevents the analysis of viral replication and its effect on thymocytes, requiring surgical removal of the tissue fragment or animal euthanasia [62]. This explains the need for the animal model that can be infected through the routes similar to those in humans, for example, through mucosae, as HIV-infected cells can transmit the virus both to cells of the immune system and to cells of mucous membranes [65, 66].

BM of NOD/SCID/ γc−/− (NSG)-hu Thy/Liv mice that have only T cells (also known as T-cell only mice or ToM) does not produce human myeloid cells and B cells [67, 68], and as the model is γc deficient, it does not manifest GVHD signs for 14 months. After the i/v or intraperitoneal infection with R5 isolate, which does not replicate in vivo in tissue macrophages and which is tropic exclusively for human T cells, the presence of viral RNA (vRNA) and p24⁺ cells is observed across the entire brain, including cerebellum, medulla oblongata and cerebral cortex [67]. High levels of viral replication maintained in peripheral blood result in moderate reduction on the population of hCD4⁺ lymphocytes; however, daily administration of combined ART can significantly decrease the viral load to undetectable levels in plasma, while preserving latently infected, resting CD4⁺ T cells [68]. Thus, the NSG-hu Thy/Liv model demonstrates the competence of T cells in establishment and support of productive HIV infection of the brain and the absent need for myeloid cells for transportation of the virus from the periphery to the brain [67].

The hu-BLT model

The possibility of rectal and ivag infection of chimeric mice with HIV depends on the presence of human cells in FRT and rectum; however, the reconstitution of these cells depends directly on the mouse strain and the humanization protocol. The lower activity of endogenous natural killer cells in NOD/SCID and NSG mice compared to SCID mice contributes to longer and more reliable reconstitution of cells involved in the human innate and adaptive immune response. In addition, T cells of these mice, similar to human T cells, demonstrate quite a versatile repertoire of Vβ T-cell receptors [17]. These mice are humanized by the mode similar to that used for the hu-Thy/Liv model: the transplantation of the human fetal thymus-liver-thymus graft sandwich under the kidney capsule and its engraftment are followed by inoculation of fetal CD34⁺ HSCs. The model is referred to as bone marrow-liver-thymus (BLT). In NOD/SCID-hu BLT mice, the transplanted thymus tissue is developed into a human thymus-like thymic organoid, in which progenitor cells of human T lymphocytes, which are limited by the human leukocyte antigen (HLA), can migrate and develop into fully functional peripheral T cells within the autologous epithelium of human thymus [69].

The analysis of tissue distribution of populations of human immune cells in NOD/SCID-hu BLT mice revealed the respective distribution of human lymphoid and myeloid cells in BM, spleen, LNs, liver, lungs, and GIT [17]. Mucosae of the vagina, exocervix, endocervix, uterus, and intestine were characterized by intensive reconstitution of hCD4⁺ T cells, monocytes/macrophages, and dendritic cells. T cells are located directly inside the epithelial layer, forming a streak on the borderline between the epithelium and LP, along the basal membrane and along the lamina propria, while macrophages and dendritic cells are located in the lamina propria along the entire FRT of mice [69]. These morphological characteristics correlate with location of T cells in humans, providing the NOD/SCID-hu BLT model with susceptibility to effective vaginal transmission of R5 isolates [69, 70] and rectal transmission of X4 and R5 isolates [71, 72].

Non-traumatic i/vag inoculation of a single cell-free dose of the primary R5 isolate of HIV-1 to NOD/SCID-hu BLT mice results in establishment of systemic infection in 88% of mice, which is demonstrated by the presence of p24 antigen and/or vRNA in plasma 2 weeks after the infection and by progressive reduction in the population of hCD4⁺ T cells in peripheral blood, being similar to the acute stage of HIV-1 infection in humans. Disseminated infection is observed along the entire GIT: Infected cells are detected in LP and small intestine epithelium, causing a sharp loss of tissue hCD4⁺ lymphocytes and effector memory cells [69].

