Cross-neutralization of SARS-CoV-2 by HIV-1 specific broadly neutralizing antibodies and polyclonal plasma

Cross-reactive epitopes (CREs) are similar epitopes on viruses that are recognized or neutralized by same antibodies. The S protein of SARS-CoV-2, similar to type I fusion proteins of viruses such as HIV-1 envelope (Env) and influenza hemagglutinin, is heavily glycosylated. Viral Env glycans, though host derived, are distinctly processed and thereby recognized or accommodated during antibody responses. In recent years, highly potent and/or broadly neutralizing human monoclonal antibodies (bnAbs) that are generated in chronic HIV-1 infections have been defined. These bnAbs exhibit atypical features such as extensive somatic hypermutations, long complementary determining region (CDR) lengths, tyrosine sulfation and presence of insertions/deletions, enabling them to effectively neutralize diverse HIV-1 viruses despite extensive variations within the core epitopes they recognize. As some of the HIV-1 bnAbs have evolved to recognize the dense viral glycans and cross-reactive epitopes (CREs), we assessed if these bnAbs cross-react with SARS-CoV-2. Several HIV-1 bnAbs showed cross-reactivity with SARS-CoV-2 while one HIV-1 CD4 binding site bnAb, N6, neutralized SARS-CoV-2. Furthermore, neutralizing plasma antibodies of chronically HIV-1 infected children showed cross neutralizing activity against SARS-CoV-2 pseudoviruses. Collectively, our observations suggest that human monoclonal antibodies tolerating extensive epitope variability can be leveraged to neutralize pathogens with related antigenic profile.


Introduction
Broadly neutralizing antibodies (bnAbs) targeting the HIV-1 envelope glycoprotein (Env) can neutralize a broad range of circulating HIV-1 isolates and have been called super-antibodies due to their remarkable potency and neutralization breadth against viruses with significant alterations in the core epitope [1]. As a result of its extensive genetic diversity, HIV-1 is subdivided in multiple clades and circulating recombinant forms (CRFs). A rare subset of HIV-1 infected individuals develops broad and potent antibody responses and have served as potential candidates for the isolation of HIV-1 bnAbs [2,3]. HIV-1 bnAbs take years to develop, have atypical features including long complementarity-determining regions (CDR) loops, high levels of somatic hypermutations (SHMs), presence of insertions and/or deletions (indels), tyrosine sulfation, and develop to tolerate significant alterations in their core epitope [1][2][3]. Notably, V2-apex bnAbs have been shown to exhibit cross-group neutralization activity with viruses derived from HIV-1 group M, N, O and P Envs. Furthermore, they even show crossneutralization of simian immunodeficiency virus (SIV) isolates [4]. Though such extensive cross-neutralization potential of several bnAbs is often associated with poly and/or autoreactivity. Several glycan-dependent HIV-1 bnAbs, such as PGT121 and PGT151, have been shown to bind uninfected/bystander cells.
Herein, we reasoned that given the unique nature of HIV-1 bnAbs and their ability to recognize and/or accommodate viral glycans, the glycan shield of SARS-CoV-2 spike protein can be targeted by HIV-1 specific bnAbs.

