Macrophages and HIV/AIDS Pathogenesis: Lessons from the Rhesus Macaque Model

Abstract

Destruction of CD4+ T cells is a primary cause of immunodeficiency in Human Immunodeficiency Virus (HIV)-infected humans and Simian Immunodeficiency Virus (SIV)-infected rhesus macaques. Tissue macrophages, however, also contribute to AIDS pathogenesis. Studies on rhesus macaque lung revealed the presence of at least two types of macrophages comprising short-lived lung interstitial macrophages in the parenchyma that are not present in bronchoalveolar lavage (BAL), and the long-lived alveolar macrophages that predominate in BAL and rarely divide. Increased blood monocyte turnover was associated with death of infected short-lived tissue macrophages and terminal disease progression during AIDS. Antiretroviral therapy (ART) treatment of SIV-infected macaques effectively prevented active infection of short-lived macrophages in tissues and delayed disease progression. Interestingly however, longer-lived macrophages remained infected and survived despite ART. This suggests that the long-lived macrophages contribute to establishing a virus reservoir and that these infected persistent cells likely become dysregulated to promote chronic inflammation. Furthermore, macrophages are the predominant immunological cells in heart, adipose tissue, and lung, and these were primarily of the long-lived macrophage subset. Information about macrophages garnered from the SIV rhesus macaque model provides a basis to further develop intervention strategies that target macrophages for reducing chronic inflammatory co-morbidities and remove a contributing viral reservoir for achieving cure.

1. Introduction

Macrophages are immune cells located in tissues throughout the body that early in life, originate from the yolk sac during development and later from bone marrow-derived blood monocytes. Macrophages regulate inflammation and tissue repair to maintain or re-establish homeostasis in response to environmental exposures [1, 2, 3, 4]. These cells exhibit tremendous plasticity and have been categorized in relation to their surface marker expression as well as pro- and anti-inflammatory responses [5, 6, 7]. Results from murine studies initially drove the categorization of polarized macrophage populations and origins. These contributed to better understanding monocyte and macrophage differentiation and plasticity in humans [5, 6]. Nonhuman primates that physiologically resemble humans, also contributed to the in vivo characterization of short-lived and long-lived macrophages using approaches that are difficult to perform in humans [8]. In addition, studies using the SIV-rhesus macaque infection model that simulates the disease pathogenesis observed in humans with HIV infection have afforded further characterization of macrophages in vivo [9, 10, 11].

Prior to the era of antiretroviral therapy (ART) application, HIV or SIV infection commonly led to AIDS resulting from a loss of CD4⁺ T cells [12, 13]. We also learned that an increase in monocyte turnover apparently required to replace damaged tissue macrophages, also correlated with terminal disease progression [14, 15]. The loss in CD4⁺ T cells and infected short-lived macrophages facilitated the development of opportunistic infections and cancers leading to death. HIV- or SIV infected hosts administered ART, however, survive with low-to-undetectable plasma viral levels, but typically develop HIV/SIV-associated non-AIDS conditions resembling chronic inflammatory diseases that occur in the elderly but at an earlier age than in non-HIV/SIV-infected individuals, a process sometimes considered accelerated or accentuated aging [16]. Chronic inflammation is likely a result of macrophage dysregulation which can be more readily evaluated in rhesus macaques infected with SIV and treated with ART. And, although, CD4⁺ T lymphocytes serve as a primary target of infection, HIV and SIV are lentiviruses that also infect macrophages, especially as CD4⁺ T cells decline in numbers leaving macrophages more available for infection [3, 17, 18, 19, 20, 21, 22, 23]. These macrophage subpopulations appear to exhibit distinct functions during progressing stages of HIV/SIV infection [22, 24, 25, 26]. Thus, studies on SIV in rhesus macaques are relevant for addressing macrophage subpopulations and their roles in pathogenesis in vivo.

