ABSTRACT

   The monkeypox (MPXV) outbreak in 2022 is more prevalent among
   individuals with human immunodeficiency virus (HIV). While it is
   plausible that HIV-induced immunosuppression could result in a more
   severe progression, the exact mechanisms remain undetermined. To better
   understand the immunopathology of MPXV in patients with and without HIV
   infection, we employed single-cell RNA sequencing (scRNA-seq) to
   analyse peripheral blood mononuclear cells (PBMCs) from six patients
   hospitalized for MPXV, three of whom had HIV infection (HIV antibody
   positive and HIV RNA level below the detection limit), and three
   patients only infected with MPXV (HIV-). We map the peripheral immune
   response in both the acute phase and the recovery period, showing the
   reconfiguration of peripheral immune cell phenotypes in acute stage
   compared with recovery stage, characterized by disturbed cell subsets
   and intense cell interactions mediated by monocytes and neutrophils.
   Importantly, we also found obviously dysregulated gene expression and
   cell subsets in HIV+ patients proposing mechanism underlying their
   serious condition. Our findings provide a comprehensive cell atlas of
   MPXV patients, shed light on the mechanisms underlying the severe
   disease progression and longer recovery time in HIV+ individuals.

   KEYWORDS: Monkeypox, HIV, single-cell RNA sequencing, immune response,
   antiviral innate immunity

Introduction

   The monkeypox virus (MPXV) causes a human disease that resembles
   smallpox and has become the most prevalent Ortho pox virus affecting
   humans since the eradication of smallpox in 1980 [[41]1]. The 2022
   global outbreaks, spanning over 100 countries, have raised significant
   concerns [[42]2]. The clinical presentation of this virus typically
   includes fever, rash, lymphadenopathy, and susceptibility to secondary
   bacterial infections. However, unlike previous outbreaks, recent study
   indicates that the 2022 MPXV shows epidemiological changes, lineage B.1
   drives the outbreak, aided by APOBEC3 in viral evolution and
   lesion-based transmission fostering variants [[43]3]. Monkeypox
   infection can occur simultaneously with HIV and SARS-CoV-2 infections
   [[44]4], and its incidence among men who have sex with men (MSM) is
   higher, leading to predominantly localized lesions rather than
   extensive disseminated lesions [[45]5,[46]6]. Actually, among monkeypox
   cases with HIV, up to about 50% are HIV positive [[47]7]. Emerging data
   has indicated that such co-infections may lead to worse clinical
   outcomes and higher mortality rates [[48]8], and the immune response to
   monkeypox virus has attracted tremendous attention [[49]9].
   Understanding the immune profiles of patients with and without HIV is
   essential for uncovering the shared pathogenesis and developing
   evidence-based strategies for managing co-infections.

   Following MPXV infection, the interactions between host immune cells
   and the pathogen evolve over time, significantly influencing the
   pathogenesis. Cytokine profiling has indicated an overproduction of
   certain cytokines in patients with severe MPXV disease, highlighting
   the critical role of an imbalanced immune response in clinical severity
   [[50]10]. Immune characteristics are intricately linked to the
   pathogenesis and disease progression of patients infected with viruses
   [[51]11]. HIV selectively infects and destroys CD4 T cells, with the
   number of circulating CD4 T cells in HIV+ individuals predicting the
   onset of overt immunodeficiency [[52]12]. HIV infection causes
   persistent depletion of CD4^+ T cells and severe immunodeficiency.
   Emerging data has indicated that such co-infections may lead to worse
   clinical outcomes, higher rates of secondary bacterial infection and
   higher mortality rates [[53]8,[54]13]. The relative study is still
   limited, and the information available regarding the in-vivo kinetics
   and crosstalk of immune cells in MPXV-infected individuals is limited.
   Given the prevalence of co-infections, there is an urgent need to
   elucidate the kinetics of the immune cells atlas in MPXV patients with
   or without HIV, to decipher the possible impact of HIV-associated
   immunocompromise on the progression of MPXV.

   Single-cell RNA sequencing (scRNA-seq) studies on diverse immune cells
   across different cohorts significantly contributed to our understanding
   of immune dysfunction in virus-related physiopathology. It could
   identify novel cell types and subtypes, discover the intricate
   diversity of immune cell functions. For instance, it provides in-depth
   understanding of the molecular and cellular dynamics underlying
   COVID-19, thereby offering crucial insights for the development of
   diagnostic tools, therapeutic strategies, and public health
   interventions [[55]14–16]. scRNA-seq has been utilized for the rapid
   acquisition of monoclonal antibodies during early outbreaks of MPXV,
   showing its versatility in pandemic response [[56]17]. However, there
   has been a notable absence of comprehensive investigations into the
   immune responses of patients afflicted with MPXV, particularly those
   co-infected with HIV.

   Here, we examined the pathological characteristics of patients with
   MPXV infection and those co-infected with HIV. We compared the
   peripheral immune cells proportion and gene expression of MPXV patients
   from acute phase to recovery phase using scRNA-seq, six hospitalized
   patients in a monocentre from January 2023 to June 2023 were included,
   three of them were co-infected with HIV. Cell subsets and gene
   expression were also compared between HIV+ and HIV– patients to
   decipher the impact of HIV-induced immune dysfunction to MPXV
   immunopathogenic. We aim to discuss our current understanding of the
   immune landscape of MPXV-infected patients and the impact of HIV on
   MPXV immunopathogenesis. Specifically, our findings identify cell
   type-specific gene expression changes in acute phase compared with
   recovery phase. We proposed mechanisms underlying the enhanced
   hospitality of HIV+ patients, including disturbed Mono/DC dynamic
   decreased NK cells and CD4 T cells, expansion of some CD8 T-cell
   subsets, and dysregulated gene expression in B cells.

Results

Characteristics and treatment outcome of patients in the present study

   Patients diagnosed with monkeypox in the Tianjin Second People’s
   Hospital since July 2023 according to the Chinese Monkeypox Diagnosis
   and Treatment Guidelines (2022 Edition). Patients were diagnosed by
   detecting monkeypox virus DNA through real-time PCR testing of
   specimens taken from the rash site, and recovery was determined when
   all rashes had scabbed off. All patients were male, with an average age
   of 38.17 ± 7.08. Among them, five patients reported a history of
   homosexual activity within 2 weeks before the onset of symptoms and
   only one patient had been vaccinated against smallpox. Patients were
   grouped based on whether the HIV virus was detectable at the time of
   enrolment. Among them, three cases of HIV+ patients who had been
   receiving regular highly active antiretroviral therapy (HAART) for over
   2 years with undetectable HIV RNA levels and positive HIV antibodies at
   enrolment. Their clinical characteristics including CD4 counts are
   listed in [57]Table 1. There is a sharp decrease in CD4 counts of HIV+
   patients. Rash areas in two groups were indistinctively observed at
   genitals, anus, extremities, torso, oropharynx, face and neck, as well
   as at inguinal lymph nodes. However, a higher incidence of gonococcal
   and purulent infections, as well as conditions such as anal fissures,
   was observed among HIV-positive patients. Since there are no specific
   antiviral drugs for monkeypox currently, besides the HAART drugs
   required for HIV patients, no other antiviral drugs were used.
   Treatment mainly focused on symptomatic support and management of
   complications, with emphasis on local disinfection using
   povidone-iodine type III, and antibacterial drugs were administered to
   patients with bacterial infections. The average duration of
   hospitalization was 15.5 days, with 13 days for the HIV– group and 18
   days for the HIV+ group ([58]Table 1).

