Deficient humoral responses and disrupted B-cell immunity are associated with fatal SFTSV infection

Severe Fever with Thrombocytopenia Syndrome (SFTS), an emerging infectious disease caused by a novel phlebovirus, is associated with high fatality. Therapeutic interventions are lacking and disease pathogenesis is yet to be fully elucidated. The anti-viral immune response has been reported, but humoral involvement in viral pathogenesis is poorly understood. Here we show defective serological responses to SFTSV is associated with disease fatality and a combination of B-cell and T-cell impairment contribute to disruption of anti-viral immunity. The serological profile in deceased patients is characterized by absence of specific IgG to viral nucleocapsid and glycoprotein due to failure of B-cell class switching. Expansion and impairment of antibody secretion is a signature of fatal SFTSV infection. Apoptosis of monocytes in the early stage of infection diminishes antigen-presentation by dendritic cells, impedes differentiation and function of T follicular helper cells, and contributes to failure of the virus-specific humoral response.

Severe fever with thromocytopenia syndrome (SFTS) is a severe infectious disease caused by the SFTSV virus. In some areas oft he world, the diesease is associated with a 10-30% frequency of fatality. The reasons why most patients survive, but 10-30% succumb to the disease is poorly understood. Previous studies mainly focussed on innate immune cells. In the present manuscript, it is shown that there is a striking lack of an anti-virus IgM and IgG response in all deceased patients (except a transient anti-Gn IgM antibody response in some deceased patients), whereas all patients who survived showed strong IgM and IgG responses against viral proteins. This is associated with a number of other immunological differences between patients who resolved the infection or died, namely i) persistent high virus titres in deceased patients compared to virus clearance in surviving patients, ii) strongly reduced frequencies of naive B cells, iii) increased frequencies of plasmablasts in the deceased patient group, iv) increased apoptosis of monocytes in the deceased patient group, v) decreased expression of MHC class II and costimulatory receptors on B cells from deceased patients, vi) alterations in TFH cells in the peripheral blood between the patient groups, and vii) alterations in cytokine levels betwen the patient groups.
Main criticism 1) In figure 3, only the relative frequencies of the various B cell subsets are presented, but it remains unclear, which subsets show changes in total numbers between the groups. For example, is there really a dramatic loss of naive B cells in the deceased patients, or is the lower frequency only due to higher number of other B cells, with total number sof naive B cells actually unchanged? Thus, the total numbers of B cells should also be presented.
2) At first glance it is unexpected and may appear even contradictory that the deceased patients have no specific antibody response, but higher frequencies of plasmablasts. This many perhaps be explained by a general B cell activation towards plasmablast differentiation in the absence of a functional specific response. However, as this is a central issue, the plasmablasts need to be better characterized. It is strange that in the deceased patients about 70% and in the surviving patients 30% of plasmablasts are IgM-IgG-. A plasmablast without strong Ig expression is highly questionable. Perhaps, there is an induction of specific class-switching to IgA? There are also socalled IgD-only B cells and plasmablasts/plasma cells in the human that produce IgD in the complete absence of IgM. Can they play a role here? Moreover, the definition of the cells as plasmablasts relies only on the staining pattern CD27-high, CD38-high. Perhaps, in the acute viral infection some activated B cells transiently upregulate CD27 and CD38 without being plasmablasts? Thus, it is recommended that the CD27++CD38++ cells are also stained for IgA and IgD, and that the cells are analyzed for key plasmablast/plasma cell markers (upregulation of transcription factors BLIMP1, IRF4, XBP1) by FACS or RT-PCR.
3) One of the peripheral blood B cell subsets that showed a significant difference between the two patient groups are CD19+CD27-IgD-B cells. In healthy individuals, it is well known that these are primarily IgG+CD27-B cells, which account for 10-20% of all IgG memory B cells. However, because of the massive B cell compartment alterations present in the patients with acute virus infection, it remains uncertain what the cells defined by lack of CD27 and IgD really are. Are they indeed IgG+ memory B cells? Can it be excluded that they are, for example, immature B cells (IgM+IgD-CD27-), released from the bone marrow, or that they are polyclonally activated naive B cells that downregulated IgD upon stimulation? It would likely be sufficient to perform flow cytometric analysis in a few patients and healthy controls to stain the IgD-CD27-B cells also for IgG and IgM to clarify their identity better. CD21 could be included to distinguish immature from unusual mature B cells, in case many of them are IgM+.
Minor points a) The complete lack of an anti-humoral immune response against the virus in those patient who died is impressive. Some of the other alterations described may explain or contribute to the failed immune response. However, the authors need to consider that these are so far only correlations. Some of the interpretations need to be tuned down. b) In figure 3B, the horizontal lines above the columns should not extend over the healthy controls, because they are distinct individuals that do not belong the patient groups. c) On page 10, line 269, it is mentioned that an anti-B220 antibody was used to eliminate B cells from a DC isolation procedure. However, whereas B220 is indeed used as a pan-B cell marker in the mouse, this marker (recognising CD45R) is expressed only by a subset of human B cells. Please explain. d) In figure 5, necrotic cells are defined as PI+ cells, and apoptotic cells as annexin V+ cells. However, one typically sees annexin V+PI-and annexin V+PI+ cells. How were the double positive cells grouped? As necrotic or as apoptotic? This needs to be better described. e) At least in the pdf file available for review, the immunofluorescence pictures in figure 5A are very dark and hardly interpretable. The figure should be improved.
f) The Discussion is with nearly six pages very long. It would profit from shortening and focussing.
Reviewer #2 (Remarks to the Author): The manuscript by Song et al showed the association of fatal SFTSV cases with a suppression of virus specific antibodies due to a lack of class switching in the plasmablast cells (PB). The authors attributed this phenomenon as the causative reason for disease aggravation and mortality, and attempted to define a mechanism that suggest a disruption of affinity maturation involving TfH, B cells and myeloid DC (mDC) activity in the lymph node or spleen. While the phenomenon of virus specific antibodies suppression is convincing, the experiments conducted to link this to dysfunctional TfH, mDC and B cells were loosely connected. Furthermore, the complex pathway that occurs within the secondary lymphoid organs and not the blood proposed in this study would require more than the surveillance of immune markers on the cells. These data should be supplemented by more conclusive functional in vitro assays or in vivo models. Comments 1. The introduction contains too much irrelevant information and could be shortened.
