Inhibition of stearoyl-CoA desaturases suppresses follicular help T- and germinal center B- cell responses

Abstract

Stearoyl-CoA desaturases (SCD) are endoplasmic reticulum (ER)-associated enzymes that catalyze the synthesis of the monounsaturated fatty acids (MUFAs). As such, SCD play important roles in maintaining the intracellular balance between saturated fatty acid (SFAs) and MUFAs. The roles of SCD in CD4⁺ T-helper cell responses are currently unexplored. Here, we have found that murine and human follicular helper T (T_FH) cells express higher levels of SCD compared to non-T_FH cells. Further, the expression of SCD in T_FH cells is dependent on the T_FH lineage-specification transcription factor BCL6. We found that the inhibition of SCD impaired T_FH cell maintenance and shifted the balance between T_FH and follicular regulatory T (T_FR) cells in the spleen. Consequently, SCD inhibition dampened germinal center B-cell responses following influenza immunization. Mechanistically, we found that SCD inhibition led to increased ER stress and enhanced T_FH cell apoptosis in vitro and in vivo. These results reveal a possible link between fatty acid metabolism and cellular and humoral responses induced by immunization or potentially, autoimmunity.

Introduction

Follicular helper T (T_FH) cells are specialized effector CD4⁺ T helper cells that promote the formation of germinal centers (GCs) and the production of class-switched high-affinity immunoglobulins [1]. T_FH cells are required for the generation of antibody-mediated protection against microbes, but aberrant T_FH responses may cause autoimmunity [2]. Therefore, it is important to understand the positive and/or negative regulators of T_FH-cell responses for the proper induction during vaccination or the inhibition of T_FH responses during autoimmunity.

T_FH differentiation in vivo is a multifactorial multistep process. The transcription factor BCL6 is considered the master regulator for the differentiation of T_FH cells and subsequent GC responses [3, 4]. Interestingly, recent evidence suggests that cell metabolic processes may play important roles in modulating T_FH development in vivo. BCL6 was shown to repress glycolysis in activated CD4⁺ T cells [5, 6]. Consistently, it has been demonstrated that T_FH cells exhibit diminished glycolysis and mitochondrial respiration compared to T_H1 cells [7]. However, a recent study has also suggested mTOR kinase complex 1 (mTORC1)-related glucose metabolism is required for the T_FH generation in vivo [8]. Therefore, the roles of glucose metabolism in negative and positive regulation of T_FH differentiation and/or maintenance remain to be elucidated.

Besides glucose metabolism, emerging evidence has suggested that lipid metabolism also plays important roles in regulating T-helper cell responses. The inhibition of acetyl-CoA carboxylase 1 (ACC1), a key enzyme for fatty acid synthesis, reduces human and mouse T_H17 cell development in favor of regulatory T (T_reg) cell differentiation [9]. Simvastatin, the inhibitor for 3-hydroxy-3-methyglutaryl (HMG)-CoA reductase, has been reported to promote Treg differentiation whereas it suppresses T_H17 development [10]. Furthermore, the suppression of fatty acid synthase (FASN) during T-cell activation reduces the development of CD4⁺ T cell memory [11]. Currently, the roles of lipid metabolism in T_FH differentiation have not been explored. Stearoyl-CoA desaturases (SCD) are endoplasmic reticulum (ER) proteins that catalyze the synthesis of monounsaturated fatty acids (MUFAs) from saturated fatty acids (SFAs). Besides the needs of MUFAs for the synthesis of complex lipids, such as diacyglycerols (DG), triglycerides (TGs), and cholesterol esters, MUFAs also have important functions in cell signaling and membrane fluidity [12]. Therefore, SCD are highly regulated and conserved enzymes that play important roles in regulating obesity, insulin resistance, inflammation, and cancer progression [13, 14]. Humans have two SCD isoforms (SCD1 and SCD5), while mouse has four (SCD1–4) that share homology with human SCD1, but not human SCD5 [15, 16]. The roles of SCD in regulating adaptive lymphocyte responses have not been explored.