Table 1.

   Characteristics of patients.
   Clinical features All patients
   (n = 6) HIV positive
   (n = 3) HIV negative
   (n = 3)
   Age (years) 38.17 ± 7.08 39.67 ± 9.07 36.67 ± 6.03
   Smallpox vaccination [n (%)] 1(16.7) 1(33.3) 0(0)
   Homosexual activity [n (%)] 5(83.3) 3(100) 2(66.7)
   Fever time(day) 3.50 ± 2.43 3 ± 2.65 4 ± 2.65
   Rash area
    Genitals/anus [n (%)] 5(83.3) 2(66.7) 3(100)
    Extremities [n (%)] 4(66.7) 1(33.3) 3(100)
    Torso [n (%)] 2(33.3) 0(0) 2(66.7)
    oropharynx [n (%)] 1(16.7) 0(0) 1(33.3)
    Face and neck [n (%)] 5(83.3) 2(66.7) 3(100)
    Inguinal Lymphadenopathy [n (%)] 4(66.7) 2(66.7) 2(66.7)
   Complications
    Gonococcal infection [n (%)] 3(50.0) 3(100) 0(0)
    Purulent infection [n (%)] 3(50.0) 3(100) 0(0)
    Anal fissure [n (%)] 1(16.7) 1(33.3) 0(0)
    Perianitis [n (%)] 1(16.7) 1(33.3) 0(0)
    Syphilis [n (%)] 2(33.3) 1(33.3) 1(33.3)
    Penile contusion [n (%)] 1(16.7) 0(0) 1(33.3)
    Hypocythemia [n (%)] 1(16.7) 0(0) 1(33.3)
   Ct value of Vesicular fluid (Median) 25.57 25.13 26.00
   ALT (U/L) 25.5 ± 14.9 13.8 ± 6.6 37.2 ± 9.9
   AST (U/L) 23.0 ± 6.9 17.8 ± 3.2 28.2 ± 5.4
   GGT (U/L) 35.2 ± 14.8 28.5 ± 9.2 41.8 ± 18.2
   ALB (g/L) 42.7 ± 2.9 43.3 ± 1.4 42.2 ± 4.3
   CD45 + (n/μL) 1989.9 ± 769.2 1811.1 ± 1065.2 2168 ± 298.8
   CD3 + (n/μL) 1435.4 ± 541.1 1270.4 ± 685.1 1600.4 ± 425.5
   CD4 + (n/μL) 737.1 ± 316.8 485.3 ± 100.6 988.9 ± 224.6
   CD8 + (n/μL) 667.6 ± 421.2 744.8 ± 613.4 590.4 ± 222.1
   CD19 + (n/μL) 117.7 ± 42.1 133.8 ± 60.1 101.6 ± 7.1
   CD16+/CD56(n/μL)+ 337.2 ± 144.3 235.5 ± 112.9 439.0 ± 90.8
   CD4/CD8 1.31 ± 0.6 0.9 ± 0.4 1.7 ± 0.4
   HIV-RNA(copies/mL)   <20 Negaitive
   HIV IgM   Positive Negative
   [59]Open in a new tab

   Note: Data for age and fever days are shown as mean ± SD.

Cellular profiles show difference between HIV+ and HIV– groups through MPXV
progression

   To profile the peripheral immune response to MPXV in HIV+ and HIV–
   patients, we performed scRNA-seq on PBMCs from MPXV individuals in
   acute and recovery phases. A total of 253,397 cells were retained for
   the following analysis by removing cells with abnormal RNA count and
   too many mitochondrial genes. We created a cells-by-genes expression
   matrix and performed dimensionality reduction by uniform manifold
   approximation and projection (UMAP) and graph-based clustering, and
   identified 41 clusters. No cell cluster shows patient individual bias
   ([60]Figure 1(a)), but cells from patients of different MPXV stage
   ([61]Figure 1(b)) or HIV infection status ([62]Figure 1(c)) exhibit
   somewhat different distribution in the embedding space, indicates
   potential influence of MPXV PBMC composition, and the impact of HIV
   infection on MPXV immunopathogenesis. Manual annotation based on the
   canonical markers obtained total of eight cell types, including B cell,
   CD4 T cell, CD8 T cell, megakaryocyte (Mega), monocyte-derived
   dendritic cell (MonoDC), neutrophil (Neu), NK cell, and
   undifferentiated T cells (undiff T), and the distance between the same
   cell type in the UMAP embedding is obviously close than those between
   distinct cell types ([63]Figure 1(d)). We next quantified the cells to
   identify MPXV-derived cell type proportions changes and impact of HIV
   on them. In MPXV patients, B cells, NK cells, and a subset of T cells
   showed an obvious increase and CD4 T cells, neutrophil and MonoDC
   showed significant decrease in the recovery stage compared to the acute
   stage ([64]Figure 1(e)). Strikingly, though the decrease of CD4 T and
   CD8 T cell in HIV+ patients was observed, not all cell subset
   proportions show significant difference (p value >0.05) between HIV+
   and HIV– patients ([65]Figure 1(f)). To find the possible impact of HIV
   on MPXV patients’ PBMC dynamic composition, we further compared the
   cell compositions between HIV+ and HIV– patients stratified by the MPXV
   infection status. A significant decrease in CD4+ T cells and expansion
   of neutrophils were observed in HIV-positive patients during the acute
   stage of MPXV infection ([66]Figure 1(g)). In the recovery phase, the
   proportion CD8^+ T cell, Mono/DC and neutrophil was higher in HIV+
   patients, while NK cell and undifferentiated T cell (undiff T cell)
   were still largely inhibited in HIV+ patients ([67]Figure 1(h)). To
   further represent the enrichment of each cell type, we calculated the
   ratio observed to the expected proportion of each of the 41 cell
   clusters ([68]Figure 1(i)). As indicated in [69]Figure 1(j), each cell
   type shows enrichment or depletion of different subsets, both
   consistent and contrasting changes in PBMC composition were observed in
   HIV-positive patients. For example, five out of the nine CD8^+ T-cell
   subclusters were significantly enriched in PBMC samples from the acute
   stage of MPXV infection, but two of those five CD8^+ T cells were
   significantly depleted in HIV+ PBMC samples ([70]Figure 1(j)). These
   data collectively indicated that MPXV infection causes comprehensive
   immune cells alteration, and HIV patients showed CD4 defects in acute
   phage, followed by dysregulated NK cells and neutrophils in recovery
   phase.

Figure 1.

   [71]Figure 1.
   [72]Open in a new tab

   Cell clustering analysis. (a–d) UMAP visualization of cell clusters
   coloured by patient cohorts, MPXV infection stage, HIV infection status
   and cell identity. (e) Cell composition comparison between acute and
   recovery stage after MPXV infection. (f–h) Cell composition comparison
   between HIV– and HIV+ samples from both acute and recovery stages of
   MPXV infection, acute stage of MPXV infection, and recovery stage of
   MPXV infection. (i) UMAP visualization of all cells coloured by cell
   subclusters. (j) Ratio of observed/expected (Roe) of each cell
   subcluster in different patients to analyse depletion or enrichment
   status of each subcluster in the corresponding patient group. P values
   in e–h were determined by Wilcoxon test. The Roe of each subcluster was
   calculated through Chi-square test.