2. Macrophages plasticity

Metchnikoff described macrophages as large (macro) cells that eat (phage) particles and phagocytize dead cells, debris, and pathogens as a first-line of innate immune responses, which remains a primary function of these cells [5, 6]. Macrophages also present antigen in the context of major histocompatibility complex antigens to promote adaptive immune responses by lymphocytes. To facilitate these functions, macrophages exhibit plasticity or polarized activation states that impact secretion of and responsiveness to a wide range of cytokines and chemokines to regulate inflammation, tissue repair, and (re)establishment of homeostasis [5, 6]. The activation of macrophages to perform these functions is multi-dimensional and impacted by intrinsic and extrinsic tissue environmental stimuli. Exposures to extrinsic stimuli including cell debris, pathogenic agents, or toxins trigger innate immune responses, recruitment of immune cells and secretion of a wide array of chemokines and cytokines. These extrinsic stimuli were further characterized using in vitro models of macrophage polarization. At one end of this spectrum, Th1 signals such as interferon stimulate M1 “classical” macrophages to promote acute inflammatory responses in response to intracellular pathogens. On the other end of the spectrum, Th2 factors such as IL4 stimulate M2 “alternative” macrophages to exhibit anti-inflammatory and anthelmintic responses as well as wound healing or tissue repair. These categorizations were further refined into a range of intermediate activation states and chronic or smoldering inflammation [6, 27]. Intrinsic factors impacting function consider macrophage origin via monocyte derivation versus resident embryonic progenitor status. Resident macrophages of embryonic yolk sac origin function in tissue remodeling and clearance of dead cells while monocyte-derived macrophages appear to function in microbial defense. It had long been believed that tissue macrophages are continually replaced by recruited blood monocytes that differentiate into tissue macrophages. More recently, studies in mice demonstrated that resident macrophages can undergo self-renewal [28]. In addition, monocyte-derived macrophages may acquire resident macrophage characteristics yet retain some epigenetic and transcriptional distinctions, suggesting that tissue environment may be more physiologically relevant than macrophage origin [6, 28, 29]. However, it is less clear and has been difficult to determine if embryonic-derived resident macrophages self-renew in nonhuman primates or humans throughout life [6, 28].

3. Tools to study macrophage populations in vivo in rhesus macaques

Rhesus macaque monocytes and macrophages express a wide range of biomarkers and cytokines / chemokines similar to those in humans, which can be identified via immune-detection and molecular biology methods. Surface biomarkers for CD14 and CD16 expression are commonly used to identify subpopulations of classical (CD14⁺, CD16⁻) pro-inflammatory, intermediate / transitional (CD14⁺, CD16⁺) and non-classical (CD14⁺/⁻, CD16⁻) anti-inflammatory monocytes that traffic to tissues where they differentiate into macrophages [6]. Among surface markers, tissue macrophages express scavenger receptors (e.g. CD163), Toll-like receptors (TLR), scavenger and lectin receptors (e.g. CD163, CD206), glycoproteins (e.g. lysosomal-associated membrane protein CD68) and MHC class II moieties for recognition of pathogens and antigen presentation.

In addition to analyzing surface biomarker expression, these cells have been characterized as short-lived macrophages having recently differentiated from trafficking monocytes or as longer-lived (i.e. resident) macrophages. An approach to identify shorter-lived macrophages in rhesus macaques is to monitor the incorporation of thymidine analogues such as 5-bromo-2′-deoxyuridine (BrdU). BrdU has a short half-life of a few hours in vivo and incorporates in cells dividing during this time, including monocyte precursors in bone marrow. Immunostaining then can be applied to detect the recently dividing monocytes and follow their trafficking patterns from bone marrow to blood to tissues, as well as their turnover rates [14, 22, 24, 25, 26, 30, 31]. To detect longer-lived macrophages, dextran, a branched polymer of anhydroglucose can be administered. Phagocytic cells take up fluorescein-labeled dextran, and long-lived macrophages containing the conjugated dextran can be identified weeks or months later whereas the shorter-lived cells with dextran die in a few days and are replaced with recently-recruited unlabeled cells [22, 32].

4. SIV infection in nonhuman primates

SIV infections in nonhuman primates have served as invaluable animal models for HIV infection and AIDS research [12, 13, 33]. Nonhuman primates are similar to humans physiologically and immunologically. Advantages to using nonhuman primates in SIV/HIV research is the ability to closely regulate parameters such as time, dose, and route of infection, as well as control for confounders such as smoking, diet, or illicit drug use. Experimental procedures for longitudinal tissue sampling, experimental vaccine testing, and therapeutic investigations also can be performed more readily in nonhuman primates. Naturally-infected nonhuman primate hosts such as African green monkeys and sooty mangabeys exhibit few clinical signs of disease from persistent or non-progressing SIV infection. Conversely, non-natural nonhuman primate hosts experimentally infected with SIV such as rhesus macaques, pigtail macaques, and cynomolgus macaques, exhibit a course of disease similar to that in humans with HIV infection. The differences between these non-pathogenic and pathogenic nonhuman primate models, respectively, provided opportunities to study comparative mechanisms of pathogenesis and immunity related to distinct outcomes of infection [34].

