DNA prime‐protein boost vaccine encoding HLA‐A2, HLA‐A24 and HLA‐DR1 restricted epitopes of CaNA2 against visceral leishmaniasis

Summary

Visceral leishmaniasis is a tropical and neglected disease with an estimated 200 000–400 000 cases and 60 000 deaths worldwide each year. Currently, no clinically valid vaccine is available for this disease. In this study, we formulated DNA and protein vaccines encoding HLA‐A2, HLA‐A24 and HLA‐DR1 restricted epitopes of CaNA2 against visceral leishmaniasis. We predicted the secondary and tertiary structures, surface properties, subcellular localization, potential binding sites and HLA‐A2, HLA‐A24 and HLA‐DR1 restricted epitopes of CaNA2. The best candidate CpG ODN (2395, M362, D‐SL03 or 685) was screened out as a DNA vaccine adjuvant. We also prepared Kmp‐11 and Kmp‐11/CaNA2 DNA and protein vaccines, respectively, for comparison. BALB/c mice were immunized with a DNA prime‐protein boost immunization strategy and challenged with a newly isolated Leishmania strain from an individual with visceral leishmaniasis. The IgG antibody titers showed that our vaccine had strong immunogenicity with a long duration, especially cellular immunity. The spleen parasite burden of each group demonstrated that the CaNA2 vaccine had a certain immune protective effect on visceral leishmaniasis in BALB/c mice, and the amastigote reduction rate reached 76%. Preliminary safety tests confirmed the safety of the vaccine. Our work demonstrates that the HLA‐A2, HLA‐A24 and HLA‐DR1 restricted epitope CaNA2 DNA prime‐protein boost vaccine may be a safe and effective epitope vaccine candidate against visceral leishmaniasis.

Abbreviations

ALT

alanine aminotransferase

CaNA2

serine/threonine protein phosphatase 2B catalytic subunit A2

CaN

calcineurin

CpG ODNs

CpG oligodeoxynucleotides

ELISA

enzyme‐linked immunosorbent assay

H₂O₂

hydrogen peroxide

IFN‐γ

interferon‐γ

IgG

immunoglobulin G

IL‐10

interleukin‐10

IL‐6

interleukin‐6

KMP‐11

kinetoplastid membrane protein‐11

LDU

Leishman Donovan Units

NK

natural killer

NO

nitric oxide

RT‐PCR

reverse transcription‐polymerase chain reaction

TLR9

Toll‐like receptor 9

TNF‐α

tumor necrosis factor‐α

VL

visceral leishmaniasis

Introduction

Visceral leishmaniasis (VL), also called Kala‐azar, is a systemic disease caused by a complex of species comprising Leishmania donovani, Leishmania infantum and Leishmania chagasi (Leishmania donovani complex). The species that causes VL in Asia and eastern Africa is mainly L. donovani, whereas the primary species in Europe, Latin America and North Africa are L. infantum or L. chagasi.1 The symptoms of VL include a long‐term high fever, hepatosplenomegaly, pancytopenia and hypergammaglobulinemia, and the disease is almost always fatal without appropriate treatment.2 At present, VL is considered a tropical parasitic disease that cannot be ignored and ranks second in mortality and fourth in morbidity among tropical diseases.3 In China, the disease is prevalent in the Xinjiang, Gansu, Sichuan, Shaanxi, Shanxi and Inner Mongolia provinces. From 2002 to 2011, a total of 3169 cases were officially reported nationwide, with a range from 140 to 509 cases per year.4 Control of VL has been hampered by the lack of an effective vaccine, the limitations and toxic side effects of drugs, and the emergence of drug‐resistant strains.5

Leishmania is an intracellular parasitic protozoan that is transmitted by its vector sandfly. The parasite's life cycle involves proliferative promastigotes in the sandfly, non‐dividing metacyclic forms before inoculation into the vertebrate host and phagocytosis by macrophages and amastigotes in the phagolysosome of human macrophages, leading to macrophage lysis and serial infection of other macrophages.6 After Leishmania infection, macrophages, which are pivotal for cellular immune responses, present and process antigens and produce a variety of cytokines.7 However, Leishmania has evolved to evade the defense mechanism of macrophages through inhibiting their activation, which enables the parasite replication and survival in the host.7 The life cycle of the parasite and the immune escape mechanism have introduced many challenges into the development of an effective vaccine against the disease. In this study, we use serine/threonine protein phosphatase 2B catalytic subunit A2 (PP2B‐A2 or CaNA2) and kinetoplastid membrane protein‐11 (KMP‐11) of Leishmania to develop a vaccine against VL. PP2B or calcineurin (CaN) of Leishmania is a Ca²⁺‐dependent and calmodulin‐dependent phosphatase that was first reported in 1999.8 This protein consists of two subunits: the CaN A subunit is the catalytic core of the holoenzyme, whereas the CaN B subunit increases the activity of subunit A.9 PP2B is involved in a number of different signaling pathways and so participates in some physical activities. In Trypanosoma, Cyclophilin–trialysin synergistically acts on parasites to activate calcineurin phosphatase signaling, which mediates cAMP recognition and evasion and adaptation of the parasite.10 Orrego et al.11 demonstrated that TcCaNA2 might play a key role in host cell invasion by Trypanosoma cruzi. In Leishmania, calcineurin interacts with other biomolecules, such as heat‐shock proteins, to provide suitable thermotolerance and virulence in Leishmania major.12 Naderer et al.9 demonstrated that a Ca²⁺ influx and calcineurin signaling activation were required for differentiation and adaptation of the parasite to the cellular stress encountered during infection of the mammalian host. These studies suggest that calcineurin has potential as a vaccine or treatment target against the parasite. However, few related studies have been performed in Leishmania. This is the first time this protein has been used as a vaccine against VL. KMP‐11, which is a highly conserved surface membrane protein present in all kinetoplastid protozoa,13 is considered a potential vaccine candidate for leishmaniasis. This protein has strong antigenicity for murine and human T cells and is capable of stimulating both innate and adaptive immune responses.14

The gene frequency of HLA class I A2 (A*02:01) is high among several ethnic groups, including Asians and Caucasians, whereas the gene frequency of HLA class I A24 (A*24:02) also comprises a large portion of the world population, especially in Asian populations.15, 16 In most humans, the most highly expressed HLA class II molecule on antigen‐presenting cells is HLA‐DR.17 HLA‐DR molecules bind microbial peptides in an endosomal compartment and present them on the cell surface for CD4⁺ T‐cell surveillance.18 HLA‐A2, HLA‐A24 and HLA‐DR1 are the major representatives of the HLA class I and II supertypes. Corresponding HLA‐restricted epitope vaccines have been reported and developed for treatment of tumors and prevention of infectious diseases, including leishmaniasis.19, 20, 21