Given the importance of lipid synthesis and metabolism in adaptive immunity, we examined the roles of SCD in T_FH responses. We found that SCD1 was highly expressed in both human and murine CD4⁺CD44⁺PD1^hiCXCR5^hi T_FH cells compared to non-T_FH cells. BCL6-deficient CD4⁺ T cells exhibited diminished Scd1 gene expression. We showed that in vivo inhibition of SCD activity impaired the maintenance of T_FH and GC B cells. Furthermore, SCD inhibitor increased apoptosis in T_FH cells in vivo and in vitro. Finally, the inhibition of SCD promoted the expression of ER stress genes whereas SCD1 overexpression showed opposite results in T_FH cells. Our results suggest that BCL6-mediated SCD expression promotes T_FH-cell maintenance and efficient GC B-cell responses in vivo.

Results

SCD inhibitor suppresses T_FH and GC B-cell responses

To explore the function of lipid metabolism in regulating T_FH and GC B-cell responses, we immunized influenza X31 virus and then treated mice with Statin (an inhibitor of HMG-CoA reductase that is required cholesterol synthesis) [17], C75 (an inhibitor of FASN that is required for the generation of long-chain fatty acids) [18], or A939572 (an inhibitor of SCD1) [19]. Since lipid metabolism can potentially regulate APC function and early T-cell activation [20, 21], we administered these inhibitors at day 4 postimmunization to minimize the effects of these inhibitors on early T-cell priming. We then checked T_FH (CD4⁺CD44⁺PD1^hiCXCR5^hi), non-T_FH (CD4⁺CD44⁺PD1^low/−CXCR5^low/−), and GC B- (B220⁺FAS⁺ GL-7⁺) cell responses in the spleen at 14 days postimmunization. We found that both Statin and A939572 markedly diminished T_FH magnitude, but only moderately decreased non-T_FH responses (Fig. 1A). A939572, but not Statin, also diminished GC B-cell responses following influenza immunization (Fig. 1A). In contrast, C75 treatment had no effects on T_FH or GC B-cell responses (Fig. 1A). SCD inhibitor did not alter the total number of CD4⁺ T and B cells (Fig. 1B), nor the production of T_H1 cytokine IFN-gamma in CD4⁺ and CD8⁺ T cells following antigen-specific stimulation (Fig. 1C). These results suggest that SCD inhibition selectively impairs T_FH but not T_H1-cell formation following influenza immunization.

Figure 1.

SCD inhibition suppresses T_FH but not T_H1 responses. WT B6 mice were immunized with X31 and treated with indicated inhibitors starting at 4 days postimmunization (d.p.i.). (A) Frequencies and absolute cell numbers of splenic T_FH, non-T_FH, GC B cells were measured by flow cytometry at 14 d.p.i. (B) Representative dot plot and absolute cell number of splenic CD4⁺ T or B cells at 14 d.p.i. (C) IFN-γ production by CD4⁺ or CD8⁺ T cells were measured through intracellular staining (ICS) following restimulation with NP_311–325 (CD4⁺ T-cell epitope) or NP_366–374 (CD8⁺ T-cell epitope) peptides in vitro at 14 d.p.i. Combined results in A are obtained from three independent experiments (two to five mice per group). Representative data in B and C are obtained from at least two independent experiments (four to five mice per group). Results are given as mean ± SE with one-way ANOVA (A) or mean ± SEM unpaired t-test (B, C). *indicates significant differences (p < 0.05), **p < 0.01, ***p < 0.001, and ****p < 0.0001.

