Abstract

   Normal glucose metabolism is critical to immune function but the
   effects of short‐term hyperglycemia on immunity are not well described.
   To study this phenomenon, we induced hyperglycemia in healthy subjects
   for 2 h with intravenous dextrose and octreotide. An RNA‐seq analysis
   of whole blood RNA demonstrated alterations in multiple immune pathways
   and transcripts during acute hyperglycemia including decreased
   transcription of IL‐6, an important component of both innate and
   adaptive immune responses. Additional in vitro studies of human
   peripheral blood mononuclear cells (PBMCs) exposed to high glucose
   confirmed decreased IL‐6 expression, most prominently in CD14^+CD16^+
   intermediate monocytes. Hyperglycemia also reduced IL‐17A expression
   suggesting further impairment of immune responses during acute
   hyperglycemia. These findings demonstrate multiple defective immune
   responses in acute hyperglycemia and suggest a novel role for
   intermediate monocytes as metabolically sensitive innate immune cells.

   Keywords: Hyperglycemia, IL‐6, IL‐17, metabolism, monocyte

Introduction

   There is increasing evidence that the immune system can influence
   metabolism [38]1, [39]2, [40]3, [41]4. However, the effects of
   alterations in host metabolism on immune function may also be
   clinically relevant. Acute hyperglycemia is commonly encountered in
   critically ill patients and is associated with increased mortality
   [42]5, [43]6, [44]7, [45]8, [46]9. Critically ill subjects,
   particularly those with sepsis or multi‐organ failure, can also present
   with dysfunctional immune responses termed “immune paralysis;” however,
   it is unclear if there is a connection between hyperglycemia and
   impaired immune responses [47]10, [48]11. In addition, brief episodes
   of hyperglycemia are noted on oral glucose tolerance tests in subjects
   at risk for developing type 1 diabetes prior to the onset of clinical
   symptoms. These episodes of acute hyperglycemia increase rates of
   progression to type 1 diabetes but the mechanism for this additive risk
   remains unclear [49]12.

   Previous studies have described immune responses during acute
   hyperglycemia and have focused on expression of IL‐6. Using THP‐1 cell
   lines, investigators have shown that in vitro hyperglycemia increased
   IL‐6 expression [50]13. In humans, data suggest that plasma IL‐6 levels
   are increased in critically ill subjects admitted with hyperglycemia
   but these levels were measured randomly and not during controlled,
   provocative metabolic testing [51]14. However, when oral glucose
   tolerance testing was performed in humans with impaired glucose
   tolerance, there was a trend for decreasing plasma IL‐6 protein levels
   during mild hyperglycemia [52]15. Given the clinical importance of
   acute hyperglycemia and the conflicting data regarding its effects on
   immune function, we chose to address this problem by combining in vivo
   transcriptome analysis and immune phenotyping of primary human cells.

Methods

   A full description of the methods is provided in the supplementary
   section.

Subjects and study approval

   For the glucose‐octreotide protocol, 10 healthy control subjects were
   recruited from the control registry of the Translational Research
   Program at the Benaroya Research Institute at Virginia Mason Medical
   Center. All subjects were negative for islet autoantibodies and had no
   personal or family history of diabetes. The registry, as well the
   glucose‐octreotide protocol, had local institutional review board
   approval prior to enrollment. Sixty percent of the subjects were
   female; the average age was 28.3 years ± 5.9. All subjects provided
   written informed consent prior to study procedures.

   For all in vitro studies, samples from the control subject repository
   at the Benaroya Research Institute were used. Subjects had donated
   blood samples for research purposes. All were negative for diabetes
   autoantibodies. All subjects had provided informed consent prior to
   blood donation. Samples were used once and were not used in multiple in
   vitro experiments.

Statistics

   For the RNA‐seq analysis, we calculated the differential expression
   between pair‐wise FPKM values at T = 0 and T = 120. Linear models were
   fitted for each gene using the Bioconductor “limma” package in R
   [53]16, [54]17. Moderated t‐statistics, fold‐change, and the associated
   p‐values were calculated for each gene. The FDR (false discovery rate)
   was computed for each gene using the Benjamini–Hochberg method [55]18.
   Over‐represented pathways were identified using the program GO Elite
   [56]19. Only pathways with a P‐value < 0.01 after correcting for
   multiple comparisons were determined to be significantly
   over‐represented. All other analyses of ELISA,
   electrochemiluminescence, and flow cytometry data were done with Prism
   (GraphPad software) using ANOVA comparisons for multiple groups and
   nonparametric pair‐wise testing to determine statistical significance.

Results and Discussion

   We began studying the effects of acute hyperglycemia on immune function
   with an in vivo clinical experiment. Ten healthy subjects without
   diabetes received intravenous octreotide (to suppress endogenous
   insulin secretion and maximize short‐term hyperglycemia) followed by a
   single bolus of intravenous dextrose (Fig. [57]1A). Glucose levels
   peaked at 10 min post‐dextrose bolus (mean peak glucose
   253 mg/dL ± 27.3) and then declined (Fig. [58]1B). Whole blood RNA was
   collected prior to intravenous dextrose and at 120 min post‐dextrose
   bolus. All subjects tolerated the octreotide and glucose bolus well
   without significant adverse events. A separate protocol only using
   octreotide was not performed as octreotide‐only infusions have not been
   found to alter cytokine levels [59]20.

Figure 1.

