Association of follicular helper T and follicular regulatory T cells with severity and hyperglycemia in hospitalized COVID-19 patients

ABSTRACT We aimed to determine the levels of follicular helper T (Tfh) and follicular regulatory T (Tfr) cells in COVID-19 patients and determine whether their levels correlated with disease severity and presence of hyperglycemia. This study was carried out in 34 hospitalized COVID-19 patients and 20 healthy controls. Levels of total circulating Tfh, inducible T-cell costimulator (ICOS)+ activated Tfh, and Tfr cells were assessed in all participants by flow cytometry. Total CD4+CXCR5+ Tfh cells and ICOS+Foxp3-activated Tfh cells increased and ICOS+Foxp3+ Tfr cells decreased in COVID-19 patients, especially in diabetic patients and those with severe disease. Activated ICOS+ Tfh cells were directly correlated with lactate dehydrogenase, D-dimer, ferritin, and respiratory rate and inversely correlated with the partial pressure of carbon dioxide. COVID-19 is associated with marked activation of Tfh cells and a profound drop in Tfr cells, especially in severe and diabetic patients. Future studies on expanded cohorts of patients are needed to clarify the relationship between SARS-CoV-2 and acute-onset diabetes.


Introduction
A new coronavirus strain (SARS-CoV-2) caused an outbreak of severe pneumonia in China, in 2019. In February 2020, WHO called the disease caused by this new coronavirus COVID-19 and later announced that COVID-19 outbreak was classified as pandemic [1, [2]].
The pathophysiology of the remarkably contagious disease COVID-19 remains obscure [3]. Several immune mechanisms, in coordination with defense mechanisms, are essential for eliminating viral infections. Immune responses mediated by T cells are critical in controlling viral infections. Upon viral infection, strong type I inflammation promotes the virus-specific CD4+ T cell differentiation into primarily T helper 1 (Th1) and follicular helper T cells (Tfh) [4].
Tfh cells are featured by the high expression of the chemokine receptor C-X-C motif chemokine receptor 5 (CXCR5) [5]. Likewise, an inducible T cell costimulator (ICOS) is essential for production and maintenance of Tfh cells [6]. The production of neutralizing antibodies (nAbs) in the germinal center (GC) requires Tfh cell help; therefore, these cells represent a major line of defense to pathogens [7]. They are specialized in supporting B cell survival, differentiation, and proliferation within the GCs in the follicles of secondary lymphoid organs via direct contact and secretion of soluble signals [8].
Tfh and Tfr exert opposing regulatory actions in GC responses, and an equilibrium of their functions is essential for maintaining immune homeostasis [12]. Moreover, animal experiments have revealed that neutralizing antibodies (nAbs) are critical components of protective immune responses in infection and following vaccination [13]. Therefore, a better understanding of the role of Tfh and Tfr in the formation of nAbs is fundamental for the design of future vaccines.
The development of severe COVID-19 has been associated with several comorbidities. Therefore, studying the relationship between immune cells and different severity markers could be of utmost importance in the early and prompt identification and control of severe cases [12]. Diabetes is the most recorded comorbidity in patients with COVID-19. SARS-CoV-2 might increase the risk of hyperglycemia in patients with and without a prior history of diabetes. However, it is not clear whether the virus can induce type-1 or type-2 diabetes or alternatively causes a new atypical form of diabetes. It's also unknown if recovering COVID-19 patients experience a higher risk of developing newonset diabetes or its complications going forward [14].
Data about the alterations of Tfh and Tfr in SARS-CoV-2 infection, especially in diabetic patients are so far limited. Thus, we aimed to determine the levels of follicular helper T (Tfh) and follicular regulatory T (Tfr) cells in COVID-19 patients and determine whether their levels correlated with disease severity and presence of hyperglycemia.