2. Wrong reference cited for line 104 -106. That study used non-monocyte derived cell lines for infection.
3. The following points should be addressed for Fig 1 and Fig 2: a. To confirm that antibodies in the survive patients play a role in the clearance of virus and protection, functional neutralization assays with the patients' serum samples should be performed. Otherwise this remains an association and its impact should not be over claimed. b. Fig 1, the profiling of NP specific antibody (Ab) is by ELISA and profiling of Gn specific Ab is by western blot. Data of Gn by ELISA is left as unpublished data. To provide better consistency, it would be better to present the ELISA data of Gn together with ELISA data of NP in the main figure and shift the data of Gn western blot to the supplementary with the data from Supp Fig 1. c. Authors should perform clustering of the data in color chart of the ELISA data. As it stands, the current figure makes it very difficult to read the data. d. Fig 2B, It is not clear how the authors selected the data to generate the correlation curve. It  would be more meaningful to convert all the data points collected in Fig 2B into the correlation plot  with just the viral load versus the Ab titre. 4. The following points should be addressed for data presented in Fig 3 and Fig 4: a. It is possible that the disease could induce leukopenia in patients and diminish the overall B cells numbers in the blood. Authors should provide data of blood counts that includes leucocytes and granulocytes. In addition, the numbers of total B cells should be compared between groups. Without this information, data in Fig 3C is not meaningful, as it is possible that while the % of a specific subset relative to total B cells is the same between survived and deceased groups, the actual numbers could be vastly different due to overall reduction of B cells. b. The rationale of having 2 different gating strategies for PB and MB was not clearly explained. Authors should consider including the CD38 marker into the analysis presented in Fig 3. In this way, the analysis of B cells subset definition could be standardized. c. Fig 4, the ELISPOT was done using total B cells. Authors should consider doing the ELISPOT with sorted PB to demonstrate the lack of virus specific Ab secreting PB cells. d. In lines 235 -239, authors concluded that the observation of increase proportion of IgM-IgG-in the deceased group meant this subset is responsible for abrogated serological response. This is similar to saying that this subset of cells is immunosuppressive, which is not true. In addition, the comparison of this subset was done in % relative to total B cells. Data of % relative to total leucocytes would be more meaningful.
5. The following points should be addressed for data presented in Fig 5: a. Author presented 2 effects to the mDCs: apoptosis/necrosis and the suppression of costimulatory molecules that could suppress Ag presentation. The manuscript would flow better if the text from lines 258 -283 is separated into 2 sub-sections based on these 2 effects. b. Regarding the apoptosis assay, authors allowed the PBMCs to adhere and subsequently performed staining on all adherent cells with the assumption they are monocytes. This is incorrect. The authors would need to either proof that this method produces homogenous monocytes in the adherent cells or add in markers to confirm the identity of the cells undergoing apoptosis. A more direct alternative method would be to track apoptosis using a annexin V kit by flow cytometry using the same gating strategy used for the mDCs. c. In the second part, authors profiles the co-stimulatory molecules and proposed that the Ag presentation of mDCs is impaired. This is inadequate and more direct function assays would be required to drive this point. Particularly, authors should sort out mDCs from the 2 groups of patients, and assess their ability to present Ag to T cells either by CFSE tagging to the T cells or by ELISOT assay. d. Lastly, an important fundamental concept must be addressed. Myeloid DC (also known as conventional DC) is derived from CDP and not monocytes. 6. Fig 6, authors proposed that B cells Ag presentation to TfH is vital for TfH expansion and the alteration of CD80 and HLA-DR disrupted this process. This unique process happens exclusively in the lymph nodes and spleen in a small percentage of B cells within the interfollicular zone2. It is questionable if the profiling of co-stimulatory molecules in the peripheral B cells would be sufficient to drive this conclusion. Direct immunofluorescent staining of B cells and TfH co-localization in the interfollicular zone would be required to show the abrogation of this interaction. 7. Fig 7, where authors argued the loss of function TfH in deceased patients would need to be further addressed: a. Representative pseudoplot provided highlighted only 2 IL-21 positive staining events. This is too low to be conclusive if we consider the noise to actual signal ratio of a generic flow cytometry machine. In view of this, authors should consider using IL-21 gene expression from sorted pTfH cells. b. Following the reference cited by the authors, functional role of pTfH in patients was shown by their ability to induce phenotypic changes in naïve B cells through co-culture system. They should consider using a similar system to assign a function difference to the pTfH beyond the profiling if IL21.
8. Lastly, it would make better sense to change the order of some of the data. Particularly, pTfH data should come after showing Ab suppression, followed by the B cell Ag presentation and mDC data which explain the change in TfH.
Reviewer #3 (Remarks to the Author): In this manuscript, Song and colleagues identify the defective serological response that correlated with fatal SFTSV infection. By focusing their study on surviving and deceased patients following SFTSV infection, the authors show that deceased patients lacked a crucial IgG mediated response to viral proteins (nucleoprotein and glycoprotein) due to a B cell class switch failure. Additionally, they characterized severe apoptosis of peripheral monocytes in patients with the fatal outcome when compared to surviving patients or healthy controls. Finally, they performed cytokine profiling of all patients and found a correlation in severity with increased levels of serum IFN-γ, IL-10, IL-6, and TNF-α. Overall, this data provides an in-depth analysis of the B-cell mediated deficiencies that lead to SFTS severity, which has not been shown before in the field.
This manuscript could be improved by making revisions in response to the following minor comments.