5. Increased monocyte turnover correlated to onset of terminal disease/AIDS in the absence of ART

Declining CD4⁺ T cells is a key biomarker for immune-deficiency, morbidity, and mortality during HIV and SIV infections of rhesus macaques [13, 35]. However, in SIV-infected rhesus macaques, we observed exceptions whereby a few animals with relatively higher levels of CD4⁺ T cells still developed AIDS while others exhibited low CD4⁺ T cell numbers (≤10% of baseline) with no obvious clinical signs of AIDS [14, 15, 36, 37] (Video 1: https://www.youtube.com/watch?v=g-RfAJyZsg0&t=23s and Video 2: https://www.youtube.com/watch?v=ynIom7fefxs&t=68s). Interestingly, while the number of monocytes in blood remained level, the percentage of recently dividing blood monocytes undergoing turnover, based on incorporation of BrdU administered 24 hours previously, drastically increased during terminal disease progression to AIDS in SIV-infected macaques [14, 25, 36, 37, 38]. To assess this observation further, machine learning algorithm modeling was applied to measure the relative contribution of covariates (singly or in combination) of monocyte turnover rate, percent CD4⁺ T-cell loss, plasma viral levels, and viral strains to survival time or days until necropsy based on clinical AIDS criteria [15]. Matched time-point data sets were used (not imputed) and results from animals surviving to experimentally-timed endpoints were excluded. By Boosted Forest regression, monocyte turnover had the highest proportion of contribution to days until AIDS-associated necropsy (0.475 or 47.5%) followed by CD4⁺ T-cell loss (0.216 or 21.6%), viral load (0.187 or 18.7%) or virus strain (below random) [15]. Classification categorization using decision tree algorithm modeling to analyze the flow chart-like proportion of the co-variates in predicting survival time in SIV-infected rhesus macaques further demonstrated that monocyte turnover (MTO) predicted a significantly shorter survival time among animals with ≥13.2% monocyte turnover rate followed by CD4⁺ T cell decline. These results led to further investigations to delineate macrophage subpopulations and their roles in SIV/HIV infection pathogenesis.

6. Tissue macrophages in the pathogenesis of HIV/SIV infection

Increased monocyte turnover during terminal disease progression to AIDS suggested that monocytes and macrophages contribute to pathogenesis. While CD4⁺ T cells are preferentially targeted by HIV/SIV, macrophages also become infected with HIV/SIV, especially if a tissue contains few T cells (i.e. brain) or after CD4⁺ T cell decline [19, 39, 40]. Infected macrophages are detected throughout the body [3, 4, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50]. Furthermore, resident macrophages infected with virus are resistant to apoptosis [17, 42, 51, 52, 53] and cytotoxic CD8 T cell lysis [54] thereby supporting their likely contribution to the virus reservoir.

1. Lung. In lung tissues and BAL fluids of rhesus macaques, two predominant populations of macrophages have been identified and characterized using in vivo administration of BrdU and dextran in conjunction with immunostaining methods to identify short-lived and long-lived macrophages, respectively [24]. Short-lived interstitial macrophages were relatively smaller in size, express CD163 (scavenger receptor) and primarily located in the lung interstitial tissue. Long-lived alveolar macrophages were relatively larger, double-positive (DP) for expressing both CD163 and CD206 (mannose receptor), and predominated in the alveolar spaces but also found in the interstitial tissue [24]. During terminal disease progression in SIV-infected rhesus macaques, the short-lived single-positive CD163⁺CD206⁻ macrophages in the interstitial lung tissues were readily infected and destroyed by SIV infection [25]. This loss in short-lived tissue macrophages may induce the increased production and turnover of monocytes to traffic and replace the damaged tissue macrophage that occurs during the transition from chronic to terminal disease progression after SIV infection. Furthermore, lung tissue lesions were progressively more severe with increasing monocyte turnover rates and macrophage accumulation (Figure 1) [25, 26]. Interestingly, resident long-lived double-positive CD163⁺CD206⁺ alveolar macrophages survived SIV infection despite efficacy in ART for reducing plasma viral loads to possibly become a virus reservoir [25]. In addition, while SIV-infected as well as uninfected CD4⁺ T cells were killed in lungs of infected rhesus macaques, macrophages did not decline and persisted, even if infected further indicating a population of long-lived macrophages that likely contribute to the virus reservoir [19].