SCD are highly expressed by human and murine T_FH cells

To identify the expression of SCD in the T_FH and non-T_FH cells, we first analyzed a published microarray data set (GEO# GSE50391) of CD45RO⁺CXCR5^high, CD45RO⁺CXCR5^int, and CD45RO⁺CXCR5⁻ from human tonsil [22]. We found that CD45RO⁺CXCR5^high GC-T_FH cells exhibit significantly higher SCD1 expression compared to non-T_FH cells (Fig. 2A). Mouse T_FH cells isolated from KLH/CFA immunized animals also exhibited higher levels of Scd1 and Scd2 relative to non-T_FH cells [23] (GEO# GSE40068) (Fig. 2B). In another published RNAseq dataset, T_FH cells, particularly LN and splenic T_FH cells, had enhanced Scd1 and Scd2 genes expressed relative to conventional non-T_FH cells (T_Conv) [24] (GEO #: GSE124883) (Fig. 2C). However, the expression of Fasn was comparable between non-T_FH and T_FH cells (Fig. 2B and C). We then sorted murine T_FH and non-T_FH (CD4⁺CD44⁺PD1⁻CXCR5⁻) from spleens of day 14 postinfluenza (X31) immunized mice. We examined Scd1, Scd2, and Fasn expression in T_FH and non-T_FH by quantitative RT-PCR. We found that T_FH cells exhibited drastic enhanced levels of Scd1 and modest increased Scd2 expression compared with non-T_FH cells (Fig. 2D). Together, these results reveal that both murine and human T_FH cells have enhanced Scd gene expression.

Figure 2.

Human and mouse T_FH express higher levels of SCD mRNA. (A) Relative SCD1 expression in published microarray data (GEO# GSE50391) CD45RO⁺CXCR5⁻ (non-T_FH), CD45RO⁺CXCR5^int (T_FH), and CD45RO⁺CXCR5^hi (GC-T_FH) cells from human tonsils (five donors per group). (B) Relative expression of Fasn, Scd1, and Scd2 in published microarray data (GEO# GSE40068) from BCL6^hiCXCR5⁺ (T_FH) and BCL6⁻CXCR5⁻ (non-T_FH) cells of the mice immunized with KLH in CFA (two samples each pooled from three mice per group). (C) Relative expression of Fasn, Scd1, Scd2, and Bcl6 in published RNAseq data (GEO# GSE124883) from LN, spleen, or blood T_FH and conventional T cells (T_conv) following NP-OVA immunization (three samples each pooled form ten mice per group). (D) Relative expression of Fasn, Scd1, and Scd2 in sorted splenic T_FH (PD-1⁺CXCR5⁺) and non-T_FH (PD1⁻CXCR5⁻) cells isolated from X31 immunized mice at 14 d.p.i. Representative results of (D) obtained from two independent experiments (five mice per groups). Results are given as mean ± SEM with one- (A) or two- (C) way ANOVA or unpaired t-test (D). *indicates significant differences (p < 0.05), **p < 0.01, ***p < 0.001, and ****p < 0.0001.

BCL6 promotes SCD expression in T_FH cells

To determine whether T_FH master transcription factor BCL6 is involved in Scd gene expression, we measured Scd1 and Scd2 expression in vitro cultured WT or BCL6-deficient (BCL6 KO) CD4⁺ T cells under the T_FH-like polarization conditions. Both Scd1 and Scd2 were diminished in BCL6 KO CD4⁺ T cells (Fig. 3A). Western blot analysis confirmed that BCL6 deficiency decreased SCD1 protein levels in CD4⁺ T cells under T_FH polarization conditions (Fig. 3B). We next preactivated naive CD4⁺ T cells for 2 days and then transduced the cells with BCL6-expressing retroviruses. Scd1 and Scd2 were measured in the isolated transduced cells. The BCL6-overexpressed cells exhibited significantly enhanced Scd1, and modestly upregulated Scd2 expression (Fig. 3C). To probe the relationship between BCL6 and SCD1 in vivo, we adoptively transferred WT OTII TCR-transgenic T cells or BCL6 KO-OTII cells into CD45.1 congenic mice. Following adoptive transfer, mice were immunized with X31-OTII virus. At day 14 postimmunization, OTII cells were sorted for measuring Scd1 and Scd2 expression. Similar to in vitro culture results, the expression of Scd1 was impaired in BCL6 KO-OTII cells (Fig. 3D). Together, these results demonstrated that BCL6 promotes the expression of SCD expression in T_FH cells.

Figure 3.