CpG oligodeoxynucleotides (CpG ODNs) are agonists of Toll‐like receptor 9 (TLR9) and have potential as vaccine adjuvants due to their strong immune‐stimulating activities. Activation of TLR9 is MyD88‐dependent and eventually leads to activation of interferon regulatory factor 7 and nuclear factor‐κB.22 Interferon regulatory factor 7 activation results in secretion of type I interferons, whereas nuclear factor‐κB activation promotes pro‐inflammatory cytokine production and dendritic cell maturation and survival.23 CpG ODNs are divided into classes A, B and C. Class C has the functions of both classes A and B such as inducing interferon‐α secretion by plasmacytoid dendritic cells and thereby indirectly activating natural killer cells, and strongly stimulating B and natural killer cells to activate and secrete cytokines.24 Therefore, we used four class C CpG ODNs (2395, M362, D‐SL03 and 685) in this study.

In this study, we selected HLA‐A2, HLA‐A24 and HLA‐DR1 restricted epitopes of CaNA2 to develop a DNA prime‐protein boost vaccine against VL and prepared Kmp‐11 and Kmp‐11/CaNA2 DNA and protein vaccines for comparison. The secondary and tertiary structures, surface properties, subcellular localizations and potential binding sites of CaNA2 and KMP‐11 were simulated. The HLA‐restricted epitopes of CaNA2 were predicted using four online analysis systems (SYFPEITHI,25 NetCTL 1·2,26 NetMHC 4·027 and Rankpep28). The best candidate of the four class C CpG ODNs (2395, M362, D‐SL03 and 685) was selected as an adjuvant for the DNA vaccine. Eukaryotic recombinant plasmids of target genes with CpGs were constructed and encapsulated by Lipofectamine as DNA vaccines. Prokaryotic recombinant plasmids of the target genes were expressed in Escherichia coli and purified as protein vaccines. BALB/c mice were immunized with the DNA prime‐protein boost immunization strategy and challenged with a Leishmania strain that was newly isolated from a VL patient in Sichuan, China. We dissected the immunogenicity, protective immunity and safety of our vaccines. Subsequent studies were performed to assess whether the HLA‐restricted epitopes of the CaNA2 vaccine could effectively prevent VL.

Materials and methods

Isolation, culture and identification of Leishmania

This study was approved by the Sichuan University Medical Ethics Committee (Approval Number: K2018056). Bone marrow puncture fluid with Leishmania amastigotes was obtained from a VL patient at West China Hospital, Sichuan University, China. Three laboratory golden hamsters (Mesocricetus auratus, 8 to 10 weeks old, female) were purchased from the Chengdu Institute of Biological Products Co., Ltd, Chengdu, China. We inoculated the golden hamsters with the patient's marrow puncture fluid (0·3 ml/hamster) by intraperitoneal injection. Five months later, the golden hamsters were euthanized. Their livers and spleens were made into imprints and stained to observe the Leishmania infection. Their spleens were also homogenized and added to M199 medium (HyClone, Logan, Utah, USA) to culture Leishmania promastigotes, which were used to extract genomic DNA. Four specific genes (Amastin, CaNA2, Kmp‐11 and PDI) were amplified by polymerase chain reaction (PCR) (the primers are listed in Table 1) and cloned into pGEM‐T (Promega, Madison, Wisconsin, USA) for sequencing. blast (http://https://blast.ncbi.nlm.nih.gov/Blast.cgi) was used to confirm the identity of the Leishmania isolate in this study.

Table 1.

Primers used in this study

Gene name	Plasmid	Primers (5′→3′)	Product	Restriction enzymes
Amastin	pGEM‐T	P1: ATGCTGTGCTCGTGCATCGTGT	552 bp
		P2: CTACATTATAATCAGCAGCACCAGGCC
CaNA2	pGEM‐T	P3: ATGAGCCGGCCCAAT	1224 bp
		P4: TCAAATGAGTGGCGTGAC
Kmp‐11	pGEM‐T	P5: ATGGCCACCACGTACGAG	279 bp
		P6: TTACTTGGACGGGTACTGCG
PDI	pGEM‐T	P7: ATGCAGCGCTCATTCCTT	1434 bp
		P8: CTACAAATCTTCCTCTTCGCTG
Kmp‐11	pET30a(+)	P9: CGC GGATCC ATG GCCACC	297 bp	BamHI
		P10: CGC GAATTC TTA CTTGGACGGGTAC		EcoRI
CaNA2 (HLA‐restricted epitopes gene)	pET30a(+)	P11: CGC GAATTC ATG ACGTCTGTAGAACG	429 bp	EcoRI
		P12: CGC GAGCTC TTA GAGTTGAAGAACGTC		SalI
Kmp‐11/CaNA2 (HLA‐restricted epitopes gene)	pET30a(+)	P9	732 bp	BamHI
		P13: CGC GAATTC ACCGCCACC CTTGGACGGGTACTGCG		EcoRI
		P14: CGC GAATTC GGTGGCGGTGGCTCC ACGTCTGTAGAACG		EcoRI
		P12		SalI
CpG/Kmp‐11	pCMV‐C‐His	P15: CGC GGATCC TCGTCGTTTTCGGCGCGCGCCG GCCACC ATGGCCACCACGTAC	322 bp	BamHI
		P16: CGC GAATTC CTTGGACGGGTACTGCG		EcoRI
CpG/CaNA2 (HLA‐restricted epitopes gene)	pCMV‐C‐His	P17: CGC GAATTC TCGTCGTTTTCGGCGCGCGCCG GCCACC ATGACGTCTGTAGAACG	454 bp	EcoRI
		P18: CGC GAT ATC GAGTTGAAGAACGTC		EcoRV
CpG/Kmp‐11/CaNA2 (HLA‐restricted epitopes gene)	pCMV‐C‐His	P15	757 bp	BamHI
		P18		EcoRV

The restriction enzyme sites are underlined. The CpG ODN 2395 sequence is marked in italics. The Kozak sequence used to strengthen the expression of eukaryotic genes is in bold.