Changes of B lymphocytes and gene expressions upon MPXV infection

   B cells secrete antibodies and mediate the humoral immune response,
   making them extremely important in protective immunity against virus
   [[73]18]. We first assessed the B-cell response in MPXV infection by
   UMAP embedding of B cells and five clusters were identified based on
   canonical marker gene expressions ([74]Figure 2(a)). We assessed the
   expression of genes associated with activation, such as CD79A and
   CD79B, and plasma markers such as MZB1, XBP1 in different B-cell
   subsets. CD79A and CD79B belong to the immunoglobulin superfamily, they
   covalently bound together serve as the main signal transduction
   molecules during antigen binding B-cell receptor (BCR). CD79A were
   mainly enriched in the subclusters 1, 3, and 5, while MZB1 and XBP1
   were mainly expressed in the subclusters 2 and 4 ([75]Figure 2(b)),
   indicate that they might represent plasma cells for antibody
   production. Meanwhile, the CD79B expression was absent from all the
   subclusters, which prompted the dysfunction of BCR organization and
   thus impacting B-cell function. Proportions of plasma cell subclusters
   2 and 4 exhibited significantly higher proportions in the acute than
   the recovery stage of patients ([76]Figure 2(c)). As expected, the
   expression of cell proliferation marker, MKI67, is mainly enriched in
   subclusters 2 and 4 and nearly absent from subclusters 1, 3, and 5,
   showing the proliferation of antibody-secreting cells in acute stage
   ([77]Figure 2(d)), and subclusters 2 and 4 were decreased in recovery
   stage ([78]Figure 2(c)). To explore the molecular perturbation upon
   MPXV infection, we performed differential gene expression analysis of B
   cells, including plasma cell, between patients that at acute and
   recovery stage of MPXV infection. Compared with recovery samples, the
   top 50 differently expressed genes (DEGs) in acute stage compared with
   recovery stage in at least 5 samples were shown in heatmap ([79]Figure
   2(e)) (FDR adjusted, p value <0.05 and absolute log2-based fold change
   >0.25 as the significant thresholds). Plasma cell marker (CD38) and
   memory cell marker (CD27) are more obviously upregulated in non-HIV
   patients ([80]Figure 2(e)), indicating that HIV infection may limit
   cell function and differentiation after monkeypox infection. KEGG
   pathway analysis based on clusterProfiler ([81]Figure 2(f)) and GSEA
   ([82]Figure 2(g)) pathway enrichment analysis for DEGs indicated
   significant activation of some neurodegenerative disease and metabolism
   related pathways. Notably, BCR pathway was significantly repressed in
   acute stage samples, which is consistent with the absent CD79B
   expression. Additionally, several immune response-related pathways,
   such as Leishmania infection, Graft versus host disease, and Intestinal
   immune network for IgA production, were repressed in the B cell of
   acute stage patients, indicated the inhibition role of MPXV on host
   B-cell immune defence. Moreover, oxidative phosphorylation pathway was
   consistently activated in B cell ([83]Figure 2(h)) as well as CD8^+ T
   cell and NK cell (the following section) in the acute stage,
   illustrated the influence of MPXV infection on lymphocyte metabolism,
   which should be closely associated with their immune function as
   previously reported. Mitogen activate protein kinase (MAPK) and WNT
   signalling pathways were significantly repressed in B cell of acute
   stage patients ([84]Figure 2(i and j)), indicating the defect in BCR
   signalling pathway [[85]19]. In addition, the MAPK pathway contributes
   to effector functions of BCR activation, including proliferation and
   cell survival [[86]20], so MPXV may decrease B-cell function by
   inhibiting MAPK pathway.

Figure 2.

   [87]Figure 2.
   [88]Open in a new tab

   Compositional analyses of B lymphocytes illustrate the expansion of
   subclusters. (a) UMAP visualization of B cells based on gene expression
   and different cell types represented by different colours. (b) UMAP
   visualization of B cells coloured by the expression of BCR related gene
   CD79A, CD79B, and antibody secreting related XBP1, MZB1. (c) Analysis
   of proportions of B-cell subsets in acute and recovery stages. (d)
   Violin plot representation of KI67 expression in different B-cell
   subsets (up) and different stage (down); each dot represents a cell,
   and horizontal lines display the mean expression value. (e) Heatmap of
   the most upregulated and downregulated 50 DEGs in acute stage compared
   to recovery stage coloured by the log2 (fold change). (f) Dot plot
   illustrating the top 20 significantly enriched KEGG pathways of the
   DEGs in acute stage. (g) The top 10 most activated and the top 10 most
   repressed KEGG pathway in acute stage of MPXV infection obtained
   through GSEA. (h–j) The enrichment plot of oxidative phosphorylation,
   MAPK and Wnt signalling pathways.

Characterization of CD8+ T-cell perturbation upon MPXV infection

   CD8^+ T cells play critical roles in viral clearance during infectious
   disease [[89]16]. Our clustering analysis identified nine subclusters
   for CD8^+ T cells based on canonical marker gene expressions
   ([90]Figure 3(a)). The observation of low expressions of the effective
   markers of T cells, such as IFNG, GZMB, TNF, IL2, in most of these
   subclusters, suggesting the sustained T-cell dysfunction of CD8^+ T
   cells in all the patients ([91]Figure 3(b)). We found the expanded
   CD8 T cells in recovery phase ([92]Figure 1(h)), and our further
   analysis suggests that subclusters 1, 2, 8, 9 were expanded in the
   recovery phase, but subclusters 3, 4, 5, 6, 7 were decreased
   ([93]Figure 3(c)), along with the reduction in expression of GZMB and
   KI67 ([94]Figure 3(d)), indicative of their less cytotoxic and
   proliferative potency after the virus were eliminated. HIV infection
   did not exhibit many influences on CD8^+ T-cell subcluster proportion
   in MPXV patients, only subclusters 1 and 9 proportions were augmented
   in HIV patients ([95]Figure 3(e)). Additionally, HIV patients exhibited
   slightly higher expression of GZMB, while expression of KI67 had no
   appreciable changes ([96]Figure 3(f)). To further determine the impact
   of MPXV infection on CD8^+ T-cell characterization, we next analysed
   genes and their related pathways involving MPXV infection progression.
   By using the same strategy as in B-cell analysis, we compared the CD8^+
   T cells of acute phase with recovery phase, and [97]Figure 3(g)
   illustrated 50 DEGs along with their log2-based fold change. Notably,
   we found that proliferation related genes KI67, MT2A, and HMGB2, which
   promote differentiation and maintenance of precursor exhausted T cells
   (Tpex) during chronic viral infection [[98]21] were consistently highly
   expressed in all the acute patients, including HIV-negative and
   positive patients; on the contrary, FOS and JUNB, that belong to
   transcription factor activator protein-1 (AP-1) family and associated
   with CD8^+ T cells effector function and cytokine production [[99]22],
   were inhibited ([100]Figure 3(g)). In addition, KEGG pathway analysis
   based on cluster Profiler enclosed these DEGs were enriched in
   biological processes concerning virus infection, such as coronavirus
   infection and Epstein–Barr virus infection ([101]Figure 3(h)). The KEGG
   pathway analysis through GSEA showed that anti-virus-related pathways
   were almost all repressed in the acute patients, which was consistent
   with that of B cell ([102]Figure 3(i)). We also noted the activation of
   oxidative phosphorylation in CD8^+ T cells ([103]Figure 3(j)), which
   was reported to limit cell effector function during chronic antigenic
   stimulation [[104]23]. Additionally, the toll-like receptor (TLR)
   signalling pathway ([105]Figure 3(k)), and the nod-like receptor (NLR)
   signalling pathway ([106]Figure 3(l)), which were both closely
   associated with type I IFN and inflammasome signalling pathways,
   leading to the production of corresponding proinflammatory or antiviral
   cytokines and chemokines [[107]24], were both significantly repressed
   in the acute stage patient, suggesting the comprehensive inhibition
   influence of MPXV on CD8^+ T-cell cytotoxicity. Collectively, our
   results demonstrated that CD8^+ T cells showed inhibited effector
   function in MPXV-infected patients.