Most of the hCD4⁺ T cells, which are present in FRT of NOD/SCID-hu BLT and NSG-hu BLT mice, express simultaneously CCR7 and CCR5, being potential target cells for HIV, similar to the human cervical tissue. As stromal cells are a major source of CCL19 and CCL21 as well as CCR7 mouse ligands, humanized mice express them at higher levels than human ligands in the cervicovaginal tract and iliac LNs. 6-10 days after the infection, the number of p24⁺ hCD4⁺ T cells increases in the cervicovaginal tract compared to their number in cervical LNs, and 14 days later, infected cells and vRNA are detected in the intestine, plasma, and spleen, while vDNA is detected in mesenteric LNs. This kinetics of processes demonstrates the absence of critical significance of CCR7-dependent migration of leukocytes for HIV spread [73].

Structures similar to the gut-associated lymphoid tissue (GALT), which are present in humans, brought up a question regarding the ability of mouse CP to initiate the genesis of human GALT. Transmission of IL-7R signals is essential for the GALT genesis; therefore, GALT cannot develop normally in mice with disturbed transmission of IL-7R signals, including NSG mice, which do not have the common γ-chain. For this reason, rectal transmission of HIV-1 is studied in NOD/SCID-hu BLT mice that have CP and can accurately reproduce depletion of hCD4⁺ T cells in GALT structures. The reconstitution of hCD4⁺ and hCD8⁺ T cells, myeloid cells and single naive T cells along the entire epithelium and LP of small and large intestine makes the NOD/SCID-hu BLT model susceptible to rectal transmission of HIV-1 [17]. In addition, in NOD/SCID-hu BLT mice, plasmatic cells secreting hIgA dominate over cells secreting hIgG, thus correlating positively with the IgA secretion in intestinal LP of humans. Furthermore, the intestine of NSG hu-BLT mice contains small numbers of plasmatic cells with high levels of hIgA and hIgG secretion [74].

Rectal infection results in a significant increase in the number of perforin-positive cells in GALT structures in NOD/SCID-hu BLT mice, in production of p24⁺ cells and in decreased numbers of hCD4⁺ lymphocytes across the entire GIT [17], thus demonstrating the consistency with depletion of hCD4⁺ T cells in intestinal LP and intraepithelial compartments during HIV infection in humans. No differences in levels of hCD8⁺ T cells in HIV-positive and HIV-negative animals have been observed [74]. The human thymic stroma, which is present in NOD/SCID-hu BLT mice, is essential for production of human thymocytes within the entire human major histocompatibility complex (MHC), thus making the model quite promising for assessment of specific immune responses to potential vaccines against HIV [75].

Unfortunately, adoptive transfer of activated Vδ2 cells, which demonstrates relative therapeutic effectiveness in treatment of some infectious diseases, does not help control HIV infection in NSG-hu BLT mice; on the contrary, it aggravates viremia, thus suggesting a possible role of Vδ2 cells as early HIV targets in promotion of virus spread [76].

Interstitial hCD4⁺ lymphocytes in lungs, which express both hCCR5 and hCXCR4 [77], unlike hCD4⁺ lymphocytes of the alveolar space or circulating blood, tend to deplete during early stages of HIV-1 infection ex vivo in the human lung tissue and in vivo in NOD/SCID-hu BLT and NSG-hu BLT mice, demonstrating the highest levels of their infection [78]. The hu-BLT mice that are i/vag infected with the R5 isolate provide the possibility to observe in vivo depletion of hCD4⁺ T cells in the interstitial lung tissue [73] during acute and early chronic HIV-1 infection [79, 80]. The viral load in the common lung tissue and sorted hCD4⁺ T cells of the lung tissue as well as the p24⁺/hCD4⁺ cell ratio were higher compared to those in the spleen tissue, while the numbers of hCD4⁺ T cells in the interstitial lung tissue and bronchoalveolar lavage fluid showed a gradual decrease during 7 weeks. The total number of hCD4⁺ and hCD8⁺ T cells correlates with the number in the bronchoalveolar lavage fluid, though not in the interstitial lung tissue, suggesting that T lymphocytes are recruited to the alveolar space. Thus, the early and severe depletion of hCD4⁺ T cells in the lungs of these models correlates with the depletion observed in the intestine [78].

Considering that microglia evolved from early progenitor cells that developed in the yolk sac, it was doubtful that human microglial cells could be reproduced by modern models of humanized mice [67]. However, NSG-hu BLT and IL34-Tg/NOG-hu BLT mouse models demonstrate effective reconstitution of human microglia and the ability to support replication of HIV in the brain [67, 81].