Broadly neutralizing antibodies targeting HIV-1 Env cross-react with SARS-CoV-2 spike protein
In the past decade, a large panel of bnAbs and non-nAbs targeting diverse epitopes on the HIV-1 Env glycoprotein (gp160) have been isolated and extensively characterized (reviewed in refs [1][2][3]). To evaluate the potential cross-reactivity of these antibodies, we first performed binding ELISA of 30 bnAbs and 7 non-nAbs with SARS-CoV-2 S2P ecto protein (pre-fusion stabilized ectodomain construct, 1-1208 amino acid residues) and receptor-binding domain (RBD, residues 319-541, also called S1 B domain). The HIV-1 bnAbs were categorized into five categories based on their epitopes on the HIV-1 viral Env (Fig 1A). CR3022 (a nAb isolated from a convalescent SARS-CoV patient [15], which has been shown to cross-react with SARS-CoV-2 [7]), and CC12.1 (a SARS-CoV-2 neutralizing antibody isolated from a convalescent donor) were used as positive control while two antibodies targeting the envelope glycoprotein of simian immunodeficiency virus (SIV) were used as negative control. Of the 30 HIV-1 bnAbs tested for binding to both S2P ecto protein and RBD of SARS-CoV-2, 6 bnAbs (VRC07.523LS, NIH45-46 G54W, N6, Z13e1, 2F5 and 4E10) showed significant binding, while one bnAb (CAP256.09) showed moderate binding to only the S2P ecto protein (Fig 1A). Non-nAbs (non-neutralizing antibodies) that target post-fusion and/or open trimeric conformation of HIV-1 Env were unable to bind both SARS-CoV-2 S2P ecto protein and RBD protein (Fig 1B), suggesting that only pre-fusion state specific antibodies that evolve via extensive somatic hypermutation and affinity maturation in response to repeated exposure to a continuously evolving antigen can cross react with other viruses.
Though reactivity against RBD was stronger than S2P ecto protein for majority of the bnAbs tested, a significant positive correlation was seen between binding to S2P ecto protein and RBD ( Fig 1C). All the 6 bnAbs that showed binding reactivity in ELISA exhibited a similar binding profile to the cell surface expressed SARS-CoV-2 S glycoprotein ( Fig 1D). Of the 30 monoclonal antibodies tested herein, bnAbs targeting the membrane proximal external region (MPER) of HIV-1 showed maximum binding to both the S protein and RBD with half-maximal effective concentration (EC 50 ) of 0.71 μg/ml and 1.71 μg/ml (Z13e1), 0.048 μg/ml and 2.91 μg/ml (2F5) and 0.79 and 0.33 μg/ml (4E10) μg/ml to the RBD and S2P ecto respectively. VRC07.523LS is an engineered variant of the VRC01 bnAb with higher SHM [16] and while it showed binding to both S2P ecto protein and RBD, both VRC01 and its somatic related clone VRC03 did not show any binding (Fig 2A and 2B) further suggesting that higher SHM in HIV-1 bnAbs might be responsible for the observed cross-reactivity.

PLOS PATHOGENS
HIV-1 bnAbs cross-neutralize SARS-CoV-2 to infection by all seven coronaviruses [21][22][23]. No infection of HEK293T cells with SARS-CoV-2 pseudoviruses was observed (Fig 3A and 3D). HEK293T/ACE2 cells and Vero-E6 cells showed comparable infection though the infection titres typically remained in the range of 1 x 10 4 to 2 x 10 5 while Huh7 cells were moderately infected with infection titre in the range of 10 3 to 10 4 ( Fig 3A). Truncation of the SARS-CoV-2 spike has been shown to enhance production of viral pseudoviruses [19,20,24,25]. We next utilized a truncated version of SARS-CoV-2 lacking the last 19 amino acids from the C terminal (Δ19). Compared to wildtype spike, Δ19 spike showed 2-3-fold increase in infection titre in all three cell lines ( Fig 3B). Since the HIV-1 proviral plasmid used (pNL4-3.Luc.R-E-) contains luciferase gene in the same frame as HIV-1 Env, we reasoned that using a separate plasmid containing luciferase gene under a strong constitutive promoter (CMV promoter) can further increase the pseudoviral titres. Next, we generated pseudoviruses using a three-plasmid system (HIV-1 proviral backbone lacking Env and luciferase, a CMV-luciferase construct with HIV-1 packaging signal and Δ19 spike protein). Compared to two-plasmid system with wildtype SARS-CoV-2 spike, three-plasmid system with
We, therefore, employed the three plasmid HIV-1 pseudovirus assay with HEK293T-ACE2 cells as they also provided as to use HEK293T cells as negative infection controls.