BCL6 promotes SCD1 expression. (A) Relative expression of Scd1 and Scd2 in WT or BCL6 KO CD4⁺ T cells under the T_FH cell polarizing condition for 4 days. (B) Western blot analysis of SCD1 protein levels in WT or BCL6 KO CD4⁺ T cells under the T_FH cell polarizing condition for 5 days. (C) Scd1, Scd2, and Bcl6 expression in the BCL6-retroviral transduced CD4⁺ T cells under the T_FH cell polarizing condition. (D) WT or BCL6 KO-OTII cells were transfer into CD45.1 congenic mice at one day before X31-OTII virus immunization. After day 14 postimmunization, the OTII cells were sorted and then Scd1 and Scd2 were measured by quantitative RT-PCR (three mice per group in two independent experiments). Representative graphs are obtained from at least two (A, B, and C) independent experiments (two mice per group). Results are given as mean ± SEM with unpaired t-test: *p < 0.05; **p < 0.01; and ***p < 0.001.

SCD inhibitor promotes accelerated contraction of T_FH and GC B cells

Next, we examined whether SCD inhibition is required for development and/or maintenance of T_FH and/or GC B cells. To this end, the frequencies and cell numbers of T_FH and GC B cells were measured 0, 8, 10, and 14 days post-X31 immunization following SCD inhibition. SCD inhibitor significantly diminished both T_FH and GC B-cell responses at day 14, but not at day 8 or 10 postimmunization (Fig. 4A and B). Notably, T_FH-cell numbers begun to diminish from day 10 postimmunization and SCD inhibitor accelerated the contraction of T_FH cells in vivo. The expression of BCL6 and ICOS in the T_FH cells between vehicle (DMSO)- and SCD inhibitor-treated group were comparable (Supporting Information Fig. 1). These data suggested that SCD inhibition may not suppress the formation of T_FH and GC B cells, but promotes accelerated contraction of T_FH cells (i.e. inhibition of T_FH maintenance). Recently, T-follicular regulatory (T_FR) (CD4⁺CD44⁺PD1⁺CXCR5⁺Foxp3⁺) cells have been reported as a major regulator repressing excessive T_FH and GC B-cell responses [25]. We found that SCD inhibition impaired T_FH but not T_FR-cell magnitude (Fig. 4C and D). As such, the frequencies of T_FR cells were inflated relative to T_FH cells following SCD inhibition (Fig. 4C). This indicates that SCD inhibition alters the balance of T_FH and T_FR responses. Consequently, influenza-specific IgG antibody production was impaired in the serum following SCD inhibition (Fig. 4E). Together, these results suggest that SCD is required for the maintenance of T_FH and GC B cells through the regulation of the balance between T_FH and T_FR cells in the spleen.

Figure 4.

Inhibition of SCD suppresses T_FH and GC B maintenance, but not T_FR responses. (A-B) Frequencies and absolute cell number of splenic T_FH and GC B cells in the mice immunized with X31 virus and treated with SCD inhibitor at indicated d.p.i. (C) Frequencies and (D) cell numbers of splenic T_FR and T_FH cells were measured in the mice immunized with X31 virus and treated with SCD inhibitor at 14 d.p.i. (E) The titer of influenza-specific IgG was measured by ELISA from blood serum at 14 d.p.i. Representative data are obtained from at least two independent experiments (two to four mice per group). Results are given as mean ± SE with two-way ANOVA (A and B) or unpaired t-test (C and E): *p < 0.05; and ***p < 0.001.

Inhibition of SCD promotes apoptosis and ER stress gene expression

To probe the underlying mechanism, we examined T-cell apoptosis and proliferation following SCD inhibitor treatment. At day 8 post-X31 immunization, active caspase 3/7 activities in T_FH cells following SCD inhibitor treatment were similar to vehicle (DMSO)-treated group. However, at day 14 postimmunization, caspase 3/7 activities were significantly higher in T_FH cells following SCD inhibition compared to DMSO-treated group (Fig. 5A). In contrast to apoptosis, SCD inhibition had no effects on the proliferation of T_FH cells at both day 8 and 14 post-X31 immunization (Fig. 5B). In vitro cell apoptosis was significantly enhanced following SCD inhibitor treatment in a dose-dependent manner under T_FH-polarizing conditions (Fig. 5C). Taken together, these data demonstrate that SCD activity promotes T_FH-cell viability. A recent study demonstrated excessive accumulation of SFAs following SCD inhibition promotes ER stress-related cell apoptosis [26]. Thus, we examined whether SCD inhibition promoted the expression of ER stress-related genes. The expression of activating transcription factor (Atf) 4 and Atf3 were increased in CD4⁺ T cells following SCD inhibition (Fig. 5D). ATF4 protein levels were also markedly increased following SCD inhibition (Fig. 5E). Furthermore, the expression of multiple ER stress-related genes including Atf4, Atf3, Perk, and sXbp1, were downregulated in CD4⁺ T cells transduced with SCD1-expressing retroviruses (Fig. 5F). These results suggest that SCD promote T_FH-cell survival likely due to their regulation of ER stress genes.