2. Brain. Four populations of microglial, meningeal, choroid plexus, and peivascular macrophages exist in the brain [40, 55]. Resident microglial macrophages and macrophages of the meninges and choroid plexus exhibit low turnover rates and although infectable, are not major sites of HIV or SIV infection. Perivascular macrophages are located at the intersection between the central nervous system (CNS) and blood where systemic stimuli are encountered. These macrophages function in antigen presentation via MHC class II expression, exhibit a relatively higher turnover rate than parenchymal microglia during homeostasus, and are repopulated by peripheral monocytes that express CD14^highCD16^intermed to then express CD206 and CD163 [56]. In SIV-infected macaques, perivascular macrophages that were replaced by trafficking infected monocytes establish CNS infection, whereas direct intrathecal injection with virus failed to produce brain infection [55, 57, 58]. The virus-infected perivascular macrophages were immunophenotypically similar to the trafficking activated (CD16⁺) blood monocyte subsets and expressed CD163 [59]. Furthermore, increased monocyte turnover via BrdU kinetics, recruitment to brain, and accumulation of macrophages correlated with severity of encephalitis and neuronal injury in SIV-infected rhesus macaques. Encephalitic lesions were primarily comprised of myeloid cells including parenchymal macrophages, i.e. microglia, and perivascular macrophage cuffs, and macrophage accumulation better correlated with neurocognitive decline severity than did virus production [60, 61]. In addition, macrophage activation in relation to production of soluble CD163 (sCD163) in plasma and cerebral spinal fluid correlated with reduced neurocognition and virus levels further supporting macrophage contributions to SIV/HIV disease [38, 59, 61]. Macrophages of the brain could be reactivated to produce infectious virus using a virus outgrowth assay to also support their potential as a virus resevoir [41].

3. Intestine. The intestinal tract is a primary and early site of HIV/SIV infection that leads to massive depletion of intestinal CD4⁺ T cells and high rate of viral replication [12]. The intestine also houses among the largest population of macrophages [62], and increased accumulation of intestinal macrophages was reported in HIV-positive patients [63] as well as in the ileum of SIV-infected rhesus macaques, especially during acute SIV infection and AIDS, but not during the chronic stage of infection [21]. The accumulated CD163⁺ macrophages exhibited reduced phagocytic function that appeared to contribute to loss of intestinal integrity fueling further recruitment of macrophages in attempt to remove debris [64]. In rhesus macaques, two subsets of macrophages that comprised CD163⁺ CD206⁺ double-positive (DP) macrophages and CD163⁺ CD206⁻ single-positive (SP) macrophages were detected in the intestines using flow cytometry analyses [22]. In uninfected macaques, DP macrophages predominated over the SP macrophages. Forty-eight hours after BrdU administration, the majority of recently-dividing labeled cells were CD163⁺ macrophages in the jejunum and colon in uninfected rhesus macaques that increased in SIV-infected animals [15]. Thus, monocytes appeared to migrate more rapidly from blood to intestine (lamina propria) during progression to AIDS.

   DP macrophages predominated over SP macrophages in the lamina propria of uninfected rhesus macaques but SP macrophages were more common in the lamina propria of SIV-infected animals with increased monocyte turnover. Interestingly, in the submucosa and muscular mucosa, DP macrophages were more frequent than SP macrophages in both infected and uninfected animals and there was a statistically significantly higher mean DP:SP macrophage ratio in the lamina propria of uninfected macaques compared to infected animals exhibiting higher monocyte turnover. In contrast, there were no significant differences in the submucosa DP:SP macrophage ratios of uninfected versus SIV-infected macaques with intermediate-to-higher monocyte turnover. Similar trends were observed for the DP:SP macrophage ratios in the jejunum of uninfected compared to infected animals with higher monocyte turnover [22]. This suggested that during terminal stages of SIV infection in animals with higher monocyte turnover, there was a loss in DP macrophages with concurrent increases in SP macrophages in the lamina propria, whereas submucosal DP and SP macrophage levels remained steady even after SIV infection and disease progression.