N6, a bnAb targeting HIV-1 CD4bs, neutralized SARS-CoV-2 pseudoviruses but failed to neutralize authentic virus
All 6 HIV-1 bnAbs that showed binding to SARS-CoV-2 S protein and RBD were then tested for their ability to block infection using a HIV-1 pseudovirus based neutralization assay utilizing SARS-CoV-2 spike protein. VSV-G and MLV pseudotyped viruses were used as negative control. Except N6, all remaining five bnAbs failed to neutralize SARS-CoV-2 ( Fig 4A). Though N6 showed neutralization of SARS-CoV-2, it failed to show complete neutralization (maximum percent neutralization of 88% with an IC 50 of 0.988 μg/ml) and compared to an anti-RBD nAb isolated from a convalescent donor, CC12.1, N6 showed~100-fold weak neutralization. N6 had a moderate binding affinity of 1.04 x 10 8 M against SARS-CoV-2 RBD ( Fig  4B). Furthermore, N6 failed to block RBD binding to soluble ACE2 by ELISA (Fig 4C), suggesting it recognizes an epitope on RBD outside the ACE2 binding site.
While pseudoviruses are a good surrogate for identifying neutralizing antibodies, we next tested the ability of N6 to block infection by authentic SARS-CoV-2 virus using a cytopathic effect (CPE) based neutralization assay. The N6 mAb did not show significant reduction in neutralization potential in cytopathic effect-based assay when tested up to 20 μg/ml against live authentic SARS-CoV-2 live virus. The control purified Anti-RBD IgG from immunized mice sera showed neutralization CPE titers upto 0.1μg/ml. The II62 scFv-Fc antibody isolated from a semi-synthetic library [26] was used as experimental negative control (Fig 5).

Polyclonal plasma from children with chronic HIV-1 infection neutralize multiple coronaviruses
We next tested plasma antibodies of children with chronic HIV-1 infection for their ability to bind SARS-CoV-2 S2P ecto protein and RBD. Ten children that had shown potent neutralization titre against a 12-virus global panel of HIV-1 isolates from previous studies in our lab were selected [27][28][29]. While all ten children showed significant binding to both S2P ecto and RBD (Fig 6A), three children showed potent and near-complete neutralization of SARS-CoV-2 pseudoviruses (AIIMS329, AIIMS330, AIIMS346) while two children (AIIMS355 and AIIMS521) showed moderate neutralization of SARS-CoV-2 (Fig 6B and 6C).
Antibodies generated against common endemic coronaviruses (HKU1, OC43 and 229E) have been shown to cross-react with SARS-CoV-2. To understand whether the SARS-CoV-2 neutralization titres observed in AIIMS329, AIIMS330, AIIMS346, AIIMS355 and AIIMS521 were due to HIV-1 bnAbs or antibodies to other coronaviruses contributed to neutralization, we next tested the neutralization of SARS-CoV-1, HKU1, OC43, MERS, NL63 and OC43 pseudoviruses. Of note, SARS-CoV-1, HKU1, OC43 and MERS belong to betacoronaviruses genus while NL63 and OC43 belong to alphacoronavirus genus. All five plasma samples that showed SARS-CoV-2 cross-neutralization potently neutralized HKU1, one of the viruses responsible for common cold while the extent of neutralization titres against OC43, NL63 and 229E were

PLOS PATHOGENS
HIV-1 bnAbs cross-neutralize SARS-CoV-2 moderately weak (Fig 6D and 6E). None of the five samples showed any significant cross-neutralization of MERS while SARS-CoV-1 was neutralized by all five samples to the similar extent as SARS-CoV-2 (Fig 6D and 6E).

Discussion
Herein, we show that broadly neutralizing antibodies targeting the envelope glycoprotein (gp160) of HIV-1 cross-react with the receptor binding domain (RBD) of SARS-CoV-2. Neutralizing antibodies engage the host immune system to clear the pathogen or infected cells and are promising candidates for combating emerging viruses [30][31][32]. The RBD of coronaviruses are highly immunogenic and infected individuals typically mount a nAb response [10,[33][34][35][36][37]. Given that several HIV-1 bnAbs showed cross-reactivity with RBD of SARS-CoV-2, vaccine efforts should focus inducing antibodies targeting the cross-reactive epitopes on RBD. HIV-1 bnAbs have several atypical features that are generally not seen in antibody responses generated against a wide variety of pathogens. One such feature of HIV-1 antibodies is their poly-and/or autoreactivity which might have been the cause behind the cross-reactivity observed with SARS-CoV-2 RBD. Furthermore, Polyclonal antibody responses generated in chronic stages of HIV-1 tend to be poly-and/or autoreactive [38,39]. These features are not typically associated with antibody responses to most pathogens and are one of the major obstacles in HIV-1 bnAb-based therapies. HIV-1 bnAbs take years to develop, have atypical features including long complementarity-determining regions (CDR) loops, high levels of somatic hypermutations (SHMs), presence of insertions and/or deletions (indels), tyrosine sulfation, and develop to tolerate significant alterations in their core epitope. Furthermore, several of HIV-1 bnAbs have been shown to be poly and/or autoreactive. Autoreactive properties have been described for several HIV-1 bnAbs, including 2F5, 4E10, !0E8, VRC01, PGT121 and PGT151 [40][41][42][43][44]. Such bnAbs have been suggested to arise, in part, due to the breakdown in tolerance pathways that would otherwise delete and/or inactivate such virus-specific B-cells.
While several HIV-1 bnAbs showed cross-reactivity in binding assay, only N6, a member of the VRC01 class of HIV-1 bnAbs, was able to neutralize SARS-CoV-2 pseudovirus though it failed to neutralize authentic SARS-CoV-2 virus in a cytopathic effect-based neutralization assay. While pseudoviruses are a great tool to understand viral biology in a biosafe environment, one of the key limitations with pseudoviruses is the enhanced neutralization