Screening CpG ODNs as the DNA vaccine adjuvant

We used four thio‐modified class C CpG ODNs (Invitrogen, Carlsbad, California, USA), which were 2395 (5′‐tcgtcgttttcggcgc:gcgccg‐3′), M362 (5′‐tcgtcgtcgttc:gaacgacgttgat‐3′), D‐SL03 (5′‐tcgcgaacgttcgccgcgttcgaacgcgg‐3′) and 685 (5′‐tcgtcgacgtcgttcgttctc‐3′). The palindrome is underlined. The CpG ODNs were added to the culture medium of RAW264.7 macrophages in 96‐well plates at different concentrations to examine their effects on macrophage proliferation, cytokine secretion and morphology. Moreover, spleen mononuclear cells were isolated from the spleen of an 8‐week‐old female BALB/c mouse provided by the Laboratory Animal Center of Sichuan University with the Mouse Spleen Mononuclear Cell Separation Kit (TBDscience, Tianjin, China). The cells were cultured in 96‐well plates, and the four CpG ODNs were added to the medium to examine their effects on mononuclear cell proliferation by the CCK‐8 method. The best CpG ODN was screened out as a nucleic acid vaccine adjuvant through these cell experiments.

Selection of HLA‐restricted epitopes

The secondary and tertiary structures, hydrophilicity plots, flexible regions, antigenic indices, surface probability plots and subcellular localizations of the CaNA2 and KMP‐11 proteins of Leishmania were predicted using dnastar software,29 PHYRE2 protein fold recognition server 30 and predictprotein.31 HLA‐A2, HLA‐A24 and HLA‐DR1 restricted epitopes of CaNA2 were calculated using four online analysis systems (SYFPEITHI, NetCTL 1.2, NetMHC 4.0 and Rankpep). According to the secondary structure, surface property and HLA‐A2, HLA‐A24 and HLA‐DR1 epitope predictions, the HLA‐restricted epitope sequence of CaNA2 was selected for subsequent vaccine preparation. According to the N‐end rule,32 the sequences of KMP‐11 and the HLA‐restricted epitopes of CaNA2 were linked by a flexible peptide fragment to obtain a KMP‐11/CaNA2 duplex protein (the sequence of KMP‐11 is in front) using a computer. Then, the tertiary structures and potential binding sites of the HLA‐restricted epitopes of CaNA2 and KMP‐11/CaNA2 protein were predicted to assess the vaccines using the PHYRE2 protein fold recognition server and the 3dligandsite server.33

Construction of eukaryotic recombinant plasmids and expression in NIH3T3 cells

The pCMV‐C‐His plasmid (Beyotime, Shanghai, China) was used to clone and express our target genes and prepare the DNA vaccines. The eukaryotic recombinant plasmids pCMV‐CpG‐CaNA2, pCMV‐CpG‐Kmp‐11 and pCMV‐CpG‐Kmp‐11/CaNA2 were constructed by PCR amplification (the primers are listed in Table 1), restriction enzyme digestion and T4 ligase connection. After identification by sequencing, the positive recombinant plasmids were encapsulated by Lipofectamine 2000 (Invitrogen) and transfected into NIH3T3 cells (Procell, Wuhan, China), and their transient expression in the NIH3T3 cells was detected by immunofluorescence. The primary antibody was a His‐tag mouse monoclonal antibody (SAB, College Park, Maryland, USA 1 : 200), and the secondary antibody was a fluorescein‐conjugated affinipure goat anti‐mouse IgG (ZSGB‐BIO, Beijing, China, 1 : 200). Then, the stable expression of these plasmids was measured by reverse transcription‐polymerase chain reaction (RT‐PCR) and Western blot after 14 days of stable screening with 300 μg/ml of G418. The primary antibody used for Western blot was also the His‐tag mouse monoclonal antibody (SAB, 1 : 5000), and the secondary antibody was a peroxidase‐conjugated affinipure goat anti‐mouse IgG (ZSGB‐BIO, 1 : 4000). These recombinant plasmids, which could be stably expressed in NIH3T3 cells, were amplified and extracted to prepare DNA vaccines for the subsequent animal experiments.

Construction of prokaryotic recombinant plasmids, protein expression and purification

We also constructed the prokaryotic recombinant plasmids pET‐CaNA2, pET‐Kmp‐11 and pET‐Kmp‐11/CaNA2 using pET30a(+) (BioDee, Beijing, China) (the primers are listed in Table 1). These recombinant plasmids were transformed into competent E. coli DH5α cells (TIANGEN, Beijing, China). After sequencing, the positive recombinant plasmids were transformed into competent E. coli BL21 cells (TIANGEN) and induced by 1 mmol of isopropyl β‐d‐thiogalactoside to express the His‐tagged fusion proteins. These fusion proteins were purified using HisTrap affinity columns (GE Healthcare, Chicago, Illinois, USA) and dialysis and identified using sodium dodecyl sulfate–polyacrylamide gel electrophoresis and Western blot. The antibodies used for Western blot are consistent with those used in the previous NIH3T3 cell experiments. The endotoxin content of these purified proteins was tested with the ToxinSensor Chromogenic LAL Endotoxin Assay Kit (Genscript, Nanjing, China), and the qualified proteins were collected for use as protein vaccines in the subsequent animal experiments.

DNA‐protein vaccination and Leishmania infection in BALB/c mice

Because BALB/c mice are susceptible to L. donovani infection and can be used to study the immunopathological changes that occur during VL, we chose these mice for the vaccine experiments. Eight‐week‐old female BALB/c mice provided by the Laboratory Animal Center of Sichuan University were maintained under standard conditions. The first immunization of the mice was performed using Lipofectamine 2000‐encapsulated plasmids as DNA vaccines. Two weeks later, the purified proteins were mixed with Freund's complete adjuvant as protein vaccines to boost immunity. In the first immunization, five groups of mice were immunized with phosphate‐buffered saline (PBS) (n = 24), pCMV‐C‐His (n = 20), pCMV‐CpG‐CaNA2 (n = 20), pCMV‐CpG‐Kmp‐11 (n = 20) and pCMV‐CpG‐Kmp‐11/CaNA2 (n = 20), respectively (25 μg of plasmid in 75 μl of PBS mixed with 25 μl of Lipofectamine 2000), by intramuscular injection. In the second immunization, these groups were immunized with PBS, PBS + Freund's complete adjuvant, CaNA2 protein, KMP‐11 protein and KMP‐11/CaNA2 protein, respectively (10 μg of protein in 100 μl of PBS mixed with 100 μl of Freund's complete adjuvant), by intraperitoneal injection. Due to the use of Freund's adjuvant, the formation of water‐in‐oil droplets is conducive to the slow release of our vaccine, and intraperitoneal injection assists with the absorption of these droplets. Two weeks after the DNA‐protein immunization, four mice per group were challenged with 1 × 10⁸ Leishmania promastigotes through an intraperitoneal injection. Another four mice in the normal group were not infected with Leishmania. Hence, the following six groups of mice were created after challenge: the normal (unvaccinated and uninfected), non‐vaccine (unvaccinated but infected), reagent control (vaccinated with empty plasmid and Freund's complete adjuvant and infected), CaNA2, Kmp‐11 and Kmp‐11/CaNA2 groups. After immunization and infection, the appearances and behavior of the mice were observed daily.