Materials and methods
Our study was approved by ethical committee of Assiut University (IRB no. 17300436). All participants provided informed written consent.
This case-control study was conducted in a sample of 34 hospitalized COVID-19 patients and 20 age-and sex-matched healthy controls. The diagnosis of COVID-19 was based primarily on reversetranscription polymerase chain reaction (RT-PCR) on throat swab samples. All healthy controls were clinically free of disease and showed no evidence of infection by history, clinical examination, and complete blood count (CBC). In addition, none of the controls had a history of close contact with a COVID-19-positive patient in the two weeks preceding the study. A schematic flow chart of the study protocol is presented in Figure 1.
Patients were subjected to a collection of a detailed medical history and clinical examinations. Laboratory tests, including CBC and COVID-19-related tests such as D-dimer, ferritin, CRP, and LDH, were performed for all participants. Oxygen saturation (SpO 2 ) and partial pressure of carbon dioxide (PaCO 2 ) were measured. Fasting blood glucose (FBG), hemoglobin A1C (HbA1C), renal function (serum urea and creatinine), and liver function (albumin, total protein, direct and total bilirubin, prothrombin time, and concentration) were also measured. Diagnosis of pneumonia was based on respiratory and systemic symptoms and clinical signs (fever, cough, dyspnea, rapid shallow breathing), as well as radiography. Severe cases were characterized as those who had clinical signs of pneumonia and met at least one of the following criteria: severe respiratory distress, tachypnea >30 breaths/min, SpO2 < 90%, or greater than 50% lung infiltrates (as evident from CT) [15][16]. For all participants, blood samples were collected into EDTA-containing vacutainer tubes nearly 10-15 days from diagnosis to assess the Tfh and Tfr cells.

Statistical analysis
Data are expressed as mean and standard error. The Mann-Whitney U test was used to compare the non-parametric groups, and the Spearman rank-order correlation was employed to evaluate the association between different variables. p-value was considered significant if <.05.

Results
The main features of COVID-19 patients are displayed in Table 1 Approximately 59% of the participating patients had a non-severe COVID, whereas 41% had severe cases. Overall, the group of COVID patients showed a very high mean FBG (222.8 mg/dl), and 23 (67.6%) of the patients were diabetic. All but one non-diabetic patient had a higher FBG than average, with a mean of 129.5 ± 8 mg/dl. The average HbA1c level was 6.7% ± .2%, and 36% of the severe cases had uncontrolled diabetes, with HbA1c > 7%. The mean TLC was 11.6 ± 1 × 10 9 /L -levels above 10.6 x 10 9 /L indicate neutrophilia -and the mean lymphocyte count was 2 ± .4 × 10 9 /L. In severe cases, the TLC reached 12.7 ± 4 × 10 9 / L (notably, numbers below .7 × 10 9 /L indicate lymphopenia). CRP, ferritin, D dimer, and LDH were markedly elevated. Renal functions were noticeably impaired, and the liver function tests showed hypoalbuminemia and elevated direct bilirubin level (8.2 ± 2 µmol/L). The mean prothrombin time and concentration were within the normal ranges; however, in severe cases, prothrombin time reached 15 ± .9 sec. The respiratory rate was increased, reaching 40.4 ± 4 in severe cases; in addition, decreased oxygen saturation (84.5% ± 2%) and PaCO 2 (27.4 ± 1 mmHg) were observed.
The levels of CD4+ T helper and Tfh cells were similar in the severe and non-severe COVID-19 patients, as shown in Figure 4 The activated CD4 +CXCR5+ICOS+Foxp3-Tfh cells were significantly elevated in severe COVID-19 patients compared to patients with non-severe disease (p < .0001). In contrast, the levels of resting CD4+CXCR5+ICOS-Foxp3-Tfh cells and CD4+CXCR5+ICOS+Foxp3+ Tfr were decreased in severe patients compared to non-severe patients (p = .009 and p = .009, respectively).
As shown in Figure 6, levels of total CD4+CXCR5+ Tfh cells were significantly increased in the nondiabetic COVID-19 patients compared with healthy Data were presented as number (percentage) or *mean ± SE (normal reference ranges. FBG fasting blood glucose, CBC complete blood count, TLC total leukocyte count, CRP C-reactive protein, LDH lactate dehydrogenase, ALT alanine aminotransferase, AST aspartate aminotransferase, SpO 2 Oxygen saturation, PaCO 2 partial pressure of carbon dioxide. volunteers (p = .003) and diabetic COVID-19 patients (p = .01). In contrast, the frequency of CD4+CXCR5 +ICOS+Foxp3-activated Tfh cells was remarkably increased in diabetic patients than control group and non-diabetic COVID patients. A lower frequency of CD4+CXCR5+ICOS+ Foxp3+ Tfr cells was found in the diabetic patients compared to controls (p = .01) and non-diabetic COVID-19 patients (p = .02).