Minor Comments:
The rationale for this manuscript is well founded based on a previous study by the same authors in 2017, where they found the reduction of both IFN-β and IL-1β in patients that have a severe or fatal outcome of the disease. Furthermore, they observed a reduction of myeloid dendritic cells and an overall suppression of TLR3 in the primary targets for SFTSV, myeloid dendritic cells, which are primary targets for SFTSV infection, and an overall suppression of TLR3 in DCs. Hence, the authors postulated that antigen presentation was inhibited in patients that had a fatal outcome following SFTSV infection. While the premise for this manuscript focuses on deficiencies concerning B-cell immunity following SFTSV infection in fatal patients, additional data that includes the state of the overall immune system of these patients (T-cells, B-cells, NK cells) could be insightful and further elucidate the mechanisms of SFTSV pathogenesis.
Furthermore, examination of viral load in B-cells could be explaining impaired B-cell-dependent immunity due to virus replication, since there has been a study to report the expression of DC-SIGN, which is known as the cellular receptor for SFTSV entry, on B cells (Rappocciolo et al. 2005 PLoS Pathogens).
Abstract: The authors state in the abstract that SFTSV is an "emerging infectious disease caused by a novel member of phlebovirus." While this is correct, the sentence should state that it is part of the phlebovirus genera.
Figure presentation: Several of the figures where statistical analysis was performed for the healthy, survived, and deceased groups lack indication of which groups are statistically significant. This can be improved by adding bars indicating which groups are significant. Additionally, the western blot in figure 1 shows the failure of the sera from all patients to react with Gc, which is indeed interesting. A positive control using commercial antibodies or an antibody against the tag epitope of the purified protein could be used as an additional control to ensure equal expression between Gc and Gn.