Figure 3.

   [108]Figure 3.
   [109]Open in a new tab

   Heterogeneous CD8 T cell in MPXV patients. (a) UMAP embedding of CD8+ T
   cells coloured by subcluster. (b) TCR and effector genes expression
   (IFNG, GZMB, TNF, IL2) in different clusters. (c) Proportion of
   different cell subsets in acute and recovery stage. (d) GZMB and KI67
   expression in acute and recovery stages. (e) Proportion of different
   cell subsets in HIV+ and HIV– samples. (f) GZMB and KI67 expressions in
   HIV+ and HIV– patients. (g) Heatmap of top enriched 50 DEGs in CD8 T
   cells in acute stage compared with recovery stage, significantly
   upregulated genes KI67, MT2A, and HMGB2 were highlighted in red, and
   significantly downregulated gene FOS and JUNB were highlighted in blue.
   (h) The top 20 most significantly enriched KEGG pathways of the DEGs.
   (i) The top 10 most significantly activated and repressed KEGG pathways
   in the acute stage samples after MPXV infection obtained through GSEA.
   (j–l) Enrichment plot of oxidative phosphorylation, toll-like receptor
   and nod-like receptor signalling pathways.

Alterations of NK cells in the peripheral blood of patients

   Natural killer (NK) cells are key elements of innate immunity and
   impaired NK-cell function were found in MPXV patients [[110]9]. To
   further explore the changes of NK cells in the disease progression, we
   conducted NK-cell composition and molecular comparison analysis between
   the acute and recovery stages of the MPXV-infected patients. NK cells
   were divided into nine distinct subclusters based on marker genes using
   UMAP ([111]Figure 4(a)). The absence of effector gene expressions was
   observed in almost all subclusters, including IFNG and TNF ([112]Figure
   4(b)), consistent with the impaired function of NK cells in previous
   study [[113]25]. Notably, subclusters 1, 2, 3, which expressed
   relatively high level of effector genes, like NKG7 (Natural Killer Cell
   Granule Protein 7), GNLY (Granulysin), and PRF1 (Perforin 1), showed a
   higher proportion during the recovery stage, indicating the
   re-establishment of the antiviral immune response in NK cells.
   Conversely, the proportions of subsets 4, 5, 6, 7, and 8 were decreased
   during the recovery stage. ([114]Figure 4(c)). PRF1 and GNLY code
   essential effector molecules for NK-cell cytotoxicity, we noted that
   only the expression of MKI67, but not PRF1 and GNLY, was decreased in
   the recovery stage, indicating sustaining NK-cell cytotoxicity even
   after the elimination of virus ([115]Figure 4(d)). DEG analysis
   identified a total of 214 DEGs in the acute stage of MPXV-infected
   patients. Heatmaps highlight the top 50 most significantly upregulated
   and downregulated DEGs in the acute phase compared to the recovery
   phase, genes associated with chronic inflammation (CLTA) [[116]26],
   immune cell dysfunction (CENPF) [[117]27] and activation (CD38) were
   all highly expressed in almost all patients ([118]Figure 4(e)).
   Similarly, KEGG pathway analysis using clusterProfiler revealed that
   the DEGs were predominantly enriched in virus-related pathways.
   ([119]Figure 4(f)). GSEA indicated that oxidative phosphorylation, DNA
   replication, cell cycle, glycolysis were significantly activated, but
   MAPK, EGFR, and WNT signalling pathways were inhibited ([120]Figure
   4(g–j)). In virus infection such as COVID-19, the expression of MAPK
   pathway genes in NK cells was also downregulated in the cured patients,
   suggesting that decreasing NK-cell function by inhibition of MAPK
   pathway could be a common mechanism in virus infection [[121]28]. Our
   analysis suggests that NK cells play an anti-MPXV role by expressing
   cytotoxic molecules. However, HIV+ patients exhibit a decreased
   proportion of NK cells, which may hinder their recovery from MPXV
   infection.

Figure 4.

   [122]Figure 4.
   [123]Open in a new tab

   Heterogeneous NK cells in MPXV patients. (a) UMAP embedding of all NK
   cells coloured by cell subcluster. (b) UMAP projection of NK cells
   coloured by expressions of canonical markers, including NKG7, KI67 and
   effector genes expression (GNLY, PRF1, IFNG, TNF) in different cell
   subclusters. (c) Proportion of different NK cell subsets in acute and
   recovery stages. (d) Violin plot illustrating GNLY, PRF1 and KI67
   expressions in acute and recovery stage samples. (e) Heatmap showing
   the log2 (fold change) of the 50 most upregulated and downregulated
   DEGs of NK cells in acute stage compared with recovery stage. (f) Dot
   plot illustrating the top 20 most significantly enriched KEGG pathways.
   (g) The top 10 most activated and repressed KEGG pathways in the acute
   stage samples obtained by GSEA (h–j). Enrichment plot of oxidative
   phosphorylation, MAPK and WNT signalling pathways.