Detection of antibody titers, cytokine responses and the spleen parasite burden

Mouse serum samples from each group were collected at 2, 4, 6 and 8 weeks after the DNA‐protein immunization and 2 months after infection. The serum total immunoglobulin G (IgG), IgG1 and IgG2a antibody titers in each group were detected by indirect enzyme‐linked immunosorbent assay (ELISA). The purified CaNA2, KMP‐11 and KMP‐11/CaNA2 proteins (100 μl, 2 μg/ml per well) were coated onto 96‐well plates (Costar, Corning, New York, USA) as antigens. Serum samples from each group were diluted in 0·1% PBST using the doubling dilution method to determine the antibody titers. Correspondingly, diluted normal mouse serum was used as the negative control. The secondary antibodies were peroxidase‐conjugated Affinipure goat anti‐mouse IgG (ZSGB‐BIO, 1 : 1000), peroxidase‐conjugated Affinipure goat anti‐mouse IgG, Fcγ subclass 1 specific (Proteintech,Rosemont, Illinois, USA, 1 : 1000) and peroxidase‐conjugated Affinipure goat anti‐mouse IgG, Fcγ subclass 2a specific (Proteintech, 1 : 1000).

The serum interferon‐γ (IFN‐γ), interleukin‐6 (IL‐6), IL‐10 and tumor necrosis factor‐α (TNF‐α) levels were measured using biotin–streptavidin double antibody sandwich ELISA kits (Cloud‐Clone Corp, Houston, Texas, USA) after immunization and infection.

Two months after infection, the mice in each group were euthanized. Their spleens and livers were weighed and made into Wright's stained imprints. The spleen parasite burden was counted with the Leishman Donovan Units (LDU) method (number of amastigotes per 1000 spleen cell nuclei × mg of spleen weight).34 Compared with that of the non‐vaccine group, we calculated the amastigote reduction rates of the reagent control, CaNA2, Kmp‐11 and Kmp‐11/CaNA2 groups. However, the number of parasites in the liver was still small after 2 months of infection, and finding Leishmania amastigotes under the microscope was difficult; therefore, we did not calculate the liver parasite burden.

Detection of serum total protein, creatinine, urea nitrogen, ALT, NO and H₂O₂ concentrations and the distribution of the DNA vaccines in each mouse tissue

The total protein, creatinine, urea nitrogen, alanine aminotransferase (ALT), nitric oxide (NO) and hydrogen peroxide (H₂O₂) concentrations in the serum samples were measured with the Enhanced BCA Protein Assay Kit (Beyotime), Creatinine Assay Kit (NJJCBIO, Nanjing, China), Urea Assay Kit (NJJCBIO), Alanine Aminotransferase Assay Kit (NJJCBIO), Nitric Oxide Assay Kit (Beyotime) and Hydrogen Peroxide Assay Kit (Beyotime), respectively. Moreover, mice in each group were euthanized, and genomic DNA was extracted from the heart, liver, spleen, lung, kidney, uterus and injection‐site muscles to determine the distribution of the DNA vaccines in each tissue using PCR amplification at different weeks post‐immunization.

Statistical analysis

The statistical analyses were performed using the IBM spss statistics 19 software Armonk, New York, USA. The data were analyzed using one‐way analysis of variance and expressed as the means ± standard deviations. Significant differences were determined and designated with asterisks as follows: * P < 0·05; ** P < 0·01; and *** P < 0·001.

Results

Identification of Leishmania

Five months after the golden hamsters were inoculated with the patient's bone marrow, a large number of Leishmania amastigotes were found in their livers and spleens (Fig. 1a,b). The spleens were homogenized and cultured in M199 medium for seven days, and Leishmania promastigotes were successfully obtained (Fig. 1c,d). Four specific genes (Amastin, CaNA2, Kmp‐11 and PDI) were amplified successfully using genomic DNA from Leishmania promastigotes as the template (Fig. 1e–g). The sequencing data for Amastin (GenBank ID: MH107836), CaNA2 (GenBank ID: MH107837), Kmp‐11 (GenBank ID: MH107835) and PDI (GenBank ID: MH107838) were aligned with blast, and the results suggested that this Leishmania strain belonged to Leishmania donovani. We named the strain Leishmania donovani MHOM/CN/2016/SCHCZ.

Figure 1.

Isolation, culture and identification of Leishmania. Three golden hamsters were inoculated with bone marrow puncture fluid with Leishmania amastigotes from a visceral leishmaniasis (VL) patient. Five months later, the spleens of the golden hamsters were homogenized to culture Leishmania promastigotes and identify this Leishmania strain. (a) Wright's staining of a spleen imprint from a golden hamster (× 400). (b) Wright's staining of a liver imprint from a golden hamster (× 400). (c) Promastigotes obtained after 7 days of spleen homogenate culture (× 400). (d) Wright's staining of Leishmania promastigotes (× 1000). (e) The extracted Leishmania genomic DNA. (f) The PDI gene was amplified by polymerase chain reaction (PCR). (g) The CaNA2, Amastin and Kmp‐11 genes were amplified by PCR.