   DP macrophages appeared to localize and remain primarily in the submucosa regardless of SIV infection status and appeared to comprise a long-lived macrophage population. This is relevant because long-lived (vs short-lived) macrophages would more likely serve as a virus reservoir, become dysregulated, and thereby contribute to chronic inflammation and pathogenesis. To explore this, we analyzed the distribution of colon macrophages that incorporated and retained the dextran (i.e. long-lived macrophages) relative to those that were labeled with thymidine analogues, BrdU or EdU (i.e. recently-dividing short-lived macrophages) [22]. BrdU⁺CD163⁺ macrophages were only in the lamina propria but not in the submucosa. Conversely, two months after dextran injection, dextran⁺CD163⁺ macrophages exclusively localized in the submucosa where there were ~ 10 times more DP macrophages than SP macrophages. These findings suggested DP macrophages in the submucosa are long-lived cells and that short-lived macrophages migrate from blood to lamina propria where they remain for shorter periods of time [22]. The presence of HIV-infected macrophages in human intestine and SIV-infected macrophages in rhesus macaque intestine supports the contribution of macrophages to gut pathogenesis and another tissue site of the virus reservoir [22, 23, 39].

4. Heart. In an early study characterizing heart tissues of rhesus macaques, macrophages were uniformly distributed throughout the heart in animals of all age groups ranging from infants to elderly adults, and were more prevalent than CD3⁺ T cells and CD20⁺ B cells [32]. Macrophages comprised approximately 2% of heart tissue cells in the younger animals and increased to a mean of nearly 4% in the older adults. CD163⁺ macrophages predominated over HAM56⁺ and CD206⁺ macrophages, and were detected at significantly higher percentage in the older animals between 13 and 24 years of age as well as in heart tissues exhibiting severe histopathology or inflammation in animals of all age groups. In vivo dextran labeling and retention indicated that at least half of the macrophages were longer-lived in healthy adult heart tissues and may comprise the tissue-resident population of macrophages.

   In heart tissues of SIV-infected rhesus macaques, increased numbers of CD163⁺ macrophages were associated with pathology and fibrosis, and macrophages infected with SIV expressed CD163 with some also expressing HAM56 [65, 66]. In our studies of rhesus macaques, the majority of heart tissue macrophages were CD163⁺ and long-lived (i.e. retained dextran), regardless of whether animals were chronically infected with or without ART [32] and Petkov et al., submitted]. Short-lived macrophages were rarely detected in the hearts of SIV-infected macaques at low monocyte turnover stage and of uninfected macaques with low histopathology scores [32]. In uninfected animals with higher histopathology scores (i.e. cardiac disease) or in SIV-infected animals with increased monocyte turnover, higher percentages of BrdU⁺ short-lived macrophages were detected in heart suggesting infiltration of monocyte-derived short-lived macrophages during disease progression. Reduction in monocyte traffic to the heart of SIV-infected rhesus macaques via anti-alpha-4 integrin antibody reduced lesion severity in the heart and decreased levels of CD163⁺CD68⁺ macrophages, implicating accumulation of short-lived macrophages as a promoter of SIV-associated heart disease [67]. Unpublished results from our lab suggested that the level of long-lived macrophage infection also influences the dynamics and recruitment of short-lived macrophages related to heart tissue damage during HIV infection. In addition, HIV and SIV infections appear to accelerate biological aging based on increased or earlier onset of chronic inflammatory diseases, including CVD [68, 69, 70, 71, 72, 73, 74, 75]. Cellularity of heart tissues declined with age in rhesus macaques [32] and there also was a reduction in heart tissue cellularity in SIV-infected younger adult macaques similar to that in uninfected older animals. Thus an increase in macrophages relative to tissue cellularity may contribute to loss of tissue homeostasis during natural aging and SIV/HIV-associated accelerated aging.