Discussion
In the mouse models infected with SARS-CoV-2, CD4+ T-cell responses showed a critical role in limiting infection [17]. Following viral infection, Tfh cells promote the production of specific nAbs by forming long-lasting memory B lymphocytes and plasma cells [18]. Circulating Tfh (cTfh) cells were found to have a similar phenotype and function to the bona fide counterparts in GCs in secondary lymphoid tissues [19].
We assessed the levels of Tfh and Tfr cells in hospitalized COVID-19 patients and healthy controls. Despite the profound reduction observed in CD4+ percentage, substantial rises were detected in the frequencies of both activated (ICOS+) and resting (ICOS-) CD4+CXCR5+ Tfh.
Levels of CD4+ T helper and total Tfh cells in our study were almost the same in both the severe and non-severe cases. While most of the CD4 +CXCR5+ Tfh cells were activated (ICOS+) in severe cases, a higher proportion of Tfh cells were resting (ICOS-) in non-severe cases. Oja and colleagues stated that the highest antibody titers and elevated frequencies of PD-1+CXCR5+ cTfh were detected in critical COVID 19 patients [30]. Additionally, Gong et al. found that patients with severe disease had a higher frequency of effectormemory cTfh cells and a lower frequency of central-memory cTfh cells than controls and non-severe disease patients [29].
On the other hand, Zhang and colleagues detected no difference in the frequency of total circulating Tfh cells between COVID-19-convalescent individuals and healthy controls. Meanwhile, they reported lower expression of CXCR3+ in total Tfh cells of both severe and non-severe COVID-19-convalescent individuals compared with the healthy controls. Additionally, the severe group showed a higher percentage of CXCR3+ Tfh cells than the non-severe group. CXCR3+ Tfh cells also positively correlated with neutralizing antibody titers proposing a propable key role of CXCR3+ Tfh cells in the early production of neutralizing antibody titer in patients recovering from COVID-19 [31].
Also, defective generation of Bcl-6+ Tfh cell and dysregulated humoral immune induction early in COVID-19 disease was reported in another study, providing a mechanistic explanation for the limited durability of antibody responses in coronavirus infections, and suggest that achieving herd immunity through natural infection may be difficult [32]. Further studies on Tfh subsets and Bcl-6 expression shall provide insight into their role in control of SARS-CoV2 infection.
In contrast, significantly lower percentage of Tfr cells was detected in our COVID-19 patients compared with healthy volunteers, especially in those with severe cases. Tfr cells can inhibit humoral immunity mediated by Tfh cells, constraining the GC response needed for high-affinity antibody production [9]. Along this line, our results showed an inverse relationship between the CD4+CXCR5+ICOS +Foxp3-activated Tfh and CD4+CXCR5+ICOS+Foxp3+ Tfr cells.
Data on the levels of circulating Tfr in patients with COVID-19 is scarce. Gong et al. [29] reported a reduced frequency of circulating Tfr cells (cTfr) in COVID-19 convalescent patients, with a similar decline observed in those recovering from non-severe and severe disease. They also found inverse relationships between cTfr cells and COVID-19 IgM and IgG titers.
In our patients, activated ICOS+ Tfh cells were directly correlated with LDH, D-dimer, and ferritin, as well as respiratory rate, and an inverse correlation with the PaCO 2 . Relationships between the resting Tfh and the Tfr subsets and the severity, clinical, and laboratory data were opposite to those observed for activated Tfh cells. This shows that these factors may have an impact on circulating activated Tfh and Tfr homeostasis, which may affect the subsequent antibody production.