Characteristics of monocytes and dendritic cells

   Monocytes and dendritic cells (DC) are crucial for antiviral immune
   responses. They sense virus by multiple pathways, secret IFN and
   present antigen [[124]29]. We next deciphered their subset
   characteristics in MPXV patients by UMAP embedding of all samples and
   identified three different subclusters ([125]Figure 5(a)). Compared
   with other major histocompatibility complex (MHC) molecules, the
   absence of HLA-DMA expression indicated the inhibition of antigen
   presentation by virus ([126]Figure 5(b)). As to the proportion of
   different subclusters in different stages, subcluster 1 decreased in
   the recovery stage, but subclusters 2 and 3 showed opposite change,
   showing that MPXV may impact their function differentiation
   ([127]Figure 5(c)). Notably, when stratified by HIV negative or
   positive, we found in HIV+ patients, subcluster 1 was elevated and
   subset 2 decreased ([128]Figure 5(d)), indicating the critical impact
   of HIV on the innate immune cells. Additionally, most of the antigen
   presenting related genes (APGs) exhibited obviously higher mRNA level
   in MPXV infection acute stage compared with recovery stage ([129]Figure
   5(e)), indicating increased antigen presentation, but HIV+ patients did
   not show obvious change of their expression ([130]Figure 5(f)). For the
   significant influence of MPXV infection on APGs, we next analysed gene
   expressions of acute phase compared with recovery phase of MPXV
   infection and identified 42 DEGs in at least 5 acute stage patients
   ([131]Figure 5(g)). We found EGR1 is among the top upregulated genes in
   monocytes. EGR1 binds enhancers of genes governing monocytic
   development, as well as a large set of inflammatory genes [[132]30]. We
   observed genes correlated with anti-virus function were upregulated in
   acute phase compared with recovery phase ([133]Figure 5(g)). For
   instance, CD14 is a pattern recognition receptor (PRR) that enhances
   innate immune responses and CD14 upregulation on monocytes is
   associated with effective control and clearance of viral and bacterial
   [[134]31]. IFI27, a classical type-1 IFN related gene and is involved
   in several types of viral infections [[135]11], and proinflammatory
   cytokines CCL3, which was previously reported to be produced by
   circulating monocytes in severe COVID-19 [[136]27], were also
   upregulated in acute stage ([137]Figure 5(g)). We also analysed the
   enrichment of KEGG pathways of DEGs through cluster Profiler
   ([138]Figure 5(h)) and found several pathways related to infection and
   inflammation were all significantly activated in acute stage of MPXV
   infection ([139]Figure 5(i–k)). Our results suggest that mono/DCs were
   activated during anti-MPXV function, while HIV changes mono/DCs
   dynamics in monkeypox infection, which may inhibit the anti-MPXV
   immunity.

Figure 5.

   [140]Figure 5.
   [141]Open in a new tab

   Heterogeneous monocytes and dendritic cells in MPXV patients. (a) UMAP
   embedding of all mono/DCs cells coloured by cell subcluster. (b) UMAP
   projection of mono/DCs coloured by canonical antigen presentation genes
   expression. (c) Proportion of different mono/DCs subsets in acute and
   recovery stage, or (d) HIV+ and HIV– patients. (e) Antigen presentation
   genes expression in acute and recovery stages, or (f) HIV+ and HIV–
   patients. (g) Heatmap of the log2 (fold change) of significantly
   upregulated and downregulated DEGs in mono/DCs between in acute stage
   compared with recovery stage samples. IF127 was identified as type-1
   IFN gene correlated with MPXV infection and was highlighted in red. (h)
   Dot-plot illustrating the top 20 most significantly enriched KEGG
   pathways of the DEGs. (i) The significantly activated or repressed KEGG
   pathways in the acute stage samples identified through GSEA. (j–k)
   Enrichment plot of leishmania infection and systemic lupus
   erythematosus.

Neutrophils in the peripheral blood of patients

   Neutrophils play critical roles in viral clearance during infections,
   they release neutrophil extracellular traps and activate adaptive
   immune response, while the overactivation of neutrophils may cause
   tissue damage and lead to poor outcomes [[142]32,[143]33]. We extracted
   the neutrophil UMAP embedding from the whole UMAP space which contained
   eight neutrophil subclusters ([144]Figure 6(a)). The visualization of
   several neutrophil developmental markers illustrates highly
   heterogeneous developmental stages of neutrophil in our samples
   ([145]Figure 6(b)). Specifically, FCGR3A and SELL, two markers of
   neutrophil maturation, exhibited obviously higher mRNA levels in acute
   stage of MPXV infection, illustrating the acceleration of neutrophil
   maturation induced by MPXV, which might be associated with viral
   elimination. Additionally, CXCR4, a marker of neutrophil senescence,
   was inhibited in the acute stage of MPXV infection, further validating
   the enhanced function of neutrophil upon MXPV infection ([146]Figure
   6(c)). In terms of the subcluster composition dynamic, we found that
   subcluster 6, which represent the only one with MKI67 expression, was
   significantly decreased in the recovery stage ([147]Figure 6(d)),
   indicating that neutrophil should be at quiescent state after the
   elimination of MPXV. In addition, with the recovery of the MPXV
   infection, subsets 2, which highly expressed IFI27 was also decreased
   ([148]Figure 6(d)). DEGs analysis obtained 17 genes in at least 5 acute
   stage patients. We observed that the expressions of MK167 and STMN1,
   which control cell proliferation and cell cycle [[149]34], were
   upregulated in neutrophils in nearly all patients in acute phase
   compared to recovery phase. We also found that MPXV infection triggered
   a robust IFN-I response in all patients, characterized by
   interferon-stimulated genes (ISGs) IFI27 and IFIT3 expression, but the
   upregulation was more significant in neutrophils from non-HIV patients
   compared to HIV patients ([150]Figure 6(e)). So, in HIV+ patients IFI27
   mediate ISGs expression and IFN-I production could be inhibited during
   MPXV infection ([151]Figure 6(e)). Some inflammatory responses as well
   as cytotoxic related genes, including IL32, GZMA, GZMB, were
   consistently upregulated. On the contrary, CCL3, which has been
   previously reported to inhibit cell differentiation, was significantly
   suppressed in acute stage ([152]Figure 6(e)), which was predictive of
   the induction of differentiation and cytotoxicity of neutrophil. CD83
   was a marker of polymorphonuclear neutrophils (PMN) with
   characteristics of dendritic cells and served as a marker of acute
   inflammation [[153]35]. Previous study indicated that herpesviruses
   could acquire immune evasion by targeting CD83 [[154]36,[155]37].
   Similarly, in our study, the expression of CD83 was also inhibited in
   all patients ([156]Figure 6(e)). KEGG pathway analysis of the 17 DEGs
   based on cluster Profiler revealed significant enrichment of strong
   antiviral pathways, such as COVID-19 signalling, cytokine signalling,
   and neutrophil extracellular trap formation ([157]Figure 6(f)). What’s
   more, GSEA also uncovered significant activation of cell cycle and
   immune response-related pathways ([158]Figure 6(g)) in the acute stage
   of MPXV infection. Collectively, we illustrated the neutrophil
   differentiation and ISGs expression could be inhibited in HIV+
   patients.

Figure 6.

   [159]Figure 6.
   [160]Open in a new tab

   Heterogeneous neutrophils in MPXV patients. (a) UMAP embedding of all
   neutrophils coloured by cell subcluster. (b) UMAP projection of all
   neutrophils coloured by canonical neutrophils marker genes expression.
   (c) Expression of neutrophils maturation related genes in acute and
   recovery stage samples. (d) Proportion of different neutrophil subsets
   in acute and recovery stage samples. (e) Heatmap illustrating the log2
   (fold change) of the most upregulated and downregulated DEGs in
   neutrophils between the acute and recovery stage samples. (f) Dot plot
   illustrating the top 20 most significant KEGG pathways of the DEGs. (g)
   KEGG pathways that were significantly activated or repressed in the
   acute stage samples obtained through GSEA.