Selection of CpG ODNs

The effects of CpG ODNs 2395, M362, D‐SL03 and 685 on the proliferation of RAW264.7 macrophages and mouse spleen mononuclear cells are shown in Fig. 2. The effect of D‐SL03 on macrophage proliferation was strongest, followed by the effects of 2395 and M362; conversely, 685 had almost no effect on macrophage proliferation (Fig. 2a,b). In mouse mononuclear cells, CpG ODN 2395 performed best on relative cell proliferation at different concentrations (Fig. 2d). The effects of these four CpG ODNs on cytokine secretion and macrophage morphology are revealed in Fig. 3. These four CpG ODNs had little effect on IL‐6 and IL‐10 secretion by macrophages at the 0·5‐μmol concentration (Fig. 3a). However, at the same concentration, 2395, M362 and D‐SL03 significantly promoted IFN‐γ and TNF‐α secretion (P < 0·05) (Fig. 3a). CpG ODN 2395 stimulated the strongest IFN‐γ secretion response, whereas D‐SL03 stimulated the strongest TNF‐α secretion response. Furthermore, we found increased nuclear division of macrophages after addition of these four CpG ODNs but decreased azurophil granules in the cytoplasm (Fig. 3b). Based on these results, we eventually chose 2395 as a nucleic acid adjuvant for the DNA vaccine.

Figure 2.

Effects of four oligodeoxynucleotides (ODNs) on macrophage and mouse spleen mononuclear cells proliferation. The phosphorothioate‐modified CpG ODNs 2395, M362, D‐SL03 and 685 were added to the culture medium of RAW264.7 macrophages and mouse spleen mononuclear cells at 0·25 μmol, 0·5 μmol, 1 μmol and 2 μmol concentration. Each group had three replicates. The absorbance was measured by the CCK‐8 method at 12, 24, 36 and 48 hr of culture. (a) The effects of the four CpG ODNs on the absolute (b) and relative proliferation of macrophages. (c) The effects of the four CpG ODNs on the absolute (d) and relative proliferation of mouse spleen mononuclear cells. *P < 0·05; **P < 0·01 and ***P < 0·001.

Figure 3.

Effects of the four CpG oligodeoxynucleotides (ODNs) on cytokine secretion and morphology of RAW264.7 macrophages. The phosphorothioate‐modified CpG ODNs 2395, M362, D‐SL03 and 685 were added to the culture medium of macrophages at a 0·5 μmol concentration. Each group had three replicates. Supernatants were collected from the cultures at 12, 24, 36 and 48 hr to detect interferon‐γ (IFN‐γ), tumor necrosis factor‐α (TNF‐α), interleukin‐6 (IL‐6) and IL‐10 secretion by ELISA. After 48 hr of culture, cell morphological changes were observed by Giemsa staining. (a) Effects of the four CpG ODNs on cytokine secretion by macrophages. (b) Effects of the four CpG ODNs on macrophage morphology (× 1000).

Protein analysis and selection of CaNA2 HLA‐restricted epitopes

The secondary structure and surface property analyses revealed that the secondary structure of CaNA2 whole protein contained multiple α‐helices, β‐sheets and β‐turns, and that the areas with high hydrophilicity, flexibility, antigenic index and surface probability were dispersed (Fig. 4a). The secondary structure of KMP‐11 was mainly composed of an α‐helix, and most areas had high hydrophilicity, flexibility, antigenic index and surface probability (Fig. 4b). The simulated tertiary structures of CaNA2 and KMP‐11 are shown in Fig. 4(c,d), respectively. The subcellular localization analysis indicated that CaNA2 was targeted to the nucleus (Fig. 4e), whereas KMP‐11 was targeted to the cytoplasm (Fig. 4f). Table 2 shows the predicted HLA‐A2, HLA‐A24 and HLA‐DR1 epitopes of CaNA2 protein. According to these results, the amino acid sequence 19~153 aa of CaNA2 was selected as the HLA‐restricted epitope sequence. The tertiary structure and potential binding sites of the HLA‐restricted epitope sequence of CaNA2 are shown in Fig. 4(g,h). The recombinant KMP‐11/CaNA2 protein was simulated by a computer, and its tertiary structure and potential binding sites were also determined (Fig. 4i,j).

Figure 4.

Secondary and tertiary structures, surface properties, subcellular localizations and potential binding sites of the target proteins. We used the dnastar software to predict secondary structures and surface properties of our target proteins, PHYRE2 to predict tertiary structures, predictprotein to predict subcellular localizations, and 3DL igandsite to predict three‐dimensional models and potential binding sites. (a) Secondary structure and surface properties of CaNA2. (b) Secondary structure and surface properties of KMP‐11. (c) Tertiary structure of CaNA2. The image is colored by a rainbow from the N→C terminus. Model dimensions (Å): X: 62·990, Y: 100·407, Z: 58·630. A total of 88% of the residues were modelled with 100% confidence. (d) Tertiary structure of KMP‐11. Model dimensions (Å): X: 25·672, Y: 13·493, Z: 21·773. (e) Subcellular localization of CaNA2 (prediction confidence 23). (f) Subcellular localization of KMP‐11 (prediction confidence 97). (g) Tertiary structure of CaNA2 HLA‐restricted epitope sequence. Model dimensions (Å): X: 43·981, Y: 37·386, Z: 37·767. A total of 99% of the residues were modelled with 100% confidence. (h) Potential binding sites of CaNA2 HLA‐restricted epitope sequence. Five potential binding sites 63ASP, 65GLN, 91ASN, 123SER and 124ASN were predicted. (i) Tertiary structure of KMP‐11/CaNA2 protein. Model dimensions (Å): X: 72·846, Y: 52·747, Z: 46·623. A total of 97% of the residues were modelled with > 90% confidence. (j) Potential binding sites of KMP‐11/CaNA2. Four potential binding sites 165ASP, 193ASN, 225SER and 226ASN were predicted.

Table 2.