NSG mice being hypersensitive to ionizing radiation demonstrate lower total numbers of human HSCs, myeloid cells and B cells in the brain after preconditioned irradiation at high dose rates. In addition, the brain of female NSG-hu BLT mice shows higher levels of reconstituted human B cells than the brain of male mice, though the number of T cells and the hCD4⁺/hCD8⁺ ratio do not depend on the gender or the dose rate. The presence of HIV target cells in the entire brain of NSG-hu BLT mice is confirmed by migration of HSCs hCD45⁺, hCD3⁺ T cells and hCD68⁺ macrophages from the olfactory bulb through brainstem base to the medulla oblongata. Subsets of hCD4⁺ and hCD8⁺ lymphocytes are present across the brain, including the olfactory bulb, cortex, caudate nucleus, thalamus, midbrain, pons, and cerebellum; the latter contains significantly larger numbers of hCD68⁺ macrophages than in the other compartments of the brain. The infection of NSG-hu BLT mice with R5 and X4 isolates through mucous membranes and parenterally results in detection of cell-associated vDNA and vRNA in 85.1% and 92.9% of the brain, respectively, regardless of the infection mode. As the infection is progressing, the increased vDNA levels in the brain entail rapid and steady reduction in the total number of human T cells and hCD4⁺ T cells as well as reduction in the hCD4⁺/hCD8⁺ ratio in the brain, which is similar to the loss of hCD4⁺ T cells in FRT and GIT mucosal tissues during early stages of infection. p24⁺ cells are detected in the cortex, cerebellum, thalamus, medulla oblongata, and midbrain. There is a direct relationship between the levels of cell-associated vRNA in the brain and the viral load in plasma as well as between the duration of infection and the number of myeloid cells [67]. In addition, the significance of atypical cell subsets in HIV persistence is demonstrated by reactivation of the persisting virus in perivascular macrophages, astrocytes, and microglia of the brain in infected mice [82].

While in NSG-hu BLT mice, human T cells develop in the human thymus epithelium and are limited by human MHC [70, 83], in NOD-Rag2−/−γc−/− (NRG)-hu BLT mice, human T cells are produced in the thymus with the mouse MHC [84–86]. In the same way, in the above model, HIV-1 infection induces activation and depletion of T cells [87–89].

Humanization of the BLT model of TKO mice lacking Rag2, Il2rg, and CD47 genes is accompanied by stable reconstitution of all major cell subsets in BM, spleen, and mesenteric LNs, including hCD4⁺ and hCD8⁺ T lymphocytes, hCD19⁺ B cells, hCD14⁺ monocytes, myeloid and plasmacytoid dendritic cells. In addition, the areas organized into follicles in the white pulp of the spleen are reconstituted with primarily mature hCD3⁺ T cells and hCD20⁺ B cells, and in GALT of the large and small intestine – with subsets of activated hCD4⁺ and hCD8⁺ lymphocytes as well as with myeloid cells and plasmacytoid dendritic cells. Rectal and intraperitoneal [90] infection of TKO-hu BLT mice with the R5 isolate of HIV-1 induces production of gp120-specific antibodies and reduction in the number of hCD4⁺ T cells along with the increasing number of activated hCD8⁺ T cells. Thus, the TKO-hu BLT model demonstrates classical immunological signs of acute HIV-1 infection in humans [91].

Depending on the selected mouse strain, hu-BLT models can reconstitute the human functional immune system [92, 93] and become infected with HIV-1 through routes similar to those of humans [94], bringing the model to the level of the gold standard for in vivo studies of HIV-1 [95, 96].

The hu-HSC model

NOD (non-obese diabetic) models based on Rag1^–/–lIl2rg^–/– and Rag2^–/–Il2rg^–/– deficient mice (lacking two genes, double knockout, DKO) that are engrafted with human hCD34⁺ HSCs demonstrate susceptibility to infection with R5 and X4 isolates of HIV-1 through the mucous membrane, resulting in chronic infection [97, 98].

Transplantation of BM cells or hCD34⁺ umbilical cord blood cells to NSG mice results in successful differentiation among various populations of cells, including human T and B cells as well as natural killer cells, monocytes/macrophages, and dendritic cells. In addition, hCD4⁺ T cells of this model are highly susceptible to R5 and X4 isolates inducing a high viral load in the plasma [99–101] for more than 40 days [101].