Interactions of immune cells

   Cell interactions are known to play an important role in infectious
   diseases. Using CellphoneDB analyses with scRNA-seq data from monkeypox
   patients, we found intense cell interactions between mono/DC and
   neutrophils with other cells, and the interactions were more obvious in
   the acute stage, indicating a critical role of mono/DC and neutrophils
   in disease progression ([161]Figure 7(a)). When we analysed
   communication network, we found strong interactions between mono/DC and
   CD4 T cells, as well as neutrophils and CD4 T cells both in acute and
   recovery stages ([162]Figure 7(b)). Ligand–Receptor (L–R) pairs between
   each immune cell type were also illustrated in both acute and recovery
   stage samples ([163]Figure 7(c)). Notably, LGALS9-HAVCR2 mediated
   interaction was only found between mono/DC and CD4 T cells, as well as
   CD4 T cells and B cells in the acute stage of MPXV infection,
   indicating the unique role of HAVCR2 in CD4 cell communication
   ([164]Figure 7(c)). We also noted strong interaction of mono/DC and
   neutrophils with several cells by LGALS9-CD44/CD45, which were
   responsible for interactions of many types of cells in this study.
   Notably, LGALS9-CD44/CD45 pairs were related to the interactions
   between CD4+ T and CD8+ T cells in only recovery but not acute samples.
   LGALS9 was reported to induce CD8+ T-cell death [[165]38] and by
   interacting with CD44, it could promote Treg function [[166]39], thus
   playing critical role in limiting T-cell response and considered as an
   immune checkpoint, so we could speculate that LGALS9 could be a
   critical factor for limiting T-cell anti-virus response after MPXV
   elimination. Additionally, the outgoing and incoming interaction
   profile analysis of each cell type also supported the stronger
   interactions exist in acute stage ([167]Figure 7(d)). Mono/DC and
   neutrophils mediated strongest interactions both in acute and recovery
   stages, there was nearly no difference between acute and recovery
   samples in terms of the incoming and outgoing signals of cell
   interactions ([168]Figure 7(e,f)). In brief, galectin, annexin, IL16,
   and CCL represented the top four molecules that were responsible for
   cell interactions ([169]Figure 7(e,f)). This suggests innate immune
   cells are critical for MPXV elimination, and our results shed light on
   identifying potential targeting cell interactions to improve viral
   elimination.

Figure 7.

   [170]Figure 7.
   [171]Open in a new tab

   Cell interaction analysis of immune cells in patients. (a) Heatmap
   showing the number of potential ligand-receptor pairs among the
   predicted cell types. (b) Interaction network constructed by cellchat.
   Each line colour indicates the ligands expressed by the cell population
   represented by the same colour (labelled). The lines connect to the
   cell types that express the cognate receptors. The line thickness is
   proportional to the number of ligands when cognate receptors are
   present in the recipient cell type. (c) Dot plot showing the
   ligand–receptor pairs that link the interacted cells. The x-axis and
   y-axis represent the cell pairs and ligand–receptor pairs,
   respectively. The colour from blue to red indicates a gradient stronger
   interaction. (d) Scatter plot illustrates the number of outgoing
   signals, i.e. ligand, and incoming signal, i.e. receptor, for each cell
   type in cell–cell interactions. (e) and (f) illustrate the number of
   incoming and outgoing signals in cell–cell interactions that each
   pathway contributed in acute stage and recovery stage samples.

Influences of MPXV and HIV infections on immune cell differentiation

   To investigate the potential influence of MPXV on immune cell
   differentiation and assess whether HIV infection might affect this
   process, we performed pseudotime analysis of B, CD4^+ T, CD8^+ T, NK,
   Mono/DC, and neutrophil cells (see Supplemental Figures 1–6).
   Expression patterns of key genes associated with cell differentiation
   and maturation were analysed along the pseudotime trajectory to detect
   the cell differentiation process. Significant enrichment was observed
   in antibody-secreting B cells, marked by the upregulation of MZB1 (see
   Supplemental Figure 1a); proliferating and cytotoxic CD8+ T cells with
   high expression of GZMB and MKI67 (see Supplemental Figure 3a); Mono/DC
   subsets with elevated expression of MHC molecules (see Supplemental
   Figure 5a); and neutrophils showing increased expression of CXCR4,
   FCGR3A, and MKI67 (see Supplemental Figure 6a) in PBMC samples during
   the acute MPXV infection stage. The influence of HIV infection on
   immune cell differentiation appeared to be minimal, possibly due to the
   clearance of HIV, as the distribution of cells from HIV– and HIV+
   patients was nearly even across the pseudotime trajectory. These
   findings suggest that MPXV accelerates immune cell activation and
   maturation.

Discussion

   In this study, we utilized single-cell RNA sequencing (scRNA-seq) to
   analyse the immune cell atlas of peripheral blood from monkeypox
   patients at both acute and recovery stages, the potential impact of HIV
   was also proposed ([172]Figure 8). Our analysis revealed dysregulated
   cell subsets, function, and robust cell interactions during MPXV
   progression. Furthermore, our study elucidated the possible and complex
   effect of HIV infection on the dynamics of several cell subsets in
   monkeypox infection, which may contribute to their decreased antiviral
   response and longer recovery time.

Figure 8.

   [173]Figure 8.
   [174]Open in a new tab

   Diagram illustrating the proposed mechanism underlying the severe
   condition of HIV-infected patients following MPXV infection. HIV
   induces CD4+ cell loss and dysfunction, impairs BCR signalling and
   humoral immunity in B cells, and reduces type 1 IFN signalling in
   neutrophils. The complex and diverse cellular interactions further
   exacerbate deficiencies in anti-MPXV immunity, contributing to
   prolonged hospitalization in HIV+ patients.

   The deficiency of CD4^+ T cells is a hallmark of HIV infection. There
   is a greater chance of severe manifestation of MPXV infection in
   immunosuppressed patients [[175]40], while the mechanism is unclear.
   Besides infecting T cells, HIV also induces the death of uninfected
   CD4 T cells. The mechanisms include the overexpression of death
   ligands, activation-induced cell death, and other processes, which have
   been extensively reviewed elsewhere [[176]41,[177]42]. According to
   U.S. Centers for Disease Control and Prevention (CDC), “Of the people
   with severe manifestations of monkeypox for whom CDC has been
   consulted, the majority have had HIV with CD4 counts <200 cells/ml,
   indicating substantial immunosuppression”. Indeed, in our study, we
   found HIV patients have lower CD4 T cell proportion compared with HIV–
   ones, especially in acute phase. So, in the current study, we mainly
   discussed the possible role of other cell subsets’ role in HIV-induced
   decreased anti-MPXV immune.

   Previous study indicates that HIV induces alterations of B-cell
   subpopulations, most B cells in healthy individuals are naive and
   memory B cells, whereas in HIV+ individuals, immature transitional B
   cells, exhausted B cells and plasma blasts are over-represented. In
   addition, HIV causes B-cell exhaustion [[178]43]. Notably, as shown in
   the heatmap of DEGs in acute phase versus recovery phase, the
   perturbations in B cells subsets is most significant in our study,
   upregulation of many genes critical in B-cell function in HIV– patients
   in acute phase was not seen in HIV+ patients, such as CD38, which is
   critical for B-cell activation, proliferation, and differentiation
   [[179]44]. Moreover, we found the absence of CD79B in all the patients
   which indicates the dysregulation of BCR signalling. Receptors initiate
   signalling through pathways and the central role of BCR in humoral
   immunity has been acknowledged [[180]45,[181]46]. High levels of many
   B-cell inhibitory receptors, including CD22, CD72 were reported to be
   expressed by tissue-like memory B cells from HIV+ individuals, and
   tissue-like memory B cells were recognized as an exhausted B-cell
   counterpart [[182]47], so we postulate that HIV impair humoral
   responses and long-term immune memory, which may responsible for severe
   condition of HIV+ patients after MPXV infection.