Whereas we did not find an association between any of these cells and SpO 2 [29], observed that the frequency of central memory cTfh cells was positively related to PaO 2 /FiO 2 and that effector memory cTfh cells were negatively associated with PaO 2 /FiO 2 , with no strong association observed between PaO 2 /FiO 2 and cTfr cells.
Our COVID patients had a remarkably high mean FBG, and 23 of them had diabetes. Furthermore, all but one nondiabetic patient had a higher FBG than average. Infection tends to increase stress-induced insulin resistance, leading to elevated blood glucose levels even in non-diabetic patients [33]. Numerous studies have described COVID-19 patients are at increased danger of developing glucose metabolic disorders. COVID-19 patients with hyperglycemia may be at a higher risk of more severe outcomes and death from COVID-19 even if they do not have preexisting diabetes. Serious metabolic complications of diabetes have been reported in COVID-19 patients, as diabetic ketoacidosis and probable new-onset diabetes [34,35].
SARS-CoV-2, like other coronaviruses, binds to angiotensin-converting enzyme 2 (ACE2) receptors, which are expressed in a variety of tissues, including pancreatic cells [36]. Fasting hyperglycemia and acute-onset diabetes were found to be more prevalent in patients with SARS-CoV-1 pneumonia than in those with non-SARS pneumonia. These findings suggest that COVID-19 may have a diabetogenic effect independent of the stress response observed during severe illness. However, it is unknown whether these diabetogenic effects will subside or persist following infection recovery [37].
While the available data have not established an association between SARS-CoV-2 exposure and new-onset diabetes [38], hypothesized that SARS-CoV-2 exposure accelerates the onset of type 1 diabetes. They noticed a rise in new cases of type 1 diabetes in children through the pandemic's peak years, and some of these children had active COVID or had a history of prior exposure to the virus [38].
Knowledge of the opposite roles of Tfh and Tfr cells in regulating GC responses, maintaining immune homeostasis requires a balance of their actions [12]. Additionally, uncontrolled Tfr or Tfh activity might lead to the production of autoreactive B cells and auto-antibodies [39]. Interestingly, our results showed that diabetic COVID-19 patients demonstrated a notable rise in the frequency of activated Tfh and a profound drop in Tfr percentage compared with both the controls and non-diabetic patients. We postulate that the Tfh/Tfr imbalance observed in COVID-19 may be implicated in the complex pathogenesis of COVID-related diabetes. However, our study has some limitations. Specifically, the age and the use of some immune-modulating agents for COVID-19, such as steroids, tocilizumab (an IL-6 receptor antagonist), and COVID-19 convalescent plasma, may have confounded the findings. Also, assessment of Tfh subsets, Bcl-6 expressing Tfh and levels of neutralizing antibodies might give a clearer insight on their role in COVID-19 pathogenesis and their relation with recovery and hyperglycemia development. It would have added a lot if Tfh and Tfr levels were also assessed in a group of diabetic non-COVID patients.

Conclusion
Our results suggest that COVID-19 is associated with marked activation of Tfh cells and a profound drop in Tfr cells, especially in diabetic patients and patients with severe disease. Future studies on expanded cohorts of patients are needed to clarify the relationship between SARS-CoV-2 and acute-onset diabetes.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Data availability statements
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request after taking a permission from our ethical committee.