BRG mice were used to create the model with consistent reconstitution and development of LNs eliminating downsides of development of the secondary lymphoid tissue. The BRGST-hu HSC model based on Balb/c Rag2⁻/⁻Il2rg⁻/⁻Sirpa^NOD (BRGS) mice expresses transgenic thymic stromal lymphopoietin, which does not depend on IL2rg and is similar to IL7 by its structure and function; the model has stable cellular and humoral human responses due to stimulation of B and T-cell responses [102]. Thus, BRGST-hu HSC mice demonstrate enhanced immunoglobulin isotype switching, development of central memory and effector memory T cells and follicular T-helper cells in secondary lymphoid tissues with pronounced B-cell zones [103].

Humanized mice devoid of human T cells (myeloid-only mice, MoM) are one of the versions of the hu-HSC model; they are characterized by stable levels of viremia accompanied by the increased number of human macrophages in the brain. Replication of macrophage-tropic R5 isolates of HIV-1 and HIV-2 takes place only in reconstituted macrophages and myeloid cells within 15 weeks. In their turn, T cell-tropic R5 isolates of HIV-1 or HIV-2 do not replicate as T cells are absent. In infected NOD/SCID-hu Mo mice, hCD68+ macrophages and p24+ cells are detected across the brain, including the cerebellum, putamen, cortex, ventral striatum, and brainstem [104]. In this model, the long-term infection supported by tissue macrophages in the absence of T cells implies that they become vulnerable as true targets and latent reservoirs of HIV infection in humans. In addition, these mice can transmit de novo the pathogen, which replicates effectively in the new host in the presence or in the absence of human T cells [104].

The results obtained in the NRG-hu Thy/HSC model, which was created by transplantation of a fetal thymus fragment beneath the kidney capsule and i/v injection of human HSCs to sublethally irradiated NRG mice, demonstrate the activation, functionality loss, and T cell depletion similar to those observed in NRG-hu HSC mice after the infection with HIV-1 as well as the similar level of CD38 and HLA expression on hCD8⁺ lymphocytes. The hCD4⁺/hCD8⁺ T cell ratio was insignificantly higher in these mice, while the percentage content of human natural killer cells, plasmacytoid dendritic cells, and monocytes was lower than in the NRG-hu HSC mice. In addition, NRG-hu Thy/HSC mice can express PD-1 – the marker of T-cell functionality loss [105].

Despite the absence of mature B cells capable of isotype switching [106], productive infection of FRT mucous membranes in NRG-hu HSC mice provides the best model of the contact route of HIV-1 transmission and offers the possibility to study changes in the microbiota [107], local pre-exposure prevention [108, 109], and vaccine therapy [106].

The hu-PBMC model

The hu-PBMC model offers the simplest method for mouse humanization; however, the heterogeneity of transplanted cell populations increases the risk of early development of GVHD and requires animals with high levels of immunodeficiency. Reconstituted hCD45⁺ cells are detected in peripheral blood, LNs, spleen, and liver of the mice approximately 4 weeks after human peripheral blood mononuclear cells have been injected intraperitoneally into NSG mice, while hCD19⁺ cells are detected only in LNs and the spleen. hCD3⁺ T cells are detected in LNs, the spleen and liver, while hCD4⁺ and hCD8⁺ subpopulations have extensive localization. Most of the hCD3⁺ T cells in the red pulp are hCD8⁺ T cells, while the white pulp contains mainly hCD4⁺ T cells. hCD8⁺ lymphocytes are also detected around periportal regions of the liver, while hCD4⁺ T cells are distributed across tissues more diffusively. hCD4⁺ and hCD8⁺ T cells reconstitute actively in LNs containing single mouse T cells. The normal hCD4⁺/hCD8⁺ ratio was detected only in the donor; it was twice as high in recipient mice and remained unchanged for 10 weeks. The intravenous infection with the dual-tropic HIV-1_DH12 strain resulted in depletion of hCD4⁺ T cells in peripheral blood 3 weeks after the infection, and in their complete loss in another 2 weeks. In the mice infected with X4 isolate, depletion of hCD4⁺ T cells is observed 4 weeks after the infection, demonstrating the relationship between the kinetics of hCD4⁺ lymphocyte depletion and the tropism of the infectious agent [110].