Minor points
a) The complete lack of an anti-humoral immune response against the virus in those patient who died is impressive. Some of the other alterations described may explain or contribute to the failed immune response. However, the authors need to consider that these are so far only correlations.
Some of the interpretations need to be tuned down.

Response:
We fully agreed with the reviewer's opinion. To illustrate the roles of the virus-specific antibodies, we performed neutralization assays to determine the virus inhibitory activity of sera from recovered patients. In addition, we generated high titer Gn-specific anti-serum in camels immunized with Gn expressed in mammalian cells..  Figure 2C in the revised manuscript.
b) In figure 3B, the horizontal lines above the columns should not extend over the healthy controls, because they are distinct individuals that do not belong the patient groups.

Response:
We have corrected this mistake in Figure 3B.
c) On page 10, line 269, it is mentioned that an anti-B220 antibody was used to eliminate B cells from a DC isolation procedure. However, whereas B220 is indeed used as a pan-B cell marker in the mouse, this marker (recognising CD45R) is expressed only by a subset of human B cells.
Please explain.

Response:
We agreed with the reviewer that B220 expresses only on a subset of human B cells. According to a previous study, there are about 80% human peripheral CD19 + B cells expressing B220 [5]. Actually, although a small part of B cells expresses low level of CD11c, its frequency in periphery blood is very small and this small subset of CD11c + B cells is located in T /B cell border in spleen [6]. Another reason we chose B220 was that, besides CD11c + MHCⅡ + B cells which need to be excluded from our gating strategy of CD11c + mDCs, recent study demonstrated that thymus-derived αβ TCR + cells could also express CD11c and MHC class II molecules [7]. Furthermore, earlier evidence has revealed that during acute virus infection, such as Ebola virus, large proportion of T cells were activated [8]. Meanwhile, another study by Bleesing et al found that activated human T cells also up-regulated the expression of B220 [9]. Therefore, using B220 marker, we can exclude most of the B cells and activated T cells. Taken together, we make a brief gating strategy of mDC by B220 marker.

d) In figure 5, necrotic cells are defined as PI + cells, and apoptotic cells as annexin V + cells.
However, one typically sees annexin V + PIand annexin V + PI + cells. How were the double positive cells grouped? As necrotic or as apoptotic? This needs to be better described.

Response:
We have performed modifications of Figure 5A to improve its visual effect in the revised manuscript.
f) The Discussion is with nearly six pages very long. It would profit from shortening and focusing

Response:
The reviewer thinks the discussion is very long and should be shortened and focused. Therefore, we have revised the discussion according to the suggestion. For reviewer #2 1. The introduction contains too much irrelevant information and could be shortened.

Response:
We have shortened the introduction, and removed some less relevant information.
2. Wrong reference cited for line 104 -106. That study used non-monocyte derived cell lines for infection.

Response:
The correct reference has been used. 3. Revision for Fig 1 and Fig 2: a. To confirm that antibodies in the survive patients play a role in the clearance of virus and protection, functional neutralization assays with the patients' serum samples should be performed.
Otherwise this remains an association and its impact should not be over claimed.

Response:
The reviewer's point is well taken. We have performed neutralization analysis to establish the roles of virus-specific serum antibody in inhibiting virus infection. In addition, we also demonstrated Gn-specific antibody generated by immunizing camels with a mammalian expressed Gn protein in inhibin viral infection.
The data are presented in Figure 2C in the revised manuscript. Please also see Answer to Reviewer #1, Minor point a.
b. Fig 1,  c. Authors should perform clustering of the data in color chart of the ELISA data. As it stands, the current figure makes it very difficult to read the data.