   Our analysis of total PBMCs identified decreased levels of NK cells in
   HIV+ patients during both the acute and recovery phases compared to
   HIV– patients. NK cells are important mediators of the innate immune
   defenders against viruses, mediate suppression of viral in many ways.
   The crucial role of NK cells and their dysfunction in antiviral
   immunity has been well established [[183]2,[184]48], including in
   monkeypox [[185]9,[186]49]. MPXV impairing NK-cell function by
   preventing its IFN-γ production [[187]9], activation, chemotaxis, and
   antibody-dependent cell-mediated cytotoxicity (ADCC) are also defective
   in HIV-infected individuals [[188]50,[189]51], which may further
   compromises host immune defences against MPXV. Our study provides
   additional evidence that HIV-induced NK-cell dysfunction may contribute
   to a decrease in protective immunity against the monkeypox virus.

   For CD8 T cells, several studies have documented their impaired
   function in HIV infection, these cells become non-cytolytic, exhibit
   lower levels of perforin, became exhausted and apoptosis
   [[190]52,[191]53]. While in our study, we found CD8 T cells increase in
   recovery phase in HIV+ patients compared with HIV– ones. Studies have
   also indicated that naïve and memory CD8 T cells turnover increases
   during HIV infection and differentiate towards senescence state
   [[192]53]. So the increased CD8 in HIV+ patients may be senescent T
   cells. Another critical question that remains to be answered is when
   T-cell exhaustion starts during HIV infection and how anti-virus
   treatment affects T-cell subsets, these may also cause the variation of
   CD8 T cells number. Since the severe condition and longer hospitality
   of HIV+ patients, the increased CD8 T cells may not promote the
   clearance of MPXV. Moreover, it is important to note that the MPXV/HIV
   co-infected patients included in our study had received prolonged
   antiviral therapy, achieving viral suppression. This could potentially
   explain why the presence of HIV does not significantly affect the
   immune cell subsets during the acute and recovery phases of MPXV
   infection, to elucidate the mechanism and role of the increased CD8 T
   cells in HIV+ patients upon MPXV infection, more mechanism study is
   needed.

   Recent studies have highlighted the potential of interferon-responsive
   neutrophil subsets in combating infections and patrolling the
   bloodstream. Following viral infections, there is often a sharp
   increase in the number of neutrophils in the local microenvironment.
   However, these neutrophils have been reported to produce Reactive
   Oxygen Species (ROS) that can damage the epithelial–endothelial barrier
   and impair virus elimination, thereby contributing to detrimental
   inflammation and increased morbidity in mice [[193]54,[194]55]. In our
   study, we observed a decrease in the proportion of neutrophils in the
   recovery stage among HIV-negative patients, whereas their percentage
   remained higher in both stages among HIV-positive patients.
   Additionally, the expression of IFGs in neutrophils was inhibited in
   HIV-positive patients, potentially exacerbating the condition of
   monkeypox patients. Nevertheless, the precise pathogenic roles of this
   neutrophil subset require further exploration through future studies
   investigating its phenotype and role in antiviral responses.

   Mono/DCs is the first line of defence against viral infections and are
   primary targets for viral assault [[195]2]. Innate immune cells
   response to HIV during the early stages of infection is an important
   contributor to persistent immune activation and chronic inflammation,
   causing the activation or immunosenescence of immune cells, and
   activated CD16^+ monocytes can disseminate HIV replication through
   transferring virions and effective infection of CD4 T cells [[196]56].
   Though no obvious changes in HLA gene expression were observed in HIV+
   and HIV– patients, dynamic change of Mono/DC subsets 1 and 2 showed
   differences between HIV-positive and HIV-negative patients from the
   acute phase to the recovery phase, indicating the persistent impact of
   HIV on Mono/DC function in MPXV infection, beyond its role in causing
   abnormal phagocytic function. In addition, we identified IFI27 as the
   most significantly upregulated ISG in both monocytes and neutrophils in
   Monkeypox infection. This highlights the pivotal role of IFI27 in the
   immune response against viral infections. Interestingly, IFI27 has been
   reported to interact with RIG-I, thereby impairing RIG-I activation and
   subsequently downregulating innate immune responses. Consequently,
   IFI27 provides a negative feedback mechanism that counteracts excessive
   inflammatory responses to RNA viral infections [[197]57]. Our results
   suggest that besides contributing to the cytokine production and
   inflammation, peripheral monocytes may also counteract excessive
   inflammatory monkeypox infection.

   We observed the upregulation of the oxidative phosphorylation pathway
   in several cell subsets in acute phase, alongside the inhibition of
   anti-virus-related signalling pathways such as MAPK and WNT compared to
   recovery phase. Metabolism correlates with cell subsets and activation
   states. For instance, T cells transit from predominantly oxidative
   metabolism to aerobic glycolysis upon activation. Alfredo et al.
   indicated that B cells chronically exposed TLR4 ligation showed
   enhanced and more rapid metabolic reprogramming towards aerobic
   glycolysis [[198]58]. The increased oxidative phosphorylation in our
   study may result from the difference of dominance B-cell subsets or the
   different stimulation type, while the detailed role of mitochondrial
   metabolism in B-cell activation and function, especially in virus
   infection remains to be further detected. MAPK cascade participates in
   regulating the immune response, but its role in viral is controversial
   [[199]59]. For example, in SARS-CoV-2 infection, TNF-α (tumour necrosis
   factor alpha), an important cytokine for proliferation and function of
   NK cells exerts its effects through MAPK pathways [[200]28]. Given its
   roles in type I IFN production, which is crucial in early innate
   responses against viral infections, the inhibition of the MAPK pathway
   genes indicates the significant impact of Monkeypox against host immune
   system. On the contrary, viruses extensively use MAPK cascade also
   implies that this pathway may be a good target for developing
   broad-spectrum antiviral drugs, for instance, p38 MAPK inhibitor,
   SB203580, inhibited the effective phosphorylation of HSP-27, CREB, and
   eIF4E in SARS-CoV-infected cells, showing promise as a new class of
   antiviral agents [[201]60]. Although disturbed MAPK signalling was
   observed in NK and B cells during the acute phase, we cannot
   definitively conclude the role of MAPK or the potential of its
   inhibitors as antiviral drugs without further research.

   The interaction of monocytes with T cells in Monkeypox patients might
   compromise cell anti-virus function. Interaction analysis indicates
   that monocytes interact with other cells largely through LGALS9. LGALS9
   is a widely expressed protein involved in immune regulation, affecting
   prognosis of various types of cancer as an immune checkpoint inhibiting
   T-cell function [[202]61]. Notably, in the SARS-CoV-2 infection, the
   serum level of LGALS9 was elevated and positively correlated with
   inflammation and tissue injury [[203]62]. Recent studies indicate that
   targeting monocytes infiltration could alleviate virus-related syndrome
   in COVID-19 infection [[204]63]. Our study suggests that LGALS9 may be
   the primary molecule mediating cell interactions in monkeypox
   infection, offering novel insights into understanding the dynamic
   interplay of immune cells in this context. Inhibiting LGALS9 holds
   promise for decreasing overactivated inflammation in such infections.