Although NSG mice do not have mature endogenous T and B cells and natural killer cells, the NSG-hu PBMC model developed GVHD signs within 4–16 weeks after the humanization [110, 111]. For this reason, the use of the model contravenes the animal-involving research policy adopted by most research institutions [112].

TKO mice are the most suitable animals for creating the hu-PBMC model [112], as the lack of CD47 and subsequent tolerance to transplanted human cells delay the GVHD development and body weight loss on average by 24 days compared to the NSG-hu PBMC model [112]. The TKO-hu PBMC mice demonstrate the capability of massive repopulation of multiple tissues and organs by hCD3⁺, hCD4⁺, and hCD8⁺ cells, including the spleen, mesenteric LNs, BM, liver, large intestine, rectum, and brain, while hCD14⁺/hCD163⁺ macrophages and hCD20⁺ B cells reconstitute in LNs and spleen. Intraperitoneal infection with the R5 isolate causes steady viremia 2 weeks later; the viremia is accompanied by reduction in the number of hCD4⁺ lymphocytes; note that the viral load in the plasma increases while the number of circulating hCD4⁺ T cells decreases proportionally, accurately mimicking HIV-1 infection in humans. Rectal and i/vag modes of infection demonstrate similar results, though with some delay in development of infection signs. The p24 antigen is detected in mesenteric LNs, BM, spleen, and rectum; however, although the model is able to reconstitute hCD4⁺ T cells in the brain, the p24 indication results are not convincing [112]. With the humanization technique, which does not require special skills, and sufficiently safe infection methods used by the operator to infect animals, this new model stands a good chance to gain recognition from the scientific community.

Transgenic animals

Resources of human tissue for creating humanized mice are limited, the transplantation procedure for successful engraftment requires profound technical skills of the personnel, and additional animal models are required due to the impossibility of AIDS pathogenesis assessment. As it turned out, activation of cell signaling pathways, inhibition of mouse MHC, increased pathogenicity of virus particles and Nef-mediated enhancement of viral replication and pathogenesis by inhibition of hCD4, which were observed during some in vitro studies, are not the most important characteristics for the Nef phenotype in vivo [113]. For example, the mice expressing the whole HIV-1 genome containing a modified long terminal repeat and the mice capable of expressing Nef, Gag, or protease in lens fibers develop depletion and eye disorders. The expression of a 3' half-HIV-1 genome causes severe nephropathy in mice, while the expression of Nef or Tat causes epidermal hyperplasia. In their turn, mice, whose T cells express CD4⁺, and CD4C/HIV^WT Tg mice, which have cells expressing the whole HIV-1 genome, are characterized by low viability due to reduced monocyte-macrophage-lymphocyte cell populations, if they have concomitant kidney and lung disorders phenotypically close to those in humans with AIDS [114]. The similarity between this model and AIDS manifestations in children points at the critical role of Nef in AIDS progression in humans, regardless of its role in virus replication [115]. The expression of Tat protein in CNS of GT-tg mice helps assess behavioral disorders associated with HIV-1 Tat effect [116].

Some models expressing myeloid hSCF, hGM-CSF, and hIL-3 cytokines (NSGS mice) [117, 118] as well as MITRG and MISTR models [119] turned out to be non-viable, demonstrating short life span of 10–20 weeks after the humanization.

Finally, it has been reported about creation of transgenic rabbits carrying the HIV-1hCD4/hCCR5 entry receptor complex through combined microinjection of both constructs. The developers of this model are planning to use primary cell cultures obtained from these animals for HIV adaptation and, subsequently, for changes in the virus [23].

Table 2 presents the summarized data on chimeric model organisms.