Prediction of HLA‐A2, HLA‐A24 and HLA‐DR1 epitopes of CaNA2

MHC supertypes	SYFPEITHI	NetCTL 1.2 server	NetMHC 4.0 server	Rankpep
HLA‐A2	308 HLFPCIISL 71 SINDTVVVV 121 ILFLLAAKV 192 GLSPDVSHV 164 TLLPQILSA 371 LMSIFLTLL 41 FLCGEQLRL 62 VLRTEPNTL 366 FLESNLMSI 370 NLMSIFLTL 91 KILAACGSL 117 NLECILFLL 110 YIGNGGFNL 204 ALIHRFRHI 363 SLPFLESNL 31 KLSLENIML 124 LLAAKVAHP 168 QILSAFNCL 181 IIRKKFFCV 289 IRAHEVHEL 332 VLVVSKETI	308 HLFPCIISL 370 NLMSIFLTL 366 FLESNLMSI 371 LMSIFLTLL 164 TLLPQILSA 41 FLCGEQLRL 117 NLECILFLL 71 SINDTVVVV YAMEIVQQA 192 GLSPDVSHV 121 ILFLLAAKV 31 KLSLENIML FNLECILFL 110 YIGNGGFNL 91 KILAACGSL NTLSINDTV 48 RLEYAMEIV	370 NLMSIFLTL 308 HLFPCIISL 366 FLESNLMSI 164 TLLPQILSA 192 GLSPDVSHV 41 FLCGEQLRL 121 ILFLLAAKV 371 LMSIFLTLL 51 YAMEIVQQA 146 FMADVLQLG 71 SINDTVVVV 266 GLSYVFNFA 110 YIGNGGFNL 31 KLSLENIML 117 NLECILFLL 171 SAFNCLPLA 116 FNLECILFL	71 SINDTVVVV 366 FLESNLMSI 370 NLMSIFLTL 121 ILFLLAAKV 54 EIVQQAALV 192 GLSPDVSHV 308 HLFPCIISL 332 VLVVSKETI 371 LMSIFLTLL 55 IVQQAALVL 164 TLLPQILSA TLSINDTVV 31 KLSLENIML 48 RLEYAMEIV 158 QLKYSSTLL 91 KILAACGSL 33 SLENIMLQF 204 ALIHRFRHI 145 KFMADVLQL 117 NLECILFLL 41 FLCGEQLRL
HLA‐A24	245 TYVPRLGLF 103 KYLFLGNYI 301 LYRPHPNHL 85 QYYDLVKIL 309 LFPCIISLF 185 KFFCVHSGL 115 GFNLECILF 145 KFMADVLQL 252 LFETRPLFI 278 RFVSANNLL 325 SFGNKGAVL 130 AHPHSIFLI 359 AFSWSLPFL 116 FNLECILFL 204 ALIHRFRHI	245 TYVPRLGLF 309 LFPCIISLF 103 KYLFLGNYI 301 LYRPHPNHL 278 RFVSANNLL 145 KFMADVLQL 252 LFETRPLFI 359 AFSWSLPFL 85 QYYDLVKIL 130 AHPHSIFLI KQRGLSYVF 268 SYVFNFACA	245 TYVPRLGLF 309 LFPCIISLF 103 KYLFLGNYI 265 RGLSYVFNF 278 RFVSANNLL 145 KFMADVLQL 263 KQRGLSYVF 370 NLMSIFLTL 301 LYRPHPNHL 130 AHPHSIFLI 358 NAFSWSLPF 302 YRPHPNHLF 354 LQGRNAFSW 252 LFETRPLFI 165 LLPQILSAF 85 QYYDLVKIL 157 CQLKYSSTL	301 LYRPHPNHL 245 TYVPRLGLF 359 AFSWSLPFL 85 QYYDLVKIL 309 LFPCIISLF 185 KFFCVHSGL
HLA‐DR1				278 RFVSANNLL 54 EIVQQAALV 19 TSVERNTLR 302 YRPHPNHLF 245 TYVPRLGLF 88 DLVKILAAC 208 RFRHIPTRG 321 NYCSSFGNK 300 KLYRPHPNH 287 CIIRAHEVH 35 ENIMLQFLC

The results include the residues numbers and peptide sequences. The algorithms of these four epitope prediction analysis systems are different. As each algorithm is not 100% accurate, we chose four analysis systems and integrated the results to improve the accuracy of the prediction. Epitope sequences predicted by all four online analysis systems are marked by with straight lines. They are the sequences of the 31–39, 41–49, 71–79, 117–125, 121–129, 164–172, 192–200, 308–316, 366–374, 370–378, and 371–379aa HLA‐A2 epitopes and 85–93, 245–253, 301–309, and 309–317aa HLA‐A24 epitopes. Epitope sequences predicted by three of the four online analysis systems are marked with bold. They are the sequences of the 91–99 and 110–118aa HLA‐A2 epitopes and 103–111, 130–138, 145–153, 252–260, 278–286, and 359–367aa HLA‐A24 epitopes.

Preparation of the DNA and protein vaccines

The immunofluorescence results suggested that transient expression of the eukaryotic recombinant plasmids pCMV‐CpG‐CaNA2, pCMV‐CpG‐Kmp‐11 and pCMV‐CpG‐Kmp‐11/CaNA2 in NIH3T3 cells was successful (Fig. 5a). RT‐PCR and Western blot demonstrated stable expression of these plasmids (Fig. 5b,c). Moreover, the molecular weights of our target proteins expressed in E. coli were predicted to be 20 700 (CaNA2), 16 700 (KMP‐11) and 32 400 (KMP‐11/CaNA2) by the DNASTAR software, which were consistent with the sodium dodecyl sulfate polyacrylamide gel electrophoresis and Western blot results (Fig. 5d). These eukaryotic recombinant plasmids were used as DNA vaccines, and the purified proteins expressed by E. coli were used as protein vaccines.

Figure 5.

Eukaryotic and prokaryotic expression of our target genes. (a) Detection of transient expression of eukaryotic recombinant plasmids in NIH3T3 cells by immunofluorescence. The expression of target genes after transfection of the cells with phosphate‐buffered saline (PBS), pCMV‐C‐His, pCMV‐CpG‐Kmp‐11/CaNA2, pCMV‐CpG‐CaNA2, and pCMV‐CpG‐Kmp‐11 at 24 hr (1–5) and 48 hr (6–10), respectively. (b) PCR and RT‐PCR amplification of target genes after 14 days of stable screening. Lane M: DNA marker (DL2000); lanes 1–5: PCR amplification using the genomic DNA of pCMV‐C‐His, pCMV‐CpG‐Kmp‐11, pCMV‐CpG‐CaNA2 and pCMV‐CpG‐Kmp‐11/CaNA2 transfected cells and normal NIH3T3 cells as templates, respectively; lanes 6–10: RNA from these five groups of cells; lanes 11–15: RT‐PCR amplification of GAPDH from these five groups of cells; lanes 16–20: RT‐PCR amplification of target genes from these five groups of cells. (c) Western blot identification of stable expression of the recombinant plasmids in NIH3T3 cells. (d) Sodium dodecyl sulfate–polyacrylamide gel electrophoresis and Western blot identification of the expression of multi‐epitope KMP‐11/CaNA2, CaNA2 and KMP‐11 proteins in Escherichia coli. Lane M: protein markers; lanes 1 and 5: whole bacterial proteins from E. coli BL21 (containing pET30a(+)); lanes 2–4: the multi‐epitope KMP‐11/CaNA2, CaNA2 and KMP‐11 proteins expressed in E. coli BL21, respectively; lanes 6–8: the purified multi‐epitope KMP‐11/CaNA2, CaNA2 and KMP‐11 proteins, respectively; lanes 9–11: Western blot identification of the purified multi‐epitope KMP‐11/CaNA2, CaNA2 and KMP‐11 proteins, respectively.