N6 failed to neutralize authentic SARS-CoV-2 virus. A) SARS-CoV-2 infected Vero-E6 cells monolayer after 72 hours post infection is shown as virus control B) Uninfected Vero-E6 cells monolayer after 72 hours showing complete absence of CPE is shown as representative cell control.
Incubation of SARS-CoV-2 virus with 20 μg/ml of C) N6, D) II62 scFv-Fc and E) Anti-RBD IgG respectively, followed by adsorption for 60 minutes on Vero E6 cells. CPE was observed after 72 hours. No significant reduction in the CPE (neutralizing activity) was observed for N6 & II62 mAb which indicates these antibodies did not block SARS-CoV-2 infection. The purified Anti-RBD IgG was used as assay positive control and no CPE was observed (upto 0.1μg/ml) which means virus is completely blocked by neutralizing antibodies and did not show infection on Vero-E6 cells Scale, 500px. https://doi.org/10.1371/journal.ppat.1009958.g005

PLOS PATHOGENS
HIV-1 bnAbs cross-neutralize SARS-CoV-2 susceptibility. Several HIV-1 bnAbs have been shown to potently neutralize several HIV-1 pseudoviruses but fail to neutralize infectious molecular clones and/or primary isolates of HIV-1 [45][46][47]. It is noteworthy that N6 recognizes HIV-1 in an unusual orientation and neutralizes HIV-1 isolates that are typically resistant to other VRC01 class bnAbs [48]. Furthermore, it can tolerate absence of key CD4bs antibody contact residues across the length of heavy chain and can tolerate escape mutations that typically provide resistance to HIV-1 from other CD4bs bnAbs. Of note, N6 has an unprecedented degree of somatic hypermutation (31% in heavy and 25% in light chain at the nucleotide level).
One of the key limitations of the current study is the lack of a detailed structural analysis of cross-reactive bnAbs with SARS-CoV-2. A detailed structural analysis of N6 binding with SARS-CoV-2 will provide an insight of the exact nature of epitope-paratope interaction that contribute to the cross-reactivity observed herein. While the current work was designed to show the ability of polyreactive and somatically hypermutated antibodies generated in chronic HIV-1 infection to bind to new and emerging pathogens such as SARS-CoV-2, our results provide impetus for future detailed structural studies.
Most HIV-1 vaccine candidates are in the stage where they typically induce tier 1B or 2 responses against autologous and heterologous viruses in rabbits and non-human primates [49][50][51]. A germline targeting HIV-1 candidate immunogen (eOD-GT8) which was designed

PLOS PATHOGENS
HIV-1 bnAbs cross-neutralize SARS-CoV-2 to prime VRC01 class CD4bs directed antibodies has been described and the frequencies and affinity of B cells from healthy HIV-1 uninfected individuals recognizing this germline-targeting immunogen showed its suitability as a candidate human vaccine prime [52,53]. Naive Bcells that recognised eOD-GT8 had L-CDR3 sequences that matched several VRC01 class bnAbs, suggesting B-cells with light chain sequences for VRC01 class exist at high frequency. Based on the above observations and the availability of sera from these immunized animals, and findings herein of the ability of HIV-1 CD4bs directed bnAbs to inhibit SARS-CoV and SARS-CoV-2 pseudovirus infection, it is pertinent that the immune sera be tested for binding and neutralization of SARS-CoV-2. Furthermore, detailed structural studies should be taken with N6 to identify its epitope and neutralization determinants, which can be used to engineer its variants as effective SARS-CoV-2 therapeutics.
Collectively, our findings highlight the ability of HIV-1 specific bnAbs and polyclonal plasma to cross-react with the newly emerged SARS-CoV-2. Understanding cross-reactive neutralization epitopes of antibodies generated in divergent viral infections can provide key evidence for engineering so called super-antibodies (antibodies that can potently neutralize diverse pathogens with similar antigenic features). Recently, using a rapid affinity maturation strategy, Zhao et.al., engineered the CR3022, a neutralizing antibody against SARS-CoV-1, to potently neutralize SARS-CoV-2 and highlighted the potential of engineering approaches that can be leveraged to refocus an existing neutralizing antibody to target a related but resistant virus [54]. Furthermore, generating libraries using the starting sequence of such cross-reactive antibodies using approaches such as rapid affinity maturation, can provide candidate antibodies in a short frame of time upon which synthetic variants can be generated in the face of future pandemics.