Figure 5.

SCD inhibition promotes T_FH apoptosis and ER stress gene expression. (A) Caspase 3/7 activity and (B) proliferation (indicated by Ki67 staining) of splenic T_FH cells from the mice immunized with X31 virus and treated with SCD inhibitor at 8 and 14 d.p.i. (C) Naïve CD4⁺ T cells were stimulated under T_FH polarization condition and treated with increased doses of SCD inhibitor, active caspase 3/7⁺ cells were measured at day 4 postactivation. (D) The expression of ER stress-related genes in CD4⁺ T cells under T_FH polarization with or without SCD inhibitor (10 μM) was determined by qRT-PCR. (E) ATF4 protein level in CD4⁺ T cells under T_FH polarization with or without SCD inhibitor. (F) The expression of ER stress-related genes in CD4⁺ T cells under T_FH polarization with or without SCD1 transduction was determined by qRT-PCR. Combine data (A and B) are obtained from two independent experiments (two to five mice per group). Representative results (C and E) are shown from two independent experiments (two mice per group). Results are given as mean ± SEM with unpaired t-test: *p < 0.05.

Enhanced Scd1 expression promotes T_FH responses in vivo

To identify whether SCD expression contributes cell-intrinsic maintenance of T_FH cells, we transduced WT OTII cells with control or SCD1-overexpressing retroviruses. We then adoptively transferred the OT-II cells into CD45.1 congenic mice and immunized the mice with X31-OTII virus (Fig. 6A). We found that SCD1 overexpression did not significantly increase T_FH-cell numbers, possibly due to the slightly diminished overall OTII T-cell magnitude following SCD1 overexpression (Fig. 6B). However, the frequency of T_FH cells within OTII cells was elevated following SCD1 overexpression, suggesting that ectopic SCD1 expression promotes T_FH differentiation and/or maintenance relative to non-T_FH cells (Fig. 6B). Ectopic SCD1 expression in OTII cells did not affect overall GC B-cell responses (Fig. 6C), which is likely due to the relatively low numbers of the transferred T_FH cells compared to endogenous T_FH cells. Together, our data suggest that ectopic SCD1 expression facilitates T_FH-cell responses relative to non-T_FH cells.

Figure 6.

Enforced SCD1 expression in CD4⁺ T cells promotes T_FH responses in vivo. Mock or SCD1-retroviral transduced OTII cells were adoptively transferred into CD45.1 congenic mice. Then the mice were immunized with X31-OTII. (A) Schematics of experimental design. (B) Frequency of T_FH cells in OTII cells and cell numbers of T_FH and non-T_FH cells in CD45.2⁺ donor cells at 14 d.p.i. (C) Frequency and cell numbers of GC B cells in spleen at 14 d.p.i. Results were pooled from three independent experiments (three to four mice per group). Results are given as mean ± SEM with unpaired t-test: *p < 0.05.

Discussion

In this report, we show that human and murine T_FH cells exhibit higher levels of the lipogenic gene expression in the form of SCD. The expression of Scd1 is regulated in a BCL6-dependent manner both in vivo and in vitro condition. Importantly, a SCD inhibitor, A939572, diminished the magnitude of T_FH and GC B-cell responses whereas it promoted the frequency of T_FR cells in vivo following immunization with X31-influenza virus. Therefore, our results have revealed a nonredundant role of SCD and lipid desaturation in regulating T_FH maintenance and GC B-cell responses.