Total IgG, IgG1 and IgG2a titers, cytokine changes and spleen parasite burdens of BALB/c mice

As shown in Fig. 6(a), the IgG1 and total IgG antibody titers of the CaNA2, Kmp‐11 and Kmp‐11/CaNA2 groups were > 10⁵ at 2, 4, 6 and 8 weeks after immunization and gradually increased over time. Compared with those of the other two groups, the IgG1 and total IgG antibody titers in the Kmp‐11/CaNA2 group were highest at 2 weeks after immunization. However, no statistical difference in the antibody titers for each group was found at the remainder of the time‐points. The IgG2a antibody titers of the three groups were > 10³ and also increased over time after immunization. The IgG2a antibody titers in the CaNA2 and Kmp‐11/CaNA2 groups were higher than that in the Kmp‐11 group. The ratio of IgG1 to IgG2a indicated that the immune response induced by the CaNA2 vaccine was biased towards cellular immunity, whereas the immune response induced by the Kmp‐11 vaccine was biased towards humoral immunity in the mice. These results demonstrated that the mice in group CaNA2 produced strong humoral and cellular immune responses after immunization and that both the level and duration of the IgG response were in line with the general requirements of a vaccine.

Figure 6.

IgG antibody titers, cytokine levels and spleen parasite burdens after DNA‐protein immunization and infection. BALB/c mice were immunized with DNA prime‐protein boost vaccines and challenged with Leishmania at 2 weeks post‐immunization. (a) Serum total IgG, IgG1 and IgG2a antibody titers and IgG1/IgG2a at 2, 4, 6 and 8 weeks post‐immunization. Normal mouse serum served as the negative control. The reagent control group was not marked in the figure because no antibody titers were detected. (b) Spleen parasite burden of BALB/c mice in each group after 2 months of infection. (c) Serum interferon‐γ (IFN‐γ), tumor necrosis factor‐α (TNF‐α), interleukin‐6 (IL‐6) and IL‐10 levels in each group at 2, 4, 6 and 8 weeks post‐immunization and 2 months post‐infection. *P < 0·05; **P < 0·01 and ***P < 0·001.

After 2 months of infection, Wright's staining imprints were made of spleens and livers from each group of mice (see Supplementary material, Figs S1 and S2), and the spleen parasite burden in the CaNA2 group was the lowest of the groups (Fig. 6b). The average spleen amastigote reduction rates in the CaNA2, Kmp‐11, Kmp‐11/CaNA2 and reagent control groups were 76·0%, 50·7%, 37·3% and 14·7%, respectively. These results suggested that our vaccines exerted a protective effect against VL in BALB/c mice.

The ELISA results revealed no significant changes in IFN‐γ, TNF‐α, IL‐6 and IL‐10 in each group at 2, 4, 6 and 8 weeks post‐immunization (Fig. 6c). However, the IL‐6 and IL‐10 levels in the Kmp‐11 and Kmp‐11/CaNA2 groups increased significantly 2 months after Leishmania infection. Unfortunately, no significant differences in IFN‐γ and TNF‐α were observed in each group after 2 months of infection.

NO and H₂O₂ assays in the mice

As shown in Fig. 7(a), the serum NO concentrations in each group appeared to undergo a transient increase 2 weeks after immunization and then returned to normal. The serum H₂O₂ concentration was higher in the Kmp‐11 group than in the other groups after immunization. After 2 months of infection, the serum NO level was highest in group CaNA2, but no significant difference was observed in the H₂O₂ levels among the groups (Fig. 7b).

Figure 7.

The serum total protein, creatinine, urea nitrogen, alanine aminotransferase (ALT), nitric oxide (NO) and hydrogen peroxide (H₂O₂) levels and the safety tests. (a) The serum total protein, creatinine, urea nitrogen, ALT, NO and H₂O₂ levels in each group were measured at 2, 4, 6 and 8 weeks post‐immunization. (b) The serum total protein, creatinine, urea nitrogen, ALT, NO and H₂O₂ levels were measured at 2 months post‐infection. (c) PCR detection of the distribution of the DNA vaccine in each mouse tissue after DNA‐protein immunization. The primers were T3 and T7. Lane M: DNA markers; lanes 1–7: genomic DNA from the heart, liver, spleen, lung, kidney, uterus and injection site muscles of BALB/c mice after immunization, respectively; lanes 8–14 and 16–22: PCR amplification using genomic DNA from the seven tissues of the reagent control group in the above order as templates at 2 and 4 weeks after immunization, respectively; lanes 15 and 23: positive controls using plasmid pCMV‐C‐His as the template; lanes 24–30, 32–38 and 40–46: PCR amplification using genomic DNA from the seven tissues of the CaNA2 group in the above order as templates at 2, 4 and 6 weeks after immunization, respectively; lanes 31, 39 and 47: positive controls using plasmid pCMV‐CpG‐CaNA2 as the template. *P < 0·05; **P < 0·01 and ***P < 0·001.

Safety test of our vaccines

None of the mice in each group exhibited any behavioral abnormalities after immunization. However, when the experimental groups of mice were dissected after immunization, we found intraperitoneal adhesions in these mice but not in the normal group of mice. Because intraperitoneal injection of Freund's adjuvant may lead to injection site pain, abscess formation, intraperitoneal granulomas and adhesions,35 we believe that the intraperitoneal adhesions in our experimental mice were caused by the Freund's adjuvant. The serum total protein, creatinine, urea nitrogen and ALT levels in each group were in the normal ranges at 2, 4, 6 and 8 weeks post‐immunization (Fig. 7a). After 2 months of infection, each group of mice exhibited visible hepatosplenomegaly. However, the serum total protein, creatinine, urea nitrogen and ALT levels in each group were also in the normal ranges (Fig. 7b). Furthermore, the results of PCR showed that the pCMV‐C‐His fragment was amplified in the muscle tissue of the injection site 2 weeks after the DNA‐protein immunization, whereas the pCMV‐CpG‐CaNA2 fragment was amplified in the same tissue at 2 and 4 weeks after immunization, and no target plasmids were detected in the other tissues (Fig. 7c). All of the DNA vaccines were undetectable by PCR at 8 weeks post‐injection, indicating that the DNA vaccines would not integrate into the mouse genomic DNA. The behavioral observations, serum total protein, creatinine, urea nitrogen and ALT concentrations and distribution of the DNA vaccines in each tissue preliminarily proved the safety of our vaccines.