Response:
The reviewer's suggestions are well taken. We moved the Gn Western blot to the supplementary (Fig S1), and instead presented dynamic ELISA data of Gn together with ELISA data of NP in the main figure in the revised manuscript.
Meanwhile, we performed clustering of the data in color chart of the ELISA results and generated dynamic profiles of virus-specific antibodies in Figure 1C and 1D in the revised manuscript.
d. Fig 2B,  Response: That's actually a false appearance caused by the superimposition of many different data points . We have altered the plot to generate a new correlation curve in the current Figure 2B. Detailed information of this new figure was described in the revised manuscript. 4. Revision for Fig 3 and Fig 4: a. It is possible that the disease could induce leukopenia in patients and diminish the overall B cells numbers in the blood. Authors should provide data of blood counts that includes leucocytes and granulocytes. In addition, the numbers of total B cells should be compared between groups.
Without this information, data in Fig 3C is not meaningful, as it is possible that while the % of a specific subset relative to total B cells is the same between survived and deceased groups, the actual numbers could be vastly different due to overall reduction of B cells. Authors should consider including the CD38 marker into the analysis presented in Fig 3. In this way, the analysis of B cells subset definition could be standardized.

Response:
The reason that two different gating strategies were used was that we applied a two-step approach to illustrate the regulation of B cells during acute phase of SFTS. In the first step (Figure 3), the combination of CD19/CD27/IgD as the B cell marker could distinguish the composition of peripheral B cells, such as naïve, memory and marginal zone-like B cells, as previously reported [3,4]. We then further defined the plasmablasts as CD19 + CD27 ++ IgDin the CD19 + CD27 + IgDgroup, since a previous study by Avery et al reported that increased expression of CD27 on human memory B cell correlated with their commitment to plasma cell lineage [5]. Due to the limitation of our FACS machinery (BD Aira II with six channels), CD38 was left to the second step gating. Therefore, in the second step (Figure 4), we included the CD38 marker in the analysis of plasmablast which are defined as CD27 high CD38 high B cells.
c. Fig 4,  d. In lines 235 -239, authors concluded that the observation of increase proportion of IgM-IgG-in the deceased group meant this subset is responsible for abrogated serological response. This is similar to saying that this subset of cells is immunosuppressive, which is not true. In addition, the comparison of this subset was done in % relative to total B cells. Data of % relative to total leucocytes would be more meaningful.

Response:
We agree with reviewer's opinion, and the word "responsible" here is actually not accurate. What we intended to say is that both the overwhelming proportion of IgM -IgG -PBs would be the results of the humoral response failure in fatal SFTSV infection. We have deleted the confusing sentence in the revised manuscript.
To maintain the consistency of the data presented in the whole article, we directly provided the numbers of CD27 + CD38 + and CD27 + CD38subsets in peripheral blood as elsewhere in this manuscript. 5. Revision for Fig 5: a. Author presented 2 effects to the mDCs: apoptosis/necrosis and the suppression of co-stimulatory molecules that could suppress Ag presentation. The manuscript would flow better if the text from lines 258 -283 is separated into 2 sub-sections based on these 2 effects.

Response:
The reviewer's suggestion is well taken, and we have separated the text into 2 sub-sections based on these 2 effects in the revised manuscript.
b. Regarding the apoptosis assay, authors allowed the PBMCs to adhere and subsequently performed staining on all adherent cells with the assumption they are monocytes. This is incorrect.
The authors would need to either proof that this method produces homogenous monocytes in the adherent cells or add in markers to confirm the identity of the cells undergoing apoptosis. A more direct alternative method would be to track apoptosis using a annexin V kit by flow cytometry using the same gating strategy used for the mDCs.

Response:
According to previous studies, the adherent cells produced by this method usually contain about 90% of monocytes [10,11], they might include a small part of dendritic cells and macrophage, even as well as a very minor fraction of CD3 + cells (＜3%) [10]. We have changed the term "monocytes" into "adherent cells of PBMC".
Considering the present study demonstrated that the apoptosis and necrosis rates of the adherent cells in the deceased patients were significantly higher than 10%, representing 32.6% and 18.3%, respectively, we could still infer that the monocytes in the deceased patients undergo the significant apoptosis and necrosis, as compared with in that the survived patients.
The reviewer's suggestion on using a annexin V kit by flow cytometry is an excellent suggestion. Unfortunately the patient PBMCs are cryogenically preserved in small quantity and the freeze-thawing has detrimental effects on APC and could induce apoptosis and necrosis of dendritic cells as previously described [12,13]. We are unable to carry out the experiments suggested by the reviewer. Instead, we isolated sufficient mDCs (purify>97%) from fresh peripheral blood of healthy donors C. In the second part, authors profiles the co-stimulatory molecules and proposed that the Ag presentation of mDCs is impaired. This is inadequate and more direct function assays would be required to drive this point. Particularly, authors should sort out mDCs from the 2 groups of patients, and assess their ability to present Ag to T cells either by CFSE tagging to the T cells or by ELISOT assay.