5. Adipose tissue. Tissue macrophages also constitute a major portion of the immune cell population in adipose [76, 77, 78]. Macrophages in subcutaneous adipose tissue of rhesus macaques and humans expressed CD68 and predominated over T cells and B cells [79]. CD68 was heavily expressed on cells of the monocyte and macrophage lineage [80] along with CD163 and CD206 that appear to reflect ‘anti-inflammatory’ macrophages [81, 82]. In subcutaneous adipose tissue of rhesus macaques, there were at least two major macrophage subsets. CD68⁺CD163⁺CD206⁺ macrophages were more common than the CD68⁺CD163⁻CD206⁻ macrophages, suggesting the predominant population exhibited an ‘anti-inflammatory’ phenotype. While triple-positive adipose tissue macrophages were found throughout the tissue, CD68⁺CD163⁻CD206⁻ were mainly detected inside crown-like structures and interstitial clusters, intimating these macrophages may be associated with pathology and inflammation [79]. As also observed in heart tissue macrophages, the vast majority of subcutaneous adipose tissue macrophages in rhesus macaques were longer-living (i.e. retained in vivo administered dextran) [32, 79]. These longer-lived subcutaneous adipose macrophages co-expressed CD68, CD163, and CD206. The CD68⁺ macrophages that did not contain dextran were generally found in the crown-like structure and interstitial spaces, and also were negative for CD163 and CD206. A small but distinct subpopulation of these macrophages simultaneously expressed Ki67 in their nuclei, located to the crown-like structure and interstitial clusters, but also to the interstitial spaces, suggesting that they were self-renewing at the time of sampling. Though only a small proportion of macrophages were self-renewing at any one time, these observations provide evidence that, like in mice, adipose tissue macrophages from rhesus macaques are predominantly long-lived and may self-maintain in tissue via slow local proliferation.

   Adipose tissue macrophages also have provided evidence that CD14⁺ adipose tissue macrophages become infected with HIV and simian immunodeficiency virus (SIV) [45]. During SIV infection of cynomolgus macaques, subcutaneous macrophages also exhibited a pro-inflammatory phenotype expressing CD14 but not CD163 or CD206 [83] thereby contributing to the chronic inflammation often observed during SIV/HIV infection.

Figure 1.

Proposed mechanism of lung tissue damage in SIV-infected macaques undergoing ART. The lung contains at least two populations of macrophages; shorter-lived interstitial macrophages (IM) and longer-lived alveolar macrophages (AM). IM become massively infected with SIV and undergo a high rate of apoptosis that correlates with increased blood monocyte turnover. Conversely, SIV infection of the longer-lived AM does not lead to high rate of apoptosis compared to that of IM. ART appears to successfully block or inhibit SIV infection in IM but not in longer-lived AM. We thus expect that elimination of SIV-infected longer-lived AM, as well as SIV-infected CD4⁺ T cells, is crucial to reduce inflammation and reverse pulmonary disease progression in SIV-infected aged macaques undergoing ART.

7. Future goals: targeting macrophages to reduce pathogenesis

Macrophages regulate inflammation and contribute to the SIV/HIV reservoir, thereby affecting pathogenesis of infection. Two general approaches have been applied experimentally to target macrophages in macaques that may benefit treatments for HIV. Bisphosphonates have been used for treating osteoporosis and other bone diseases, and target bone macrophages or osteoclasts to inhibit bone resorption [84]. In rhesus and cynomologus macaques, liposome-encapsulated alendronate administered intravenously reduced levels of CD14⁺CD16⁻ and CD14⁻CD16⁺ monocytes as well as CD163⁺ tissue macrophages with relatively low or no toxicity [85]. Liposome-clodronate inoculated intracisternally selectively depleted CD206⁺ perivascular macrophages of the brain as well as CD14⁺ monocytes in SIV-infected rhesus macaques [56]. A second approach was to apply natalizumab, an antibody to α-4-integrin that inhibits monocyte/macrophage trafficking. This antibody reduced neuropathology and cardiac pathology in SIV-infected rhesus macaques [56, 67]. Increasing evidence for the presence of SIV/HIV-infected and persistent macrophage reservoirs supports further studies and approaches for modulating or targeting macrophages to achieve cure.

Acknowledgments

The authors gratefully acknowledge research funding from the National Institutes of Health (grants AI097059, AI110163, and MH108458, to M.J.K., AG052349 to E.S.D., HL139278 to E.S.D. and M.J.K., OD011104, OD010568, and OD024282 to the Tulane National Primate Research Center and OD011107 to the California National Primate Research Center).

Author contributions

E.S.D. and M.J.K. conceived and wrote the manuscript.

Declaration of interest

The authors declare no competing interests.