Ethics statement
The study was approved by the institute ethics committee of All India Institute of Medical Sciences (IEC/NP-295/2011 & IEC/59/08.01.16).

Study design
The current study was designed to assess the cross-reactivity of HIV-1 broadly neutralizing antibodies and plasma antibodies from children with chronic HIV-1 infection against the SARS-CoV-2. After written informed consent from guardians, blood was drawn in 3-ml EDTA vials, and plasma was aliquoted for plasma neutralization assays and binding ELISAs.

Bacteria
E. coli DH5α, DH10β and STBL3 for propagation of plasmids were cultured at 37˚C (30˚C for STBL3) in LB broth (Sigma-Aldrich) with shaking at 220 rpm.

Plasma from children with chronic HIV-1 infection
Well-characterized plasma sample from ten children that had shown potent neutralization titre against a 12-virus global panel of HIV-1 isolates from previous studies in our lab were selected.

Recombinant protein production and purification
SARS-CoV and SARS-CoV-2 ectodomain and RBD constructs were transiently transfected in Expi293F cells at a density of 2 million cells/mL using polyethylenimine and expression plasmids at a molar ratio of 3:1 and purified from clarified transfected culture supernatants 4-days post-transfection using Ni 2+ -NTA affinity chromatography (GE Life Sciences). Proteins were eluted from the column using 250 mmol/L imidazole, dialyzed into phosphate buffered saline (PBS), pH 7.2 and concentrated using Amicon 10-kDa (RBD) and 100-kDa (S2P ecto ) Amicon ultra-15 centrifugal filter units (EMD Millipore). Protein concentration was determined by the Nanodrop method using the protein molecular weight and molar extinction coefficient as determined by the online ExPASy software (ProtParam).

Binding ELISA
96-well microtiter plates were coated overnight with 2 μg/ml of purified SARS-CoV S2P ecro , SARS-CoV-2 S2P ecto , SARS-CoV RBD and SARS-CoV-2 RBD. Plates were blocked with 1% BSA for 3 hours. Monoclonal antibodies were added at a starting concentration of 10 μg/ml, with 11-point titration, and incubated for 2 hours at room temperature. Horseradish peroxidase conjugated goat anti-human IgG was used as secondary antibody and TMB substrate was used for color development. Absorbance at 450 nm was measured using a spectrophotometer. CR3022 (a nAb isolated from a convalescent SARS-CoV patient, which has been shown to cross-react with SARS-CoV-2), and CC12.1 (a SARS-CoV-2 neutralizing antibody isolated

PLOS PATHOGENS
HIV-1 bnAbs cross-neutralize SARS-CoV-2 from a convalescent donor) were used as positive control while two antibodies targeting the envelope glycoprotein of simian immunodeficiency virus (SIV) were used as negative control.