 

Table 2. Chimeric models

Mouse model	Method of infection; pathogen	Anatomical and physiological features	Pathophysiological features
SCID-hu Thy/Liv	HIV-1 inoculation into graft tissue	Grafts’ thymocytes subpopulations composition and hCXCR4 and hCCR5 expression are similar to the normal human fetal thymus tissues, the functionality is kept for 6 to 15 months [6, 12, 61–63]	MHC-limited T cells infection and depletion, up to their complete disappearance [6, 12, 61–63]
NSG-hu Thy/Liv	I/v and intraperitoneally; T-cell-tropic R5 HIV-1 isolate	System reconstitution of human T-cells [68]	Productive infection of all brain divisions. Myeloid cells are not required to transport HIV from the periphery to the brain [67]. Maintaining high levels of virus replication in peripheral blood, leading to a moderate decrease in the hCD4+ T cell population. Administration of combined antiretroviral therapy leads to a significant suppression of virus replication up to undetectable in plasma with the preservation of latently infected resting hCD4+ T cells [68]
NOD/SCID-hu BLT	Intravaginally (i/vag) [17, 69] or rectally [70, 71, 74]; HIV-1	HLA-limited intensive reconstitution, corresponding distribution and functionality of human lymphoid and myeloid cells in the tissues of the thymic organoid [69], BM, spleen, LN, liver, lungs, FRT [17, 69], gastrointestinal tract [17, 70, 71, 74] and thymic organoid [69] in the context of human MHC [65]. T cells demonstrate a diverse repertoire of Vß TCR receptors [17, 69]. The predominance of hIgA-secreting plasma cells over hIgM-secreting cells [74]	System infection and progressive depletion of hCD4+ T cells [17, 71] including interstitial lung tissue [74]. The number of human perforin-positive cells was significantly increased [74]
NSG-hu BLT	I/v [76, 88, 89, 104], i/vag, rectally or orally [94]; HIV-1	The genesis of GALT is impossible [74]. hCD34+ HSCs differentiate to mature forms capable of HIV infection [76]. Effective reconstitution of human microglia due to migration of hCD45+ cells, hCD3+ T cells and hCD68+ macrophages from the olfactory bulb through the base of the brainstem to the medulla oblongata [67]. There are a large number of human B cells in the brain of females than in males [67]. The presence of hCD4+ and hCD8+ T cells in the entire brain, while the cerebellum contains a significantly larger number of hCD68+ cells than in the brain. A small number of plasmocytes with a high level of hIgA and hIgG secretion [74]	Activation and depletion of hCD4+ T cells [74, 88, 89]. Maintenance of HIV replication in bone marrow hCD68+ macrophages [67, 104], regardless of the method of infection [67]. There is a direct relationship between the viral load in plasma and the levels of cell-associated RNA in the brain, and between the number of myeloid cells and the duration of infection [67]. Adoptively transferred activated Vδ2 cells promote the spread of the virus as early HIV targets [76]
NRG-hu BLT	I/v; HIV-1	Reconstitution of all major leukocyte subgroups hCD45+, including in the thymus in the context of mouse MHC [84–86, 105]	Activation and depletion of T cells [87–89]. Persistent HIV-1 infection leads to stable and systemic induction of IFN-I [87]
DKO-hu HSC	I/vag; HIV-1	Reconstitution of T-, B-, myeloid cells and natural killers in central and peripheral lymphoid organs [97, 98]	Susceptibility to mucosal viral transmission. hCD4+ T-cells depletion [97, 98]
NRG-hu HSC	I/v; HIV-1	Human hemopoiesis reconstitution, including T-, B-, myeloid and plasmocytoid dendritic cells, natural killers and HSCs [84, 86]. The percentage of hCD45+ cells engraftment in the bone marrow is 6-fold higher than in NOD/SCID mice [86]	Depletion, activation and exhaustion of T cells [87–89]. Expression of the PD-1 T-cell depletion marker [105]
NRG-hu HSC	I/vag; HIV-1		System infection and progressive depletion of hCD4+ T-cells in the gastrointestinal tract and FRT [107]
NRG-hu Thy/HSC	I/v; HIV-1	The ratio of hCD4+/hCD8+ T cells is slightly higher, and the percentage of natural killer cells, plasmacytoid dendritic cells and monocytes is lower compared to NRG-hu HSC mice [86, 105]	Depletion, activation and exhaustion of T cells, CD38 and HLA expression on hCD8 T cells, expression of PD-1 T cell depletion marker [105]
NOD/SCID-hu Mo	I/v; macrophage-tropic R5 HIV-1 and HIV-2 isolates [104]	Reconstitution of hCD68+ macrophages throughout the brain [104]	HIV replication occurs only in myeloid cells and hCD68+ macrophages, increasing the number of the latter. The absence of a significant decrease in the level of viremia after reaching the peak. The possibility of de novo transmission of infection and effective replication in the new host in the presence or complete absence of human T-cells [104]
NSG-hu PBMC	I/v; HIV-1DH 12	Recovery of hCD45+ cells in peripheral blood, lymph nodes, spleen and liver. hCD19+ cells are detected only in the lymph nodes and spleen. T cells are actively regenerated in lymph nodes containing single mouse T cells [110]	Depletion and complete disappearance of hCD4+ T-cells in peripheral blood [110]
IL34–Tg/NOG-hu BLT		Effective reconstitution of human microglia [81]	Maintenance of HIV replication in the brain [81]
BRGST-hu HSC		Reconstitution of central and effector memory T cells and follicular T helper cells providing stable human cellular and humoral reactions [118] in secondary lymphoid tissues with pronounced B-cell zones [119]. The presence of HLA class I and II transgenes improve the development and functionality of T and B cells, including switching of Ig isotype classes and antigen-specific responses [103]	–
CD4C/HIVWT Tg		HIV-1 expression in mouse cells corresponding to cells of HIV-positive humans [114]	Low viability of mice, a significant decrease in monocyte-macrophage-lymphocyte cell populations in the presence of concomitant lung and kidney lesions [114]
GT-tg		Tat protein expression in the central nervous system [116]