Excessive or prolonged ER stress, also known as unfolded protein response (UPR), suppresses normal cell growth through inhibition of translation and ultimately apoptosis. Lipotoxicity, induced by overload of lipids and their metabolites, disturbed ER homeostasis and caused metabolic abnormalities and cell death [27]. Knockdown of SCD1 enhanced the accumulation of SFAs, resulted in increased expression of ER stress-related genes such as C/EBP homologous protein (CHOP), glucose-regulated protein 78 (GRP78), and splicing of Xbox-binding protein 1 (sXBP1) [26]. However, these enhanced ER stress-induced genes and cell death were rescued by treatment of unsaturated fatty acid in Hela cells [28]. Interestingly, oleate and palmitoleate, the major products of SCD1 activity, have different metabolic functions. The oleate, but not palmitoleate, normalized the excess ER stress in SCD1 liver-specific deficient mice [12]. Thus, SCD expression in T_FH cells may contribute to cellular homeostasis through the maintenance of balance between SFAs and MUFAs. It must be noted that even though SCD inhibition induces the upregulation of ER stress genes in T cells, and excessive ER stress has been associated with cellular apoptosis. Evidence presented in this study linking ER stress and T_FH apoptosis is correlative in nature. Future mechanistic investigations are needed to test whether excessive ER stress is the actual cause of enhanced T_FH apoptosis following SCD inhibition in vitro and in vivo.

Nevertheless, SCD1 have been reported to be highly expressed in several human cancer cells including breast, prostate, lung and ovary [29]. SCD inhibitor decreased the proliferation and increased apoptosis in clear cell renal cell carcinoma (ccRCC) in vivo and in vitro [30]. SCD1 inhibition promoted autophagy-induced apoptosis accompanied with autophagy formation in the human hepatocellular carcinoma (HCC) cells via AMPK activation [31]. Thus, SCD inhibition might be utilized to targeted cell therapy against SCD-expressing cells including malignant cells and T_FH cells.

Enhanced T_FH cells have been reported to be associated with the development and progression of many autoimmune diseases [32]. Recently, as the potential target of autoimmune disease therapy, inhibition of lipid metabolism gene has been reported in human patients and animal models. Inhibition of FASN reduced the development of T_H17 cells, resulting in diminished EAE disease progression [33]. Simvastatin, inhibitor of HMG-CoA reductase, diminished the serum TNF-α and CRP levels in RA patients [34]. In our model, compared to C75 (fasn inhibitor) and fluvastatin, SCD inhibitor (A939572) exhibited superior activities in terms of the inhibition of T_FH and GC B-cell responses. Interestingly, it has been reported that obese humans have higher levels of SCD1 expression and MUFA contents compared to lean objects [35]. SCD1 has been associated with obesity-related abnormal metabolic diseases such as hyperlipidemia, nonalcoholic liver disease, and type 2 diabetes in animal model and human [12]. These reports suggest that T-cell lipid metabolism, particularly SCD pathway, may represent a promising target to dampen autoimmunity, especially in those obesity-associated autoimmunity [36].

In this study, we show that inhibition of SCD reduces the number of T_FH and GC B-cells postinfluenza virus immunization. It is possible that SCD inhibitor may have off-target effects in vivo to attenuate T_FH and/or GC B-cell responses. Furthermore, SCD inhibitor may directly inhibit B cells or other cell types, such as DCs, to diminish T_FH maintenance. Particularly, B cells and DCs have been shown to be important in T_FH-cell generation and/or maintenance [1, 37, 38]. Therefore, future studies employing cell-type specific deletion of SCD in different cell types (such as T cells, DCs, and/or B cells) are warranted to address these potential limitations of our studies. However, as shown in Figure 6, Scd1 overexpression promoted T_FH responses in vivo, indicating that the promotion of the T cell-intrinsic SCD1 expression facilitates T_FH responses in vivo.

In conclusion, we found SCD activity promotes T_FH-cell maintenance by inhibiting their apoptosis and sustaining GC B-cell responses. Therefore, SCD may be considered as a novel target to modulate excessive T_FH responses during vaccination and/or autoimmunity.