Neutralization assays
Full length spike proteins of all seven coronaviruses were co-transfected with an HIV-1 backbone and helper plasmid expressing firefly luciferase and for SARS-CoV-1 and -2, serine protease TMPRSS2 (CMV-Luc, RΔ8.2 backbone plasmid, pTMPRSS2) in 1.25 x 10 5 HEK293T cells for 72 hours. Post-transfection, culture supernatants were harvested, filtered and stored at -80˚C. For determination of neutralization potential of bnAbs against SARS-CoV-2, eightpoint titration curves with 2-fold serial dilution starting at 10 μg/ml, were performed. For neutralization potential of pediatric plasmas with chronic HIV-1 infection against all seven coronaviruses, eight-point titration curve starting at plasma dilution of 1:100 and going upto 1:51200 were used. Serially diluted bnAbs and plasmas were mixed with respective pseudotyped viruses for 1 hour at 37˚C. pseudovirus/bnAb combinations were then added to 293T-ACE2 cells pre-seeded (24-hours) at 20,000 cells/well. For plasma samples, plasma/pseudovirus combination was added to Vero-E6 cells (20,000 cells/well). After 48-72 hours, supernatant was removed and luminescence was measured on Tecan luminescence plate reader using Bright Glow reagent. The percent infectivity was calculated as ratio of relative luminescence units (RLU) readout in the presence of bnAbs and plasmas normalized to RLU readout in the absence of mAb and plasmas. The half maximal inhibitory concentrations (IC50 for antibodies, ID50 for plasmas) were determined using 4-parameter logistic regression (Graph-Pad Prism version 8.3). CC12.1, a neutralizing antibody against SARS-CoV-2, was used as positive control. Amphotropic envelope protein from Vesicular Stomatitis Virus (VSV-G) and murine leukemia virus (MLV) were used as positive virus infection controls.

Biolayer interferometry analysis of the SARS-CoV-2 RBD binding affinity with N6 bnAb
Biolayer interferometry was performed using an Octet Red96 instrument (ForteBio, Inc.). A 5 μg/ml concentration of SARS-CoV-2 RBD-His was immobilized on a Ni-NTA coated biosensor surface. The baseline was obtained by measurements taken for 30 s in running buffer (1x PBS, 0.1% BSA and 0.02% Tween-20), and then, the sensors were subjected to association phase immersion for 300 s in wells containing N6 bnAb diluted in running buffer. Then, the sensors were immersed in running buffer for 600 s to measure dissociation. Biosensor was then regenerated by dipping it in EDTA followed by nickel sulfate solution. The mean K on , K off and apparent K D values of the SARS-CoV-2 RBD binding affinity for N6 bnAb were calculated from all the binding curves based on their global fit to a 1:1 Langmuir binding model.

Cytopathic effect (CPE) based SARS-CoV-2 neutralization assay
CPE-based neutralisation assays were carried out as previously described in Parray et al; 2020. In brief, 1 x 10 2 TCID 50 isolate USA-WA1/2020 virus was passaged once in Vero-E6 cells and then treated with antibody dilutions ranging from 20 to 0.3μg/ml for 90 minutes before adsorption on Vero-E6 cells for 1 hour. Following cell washing, DMEM supplemented with 2% FBS was added. After 4-5 days of incubation at 37˚C with 5% CO2, the presence of cytopathic effect (CPE) in cells was observed using a microscope. Non-infected Vero-E6 cells and purified IgG from RBD immunized mice were used as a positive control, and infected Vero-E6 cells and II62 scFv-Fc were used as a negative control.

Cell surface binding assay
1.25 x 10 5 HEK293T cells seeded in a 12-well plate were transiently transfected with 1.25 μg of SARS-CoV-2 S full-length protein using PEI-MAX. 48 hours post-transfection, cells were harvested and per experimental requirement, distributed in 1.5 ml microcentrifuge tubes. For monoclonal antibody staining, 10 μg/ml of antibody was used and titrated 2-fold in staining buffer. 100 μl of primary antibody (HIV-1 specific monoclonals) were added to HEK293T cells expressing SARS-CoV-2 S, and incubated for 30 minutes at room temperature. CR3022 and CC12.1 were used as positive control for surface binding assay. After washing, 100 μl of 1:500 diluted PE conjugated mouse anti-human IgG Fc was added, and after 30-minute incubation, a total of 50,000 cells were acquired on BD LSRFortessa X20. Data was analyzed using FlowJo software (version v10.6.1).

Statistics and reproducibility
All statistical analyses were performed on GraphPad Prism 8.3. A p-value of <0.05 was considered significant. Neutralization assays were performed in triplicates and repeated thrice. Average IC50 values are shown and used for statistical comparisons. Binding ELISAs were performed in duplicates and repeated thrice. Average OD 450 values were used for plotting curves. Surface binding assay was performed thrice and average PE-MFI (phycoerythrin-median fluorescence intensity) values were used for plotting curves.