 

Conclusion

SIV shares only 50% identity with HIV; there are significant differences in the composition of the γδ-lymphocyte subset in the monkey and human phenotype [76]; however, SIV-infected non-human primates are used as the main animal model for in vivo assessment of the effectiveness of potential vaccines and microbicides [120, 121]. Although in vaccines, epitopes must be compatible with the infection-inducing virus, for a long time SIV of rhesus macaques has been used most frequently for preclinical assessment of vaccines against HIV [75].

When using humanized mice in ART studies, close attention should be given to preparation and provision of viral epitopes in the context of human rather than mouse MHC. For example, immunodominant epitopes presented by mouse MHC for human T-cell receptors may not be directly relevant to human immunodominant epitopes presented by MHC, let alone to human T-cell receptors; therefore, the process of development of T cells in humanized mice in the absence of human stroma should be thoroughly studied [75].

Compared to NSG-hu HSC, NSG-hu Thy/Liv, and NRG-hu HSC mice, NSG-hu BLT mice have better reconstitution of the functional human immune system, including mucosal immune system [92, 93], and can be infected with HIV-1 through intravaginal, rectal, and oral routes [94]. For this reason, hu-BLT mice are currently considered the gold standard for HIV-1 research in murine models [95, 96].

In humans, the transition of T cell-tropic CNS viruses to a macrophage-tropic phenotype generally corresponds with a longer timer of infection and extensive replication leading to viral evolution. However, the overall lifespan of mice is relatively short compared with that of humans; therefore, there is no possibility to follow the animal models over several years after their infection [67], and the problem of the relevant animal model for HIV infection remains unsolved.

About the authors

Aleksey M. Nagornykh

Central Research Institute for Epidemiology

Email: nagornih@cmd.su
ORCID iD: 0000-0002-8320-2999

Cand. Sci. (Vet.), senior researcher, Scientific group of experimental models’ development, Laboratory of experimental pharmacology, Department of molecular diagnostics and epidemiology

Russian Federation, Moscow

Marina A. Tyumentseva

Central Research Institute for Epidemiology

Email: nagornih@cmd.su
ORCID iD: 0000-0002-3145-3702

Cand. Sci. (Biol.), Head, Laboratory of genome editing, Department of molecular diagnostics and epidemiology

Russian Federation, Moscow

Aleksandr I. Tyumentsev

Central Research Institute for Epidemiology

Email: nagornih@cmd.su
ORCID iD: 0000-0003-0537-2586

Cand. Sci. (Biol.), Head, Laboratory of experimental pharmacology, Department of molecular diagnostics and epidemiology

Russian Federation, Moscow

Vasily G. Akimkin

Central Research Institute for Epidemiology

Author for correspondence.
Email: nagornih@cmd.su
ORCID iD: 0000-0003-4228-9044

D. Sci. (Med.), Professor, Full Member of the Russian Academy of Sciences, Director, Central Research Institute for Epi- demiology

Russian Federation, Moscow

References

Supplementary files