Response:
The reviewer's point is well taken. As mentioned in the last paragraph, there were some difficulties in performing functional assay using the cryo-preserved patients' samples. Therefore, we performed the allogenic DC/T-cell coculture and stimulation assay in vitro using the peripheral blood of healthy donors to investigate the impaired Ag presentation of mDCs caused by SFTS virus infection as previously described [15][16][17]. The proliferation rate of CFSE-tagged CD4 + T cells and the level of IFN-γ in the supernatant after allogenic DC stimulation are determined and presented in Figure 6I-6J. The results provided compelling evidence that Ag presentation function of mDCs was impaired due to SFTS virus infection.
d. Lastly, an important fundamental concept must be addressed. Myeloid DC (also known as conventional DC) is derived from CDP and not monocytes.

Response:
We fully agree with the reviewer's comment. Although monocyte derived DC (mo-DC) share some common phenotypic markers with mDC, including CD11c and MHC class II molecules, they actually originate from different precursor cells [18]. To avoid conceptual confusion between mDC and mo-DC, we have deleted the incorrect expression and corrected the sentence "Severe apoptosis of monocytes followed by the failure of myeloid DC differentiation" in the revised text. 6. Fig 6, authors proposed Figure 5A in the revised manuscript. To further address reviewer's concern, we have sorted pTfh cells of SFTS patients, and measured IL-21 gene expression by RT-PCR. The result presented in Figure 5D  were presented in Figure 5E-5F in the revised manuscript. 8. Lastly, it would make better sense to change the order of some of the data. Particularly, pTfH data should come after showing Ab suppression, followed by the B cell Ag presentation and mDC data which explain the change in TfH.

For reviewer #3
Minor Comments: The rationale for this manuscript is well founded based on a previous study by the same authors in 2017, where they found the reduction of both IFN-β and IL-1β in patients that have a severe or fatal outcome of the disease. Furthermore, they observed a reduction of myeloid dendritic cells and an overall suppression of TLR3 in the primary targets for SFTSV, myeloid dendritic cells, which are primary targets for SFTSV infection, and an overall suppression of TLR3 in DCs.
Hence, the authors postulated that antigen presentation was inhibited in patients that had a fatal outcome following SFTSV infection. While the premise for this manuscript focuses on deficiencies concerning B-cell immunity following SFTSV infection in fatal patients, additional data that includes the state of the overall immune system of these patients (T-cells, B-cells, NK cells) could be insightful and further elucidate the mechanisms of SFTSV pathogenesis.
Furthermore, examination of viral load in B-cells could be explaining impaired B-cell-dependent immunity due to virus replication, since there has been a study to report the expression of DC-SIGN, which is known as the cellular receptor for SFTSV entry, on B cells (Rappocciolo et al.

PLoS Pathogens).
Abstract: The authors state in the abstract that SFTSV is an "emerging infectious disease caused by a novel member of phlebovirus." While this is correct, the sentence should state that it is part of the phlebovirus genera. survived, and deceased groups lack indication of which groups are statistically significant. This can be improved by adding bars indicating which groups are significant. Additionally, the western blot in figure 1 shows the failure of the sera from all patients to react with Gc, which is indeed interesting. A positive control using commercial antibodies or an antibody against the tag epitope of the purified protein could be used as an additional control to ensure equal expression between Gc and Gn.

Responses:
1. To address the reviewer's questions on the state of the overall immune system of SFTS patients, such as T-cells, B-cells and NK cells, we analyzed the subsets of lymphocytes from SFTS patients by flow cytometry, and found that the main subsets of peripheral lymphocytes, including total T cells, total B cells and NK cells, manifested no significant differences among patients' groups regardless of the disease severity (presented in the