Materials and methods

Mouse and immunization

Mouse studies were approved by the Institutional Animal Care and Use Committee at the Mayo Clinic or Indiana University School of Medicine (IUSM). All methods were performed in accordance with the relevant guidelines and regulations. WT C57/BL6, CD45.1 congenic, OTII and CD4-cre mice were originally purchased from the Jackson Laboratory and bred in house. BCL6 KO mice (CD4-cre Bcl6 fl/fl) were generated by crossing CD4-cre transgenic mice with Bcl6 fl/fl mice [39]. BCL6 KO OTII mice were generated by crossing OTII transgenic mice to BCL6 KO mice. Influenza immunization was achieved i.p. with Influenza H3N2 A/X31 or X31-OTII (~1.2 × 105 pfu/mouse). A939572 (6 mg/kg, BioFine International, WA, USA), C75 (15 mg/kg, Cayman chemical, MI, USA), or Fluvastatin (20 mg/kg, Cayman chemical) was injected to the X31 immunized mice from day 4 to day 13 for daily via i.p. route.

Flow cytometry

Splenic T_FH and T_FR cells were stained with antibodies against mouse CD4 (Clone: RM4–5, Biolegend, San Diego, CA, USA), CD44 (Clone:IM7, Biolegend), PD-1 (Clone: 29F.1A12, Biolegend), CXCR5 conjugated with biotin (Clone: 2G8, BD Biosciences, San Jose, CA, USA), streptavidin conjugated with APC (Biolegend), Ki67 (Clone: SolA15, Thermo Fisher Scientific, Waltham, MN, USA), Foxp3 (Clone: FJK-16s, Thermo Fisher Scientific), BCL6 (Clone: K112–91, BD Biosciences), ICOS (Clone: C398.4A, Biolegend), and GC B cells were stained with antibodies against mouse B220 (Clone: RA3–6B2, Biolegend), GL7 (Clone: GL7, Biolegend), and CD95 (FAS, Clone: SA367H8, Biolegend). For BCL6-retroviral transduced cell gating, anti-mouse H2K antibody (Clone: 36–7-5, BD Biosciences) was used and for SCD1-retroviral transduced cell gating anti-human CD4 antibody (Clone: OKT4, Biolegend) was employed. For intracellular staining, cells were restimulated with NP 311–325 peptide or NP 366–374 peptide (AnaSpecs, CA, USA) in the presence of 1 μg/mL of monensin (BD Biosciences) and brefeldin A (Biolegend) for 4 h. Then, cells were stained with surface antibodies and fixed/permeabilization buffers (Biolegend) for staining with anti-IFN-γ antibody as previously described [40]. To detect cell apoptosis, the CellEvent Caspase 3/7 Green Detection Reagent (Thermo Fisher Scientific) was added to the cells with complete media for 30 min in 37°C and the caspase 3/7 positive cells were measured on an Attune NXT flow cytometer (Life Technologies, Carlsbad, CA, USA). For T_FH-cell sorting, CD4⁺ T cells were enriched from spleen with CD4 microbeads (Clone: L3-T4, Miltenyi Biotec, San Diego, CA, USA). The enriched CD4⁺ T cells were stained with antibodies to define T_FH cells as described in flow methods below. And the cells were sorted on a FACSAria (BD Biosciences,). All the flow cytometry analysis was conducted accordingly to the “Guidelines for the use of flow cytometry and cell sorting in immunological studies” [41].

In vitro T-cell culture

Naïve CD4⁺ T cells were isolated from WT (C57BL/6) or Bcl6 fl/fl CD4-Cre (BCL6 KO). Cells were cultured with T_FH-polarizing condition (precoated with anti-CD3 [1 μg/mL], soluble anti-CD28 [2 μg/mL], anti-IFN-γ [20 μg/mL], anti-IL-2 [10 μg/mL] antibodies, rmIL-21 [20 ng/mL], and rmIL-6 [20 ng/mL]) for 4–5 days.