Discussion

HLA‐restricted epitope vaccines and CpG ODNs as nucleic acid adjuvants have shown good performance in combating VL.36, 37 The Lipofectamine‐encapsulated DNA vaccine was proven to slow the degradation of the DNA vaccine, prolong the duration of the vaccine, and achieve better immune responses.38 The DNA prime‐protein boost immunization strategy was shown to produce good immunogenicity and protective immunity by many vaccine studies.39, 40 However, these methods have not been applied by an experiment at the same time. In addition, due to the long‐term preservation and passage of microorganisms in laboratories, their virulence may be lost or decrease.41 Therefore, newly isolated clinical strains can better reflect the infectivity and pathogenicity of pathogenic microorganisms. In our study, we successfully designed a CaNA2 HLA‐restricted epitope vaccine, used a class C CpG ODN as a DNA vaccine adjuvant, encapsulated the DNA vaccine with Lipofectamine, immunized mice with the DNA prime‐protein boost immunization strategy, and infected the animals with a newly isolated Leishmania strain. The results of IgG antibody titers and spleen parasite burden demonstrated that the CaNA2 HLA‐restricted epitope vaccine had strong immunogenicity and provided an immune protective effect against VL in BALB/c mice. The IgG1 and IgG2a levels after immunization showed that the vaccine induced both strong humoral and cellular immunity. The serum total protein, creatinine, urea nitrogen and ALT levels and distribution of the DNA vaccine in each tissue of the mice suggested that our vaccine was safe.

Some interesting data in this study deserve our analysis and discussion. The serum IFN‐γ, TNF‐α, IL‐6 and IL‐10 levels were almost unchanged in each group at 2, 4, 6 and 8 weeks after DNA‐protein immunization. Therefore, we can reasonably postulate that the immunization dose was small and not sufficient to cause changes in cytokines or that the maintenance time of the cytokine changes was too short to be detected. In addition, after 2 months of infection, the IFN‐γ level was not significantly different among the groups, indicating that the T helper type 1 (Th1) immune response in the mice was not strong. As control of VL depends on a successful cell‐mediated immune response in which IFN‐γ and natural killer cells lead to stimulation of the microbicidal action mediated by NO,42 this result suggested that our vaccines were not ideal. However, we consider that this observation might also be caused by the dose and inoculation route. Kaur et al.43 found that the dose and inoculation route for Leishmania had significant impacts on the type of immune responses in mice. When mice were inoculated with a high dose (10⁷) of Leishmania through the intraperitoneal route, the Th1 immune response in the mice was weak and the Th2 immune response was strong. Moreover, VL is strongly correlated with high production of IL‐10, which is an immunosuppressive cytokine that inhibits leishmanicidal immune functions.3, 36 However, in this study, after 2  months of infection with Leishmania, the IL‐10 levels in the Kmp‐11 and Kmp‐11/CaNA2 groups were significantly increased, especially in the Kmp‐11 group, which was inconsistent with the results of some other studies. After consulting some articles, we found that this effect might be caused by the combination of Kmp‐11 and the CpG ODNs. Agallou et al.44 immunized BALB/c mice with PBS, bone marrow‐derived dendritic cells (BM‐DCs), Kmp‐11‐pulsed BM‐DCs, CpG ODNs and (Kmp‐11 + CpG ODNs)‐pulsed BM‐DCs and then challenged them with L. infantum and obtained spleen cells to detect cytokines. These authors also observed a higher IL‐10 level in the (Kmp‐11 + CpG ODNs)‐pulsed BM‐DC group, whereas no significant differences were found among the other four groups. However, few articles have investigated this topic. Therefore, why the combination of Kmp‐11 and CpG ODNs produced a high IL‐10 level is not clear.

Leishmania membrane and extracellular proteins may be related to parasite immune escape, so the application of intracellular proteins, especially some enzymes, as vaccines has been studied often and has shown good immune effects.45, 46 This study represents the first use of CaNA2 as a vaccine in VL experiments. The IgG antibody titers after immunization and the 76% amastigote reduction rate after infection indicated that the CaNA2 HLA‐restricted epitope vaccine had strong immunogenicity and immune protection against VL. However, after infection, the IFN‐γ, TNF‐α, IL‐6 and IL‐10 levels had no significant difference between the CaNA2 and normal group, and the amastigote reduction rate was significantly reduced when the vaccine was combined with Kmp‐11. The reasons for these discrepancies are not clear, probably because we only measured these items after 2 months of infection and could not observe dynamic changes in cytokines and the visceral parasite burden. As CaN is involved in a variety of cellular signaling events and activation processes47 and is considered a potential target for chemotherapeutic intervention in Chagas disease,11 we think that it may also have a good application prospect in VL, but further studies are needed.

Considering all of the work in our study, there were some limitations. The concentration of the DNA vaccine was only 25 μg per mouse, and after infection, the cytokine secretion and IgG titers were measured only at 2 months. We will make improvements in future experiments. Our next plan is to optimize our vaccine to increase IFN‐γ secretion in the mice, to challenge mice with different strains and amounts of Leishmania, and to use different routes of infection and different strains of mice to verify the effectiveness of our vaccines.

Our study successfully obtained DNA and protein encoding HLA‐A2, HLA‐A24 and HLA‐DR1 restricted epitopes of CaNA2 as vaccine candidates against visceral leishmaniasis. The results demonstrate that the DNA prime‐protein boost vaccine may be a safe and effective vaccine candidate against VL, and that vaccine preparation, immune adjuvants and the immunization strategy may provide new ideas for future vaccine experiments. However, the mechanism and immune response of the vaccine should be further studied.

Disclosures

The authors declare no conflict of interest.

Supporting information

Figure S1. Spleen imprints of the mice from each group at 2 months post‐infection.

Figure S2. Liver imprints of the mice from each group at 2 months post‐infection.