   Our study also has some limitations, first, the limited sample size and
   sequencing depth is a major shortcoming of this study, potentially
   affecting the robustness of the results and leading to biased or less
   representative conclusions. For example, although we observed decreased
   NK cells in HIV-positive patients compared to HIV-negative ones in
   acute phase of MPXV, we were unable to identify a statistically
   significant difference in our analysis. In addition, B cells subset 4
   shows the highest Ki67 expression and may exert unique anti-virus
   function, and biological processes and signalling pathways associated
   with T-cell dysfunction should be further analysed. Further study is
   needed to provide more sufficient resolution into the various cell
   compartments and interaction. In addition, our conclusion in the
   current study is mainly drawn from scRNA-seq data, clinical validations
   will be carried out in our further research.

   Overall, our work proposes the potential mechanisms underlying the
   severe condition of HIV-bearing MPXV patients ([205]Figure 8) and
   provides understanding of pathogenesis in HIV patients co-infected with
   MPXV, thus deepening current understanding of the MPXV infection. Our
   data identifies potential mechanisms of immunocompromised individuals,
   such as HIV-infected patients, being more probably to experience severe
   manifestations. Individuals with acquired immunodeficiency may require
   more active medical intervention rather than spontaneous resolution.

Material and methods

Subjects and specimen collection

   From January to June 2023, peripheral blood samples were collected from
   six monkeypox patients at Tianjin Second People’s Hospital after
   obtaining written informed consent from the patients (Tianjin Second
   People’s Hospital Ethics Committee Approval No. 2023-075). The sample
   collection criteria for this study are defined as follows: during the
   acute phase, samples were obtained in the febrile and eruptive stages,
   while in the recovery phase, samples were collected on the day complete
   detachment of rash scabs. All scRNA-seq samples from the enrolled
   patients were collected during the acute phase of infection, prior to
   the initiation of antibiotic treatment. Samples collected during the
   recovery phase were obtained at least 1 week after the completion of
   antibiotic therapy to avoid the impact of antibiotics on the immune
   profiles in our study. Eligibility criteria included age >18 years and
   hospitalization at Tianjin Second People’s Hospital with a positive
   monkeypox virus RT–PCR test from vesicular fluid. All enrolled patients
   provided informed consent. Blood samples were processed by isolating
   PBMCs using Ficoll-Paque Plus medium (GE Healthcare) through density
   gradient centrifugation and washed with Ca/Mg-free PBS. All blood
   samples were processed within 4 h of collection, with the majority
   processed within 1 h.

scRNA sequencing by Seq-Well

   The Seq-Well platform for scRNA-seq was utilized as described
   previously11,12,42. Immediately after Ficoll separation, 50,000 PBMCs
   were resuspended in RPMI + 10% FCS at a concentration of 75,000 cells
   per ml. A 200-μl volume of this cell suspension (15,000 cells) was then
   loaded onto Seq-Well arrays pre-loaded with mRNA capture beads
   (ChemGenes). Following four washes with Dulbecco’s phosphate-buffered
   saline (DPBS) to remove serum, the arrays were sealed with a
   polycarbonate membrane (pore size of 0.01 μm) for 30 min at 37°C and
   then frozen at −80°C for no less than 24 h and no more than 14 days to
   allow batching of samples processed at irregular hours. Next, arrays
   were thawed, cells lysed, transcripts hybridized to the mRNA capture
   beads, and beads recovered from the arrays and pooled for downstream
   processing. Immediately after bead recovery, mRNA transcripts were
   reverse-transcribed using Maxima H-RT (ThermoFisher, EPO0753) in a
   template-switching-based rapid amplification of cDNA ends (RACE)
   reaction, excess unhybridized bead-conjugated oligonucleotides were
   removed with exonuclease I (NEB M0293L) and second-strand synthesis was
   performed with Klenow fragment (NEB M0212L) to enhance transcript
   recovery in the event of failed template switching. The whole
   transcriptome amplification (WTA) was performed with KAPA HiFi PCR
   Mastermix (Kapa Biosystems KK2602) using ∼6000 beads per 50-μl reaction
   volume. Resulting libraries were then pooled in sets of six (∼36,000
   beads per pool) and products purified by Agencourt AMPure XP beads
   (Beckman Coulter, A63881) with a 0.6× volume wash followed by a 0.8×
   volume wash. The quality and concentration of WTA products were
   determined using an Agilent Fragment Analyser (Stanford Functional
   Genomics Facility), with a mean product size of >800 bp and a
   non-existent primer peak indicating successful preparation. Library
   preparation was performed with a Nextera XT DNA library preparation kit
   (Illumina FC-131-1096) with 1 ng of pooled library using dual-index
   primers. Tagmented and amplified libraries were again purified by
   Agencourt AMPure XP beads with a 0.6× volume wash followed by a 1.0×
   volume wash, and quality and concentration were determined by fragment
   analysis. Libraries between 400 and 1000 bp with no primer peaks were
   considered successful and pooled for sequencing. The sequencing was
   performed on a NovaSeq S2 instrument (Illumina; Chan Zuckerberg
   Biohub). The read structure was paired-end with read 1 beginning from a
   custom read 1 primer11 containing a 12-bp cell barcode and an 8-bp UMI,
   and with read 2 containing 50 bp of mRNA sequence.

Bioinformatics analysis

   The raw sequencing data were preprocessed by CellRanger v8.0, mainly
   including the removal of low-quality reads, alignment to the reference
   genome, i.e. GRCh38 in this study, and generation of gene-cell
   expression matrix. Seurat v5 was then used for the integration of data
   from different samples, dimension reduction, cell clustering and
   differential expression analysis by using the raw expression matrix as
   input. In detail, quality control of the raw data, mainly including
   removal of cells that contained less than 200 or more than 4000 genes
   as well as those with more than 25% mitochondrial genes. The genes that
   expressed in no more than three cells were also removed. Scaled
   normalization of gene expression was conducted by using the
   NormalizeData function. The normalized data from different samples were
   integrated by anchors screened by FindIntegrationAnchors function to
   remove the batch effect. Data dimension reduction was performed by
   three methods, i.e. PCA, UMAP and TSNE, and cell clustering was
   performed based on the low-dimension data by using the FindClusters
   function with the most optimal cluster number identified by cluster R
   package. Differential expression analysis between different cell
   clusters or patient groups was conducted based on the FindMarkers
   function with the thresholds of FDR adjusted p value <0.05 and absolute
   log2-based fold change >0.25. Besides, only genes that expressed in at
   least 10% cells of each sample group were retained for differential
   expression analysis. Cell identity was manually annotated based on the
   canonical markers as well as the specifically expressed genes of each
   cell cluster. Both CellPhenoDb and cellchat were applied for the
   cell-cell interaction analysis with the default parameters.

Supplementary Material

   supplementary.docx
   [206]TEMI_A_2459136_SM6310.docx^ (1.7MB, docx)

Acknowledgements