Quantitative RT-PCR

RNA was extracted by RNA miniprep kit (Sigma-Aldrich) according to the manufacturer’s instructions. RT was performed using random primers (Invitrogen, Grand Island, NY, USA) and MMLV (Invitrogen). cDNAs were amplified using SYBR Green PCR Master Mix (Applied Biosystems, Grand Island, NY, USA) using quantistudio 3 PCR system (Applied Biosystems). Data were generated with the comparative threshold cycle method by normalizing to hypoxanthine phosphoribosyltransferase (HPRT), as we previously reported [42].

Retroviral transduction

Naïve polyclonal CD4⁺ T or OTII cells were preactivated with precoated anti-CD3 (5 μg/mL) and soluble anti-CD28 (2 μg/mL) antibodies for 2 days. Then cells were transduced with bicistronic retroviruses through spin transduction (2,500 rpm, 90 min) as reported [43]. After transduction, cells were cultured with T_FH polarized condition for an additional 2 days.

Adoptive cell transfer

OTII, BCL6 KO-OTII cells, control retroviral-transduced or SCD1-retroviral transduced OTII cells were adoptively transferred into CD45.1 congenic mice. One day post-transfer, A/X31-OVA (X31-OTII) (~5.0 × 105 pfu/mouse) was immunized for 14 days via i.p. route.

ELISA

Levels of X31-specific IgG antibody in the blood serum were measured by ELISA. In brief, 96-well plates (Biolegend) were precoated with 100 μL/well of 7 × 10⁴ PFU of X31 virus in PBS overnight at 4°C. After blocking with PBS containing 3% of BSA (Sigma) for 1 h at room temperature (RT), each serum sample was serially diluted in blocking buffer to each well and incubated for 2 h at RT. After three times wash with wash buffer (0.05% of Tween 20 in PBS), HRP-conjugated goat anti-mouse IgG (Promega, Madison, WI, USA) was added into each well. After incubation for 2 h at RT, tetramethylbenzidine (TMB, Thermo Fisher Scientific) was added to develop the color and then the reaction was stopped by adding 2 M H₂SO₄. The absorbance at wavelength 450 nm was measured by an ELISA reader (Molecular Devices, Sunnyvale, CA, USA). Endpoint titers were calculated as the titer of dilution that gave an absorbance value (O.D.) of 0.2.

Western blot analysis

Naïve CD4⁺ T cells were purified from WT or BCL6 KO mice with naïve CD4⁺ T-cell isolation kit (Miltenyi Biotec). Then, the cells were activated and cultured under the T_FH polarizing condition as above. The cultured cells were lysed in sample lysis buffer (Cell Signaling Technology) with a protease inhibitor cocktail (Thermo Scientific) and incubated on ice for 30 min. Proteins in the supernatants were heat denatured, separated by 4% to 12% gradient Bis-Tris gel (Invitrogen) and transferred into a nitrocellulose membrane (Bio Rad). Membrane was blocked with 5% of nonfat milk in TBST for 1 h at RT. After washing with TBST (Tris-buffered saline with 0.1% Tween 20), membrane was incubated with following primary antibodies for overnight: Rabbit-anti SCD1 (Cell signaling), rabbit-anti ATF4 (Cell signaling) or mouse-anti actin (Santa Cruz). Membrane was then rinsed with TBST and incubated with the HRP-conjugated secondary antibodies (Promega) for 1 h at RT. ECL kit (Thermo) was used for film (Fisher Scientific) development.

Statistical analysis

Graphs were generated by Graph Pad Prism software. Statistical significance was evaluated by calculating ρ-values using one/two-way ANOVA or Student’s t-test. Significance between the groups was judged based on p-value< 0.05 (two tailed).

Abbreviations:

ACC1

acetyl-CoA carboxylase 1

Atf

activating transcription factor

DG

diacyglycerols

FASN

fatty acid synthase

mTORC1

mTOR kinase complex 1

MUFAs

monounsaturated fatty acids

RT

room temperature

SCD

Stearoyl-CoA desaturases

SFAs

saturated fatty acids

T_Conv

conventional non-T_FH cells

T_FH

follicular helper T

T_FR

follicular regulatory T

TG

triglycerides