   Figure 1
   [60]Open in a new tab

   Suppression of immune gene expression during in vivo acute
   hyperglycemia in human subjects. (a) Clinical protocol for 10 healthy
   control subjects. Participants were admitted to the clinical research
   center and received a bolus and continuous infusion of octreotide for
   120 min. Dextrose bolus occurred at t = 0 min and collection of whole
   blood RNA occurred at t = 0 and t = 120 min. (b) Glucose excursion over
   2 h in all 10 subjects with standard deviations. (c) Pathways
   significantly altered (P‐value < 0.01 via Benjamin–Hochberg method)
   between the whole blood RNA‐seq profiles at t = 0 and t = 120 min. (d)
   Proportion of transcripts that increased and decreased in the immune
   system process pathway (GO: 0002376). (E) Heatmap showing fold changes
   (in log 2) in the immune system process pathway (GO: 0002376). Arrow
   indicates IL‐6 expression in RNA‐seq analysis. (f) Transcripts from (e)
   with the linear fold change > 1. (g) Serum IL‐6 levels did not change
   during the 2‐h protocol.

   An RNA‐seq analysis was performed on the whole blood RNA comparing the
   baseline and 120‐min RNA samples. A pathway enrichment analysis
   demonstrated 24 statistically significantly altered pathways during
   acute hyperglycemia of which 4 were immune based pathways (Fig.
   [61]1C). We examined the immune system process pathway (GO: 0002376,
   P‐value = 0.002 after correcting for multiple hypothesis testing) as it
   contained the largest number of differentially expressed immune
   transcripts in our analysis. We observed that the majority of
   transcripts, including IL‐6, decreased with acute hyperglycemia (Fig.
   [62]1D and 1E). We focused on validating the IL‐6 signature given the
   magnitude of its observed fold change in relation to other transcripts,
   its importance in innate and adaptive immunity, and its association
   with hyperglycemia in previous studies (Fig. [63]1F). We analyzed serum
   from patients in the hyperglycemia protocol, collected at the same time
   as the RNA, and observed no changes in IL‐6 protein levels with the
   acute hyperglycemia (Fig. [64]1G). Given the short duration of the
   clinical intervention, we decided to model acute hyperglycemia in vitro
   to allow us to observe changes in IL‐6 protein levels with longer
   periods of hyperglycemia.

   We next studied another cohort of 10 healthy control individuals (see
   methods) not studied in our hyperglycemia protocol. Frozen human
   peripheral mononuclear cells (PBMC) samples were equally divided into a
   normal glucose (100 mg/dL) environment and a high glucose (450 mg/dL)
   environment for 5 h. Cell numbers were equivalent in both cultures.
   Cell culture supernatants were harvested and analyzed for IL‐6 by
   ELISA. IL‐6 levels significantly decreased in high‐glucose environments
   (Fig. [65]2A). Of the 10 subject samples analyzed under these
   conditions, 9 showed reduced IL‐6 expression with hyperglycemia.
   Exposing PBMCs to osmotically equivalent amounts of mannitol did not
   affect IL‐6 expression (Supplementary Fig. S1).

Figure 2.

   Figure 2
   [66]Open in a new tab

   Acute hyperglycemia decreases IL‐6 expression in intermediate
   monocytes. (a) IL‐6 levels in PBMC culture supernatants after 5 h from
   10 individuals analyzed with a Wilcoxon matched‐pairs sign rank test
   (b) Representative flow plots showing decreased intracellular IL‐6 in
   HLA‐DR^+CD14^+ monocytes with acute hyperglycemia. Whole PBMCs were
   gated for HLA‐DR^+CD14^+. Double positive cells were selected (circles)
   based on fluorescence minus one (FMO) control for both HLA‐DR and CD14.
   The FMO control for CD14 expression is shown in gray. Normal
   glucose = purple, high glucose = orange, fluorescence minus one (FMO)
   IL‐6 control = grey. (c) Representative flow plots showing monocyte
   subsets based on CD14 and CD16 staining. Gates drawn based on FMO
   controls (not shown). Normal glucose = purple and high
   glucose = orange. Histograms of intracellular IL‐6 staining in monocyte
   subsets from sample shown in C. Significant IL‐6 is only seen in
   intermediate monocytes and decreases with acute hyperglycemia. (d)
   Whole PBMC culture supernatants of five individuals exposed to normal
   glucose media and normal glucose media with a p38 inhibitor, SB203580,
   showed reduced IL‐6 expression (testing with a Wilcoxon rank sign
   test). (e) Whole PBMCs (n = 5) exposed to normal and high glucose show
   reduced mean p38 protein expression (testing with a Wilcoxon rank sign
   test, standard deviation bars shown). (f) Whole PBMC (n = 5) exposed to
   high glucose showed no significant change in mean phospho‐p38 levels
   (testing with a Wilcoxon rank sign test, standard deviation bars
   shown).

   We next analyzed the cell types secreting IL‐6 by flow cytometry after
   the 5‐h culture described above. When exposed to high glucose,
   monocytes (CD14^+HLA‐DR^+) demonstrated decreased IL‐6 expression,
   correlating with the ELISA data (Fig. [67]2B). Purified T‐cells did not
   show significant IL‐6 expression by ELISA or flow cytometry in our 5‐h
   culture experiment (Supplementary Fig. S2).

   As only a subset of CD14^+ cells seemed to respond to high glucose, we
   divided the monocyte population into subsets based on CD14 and CD16
   staining to identify classical (CD14^+CD16^−), non‐classical
   (CD14^−CD16^++), and intermediate monocytes (CD14^+CD16^+) [68]21,
   [69]22. Whole PBMCs were gated for CD14 and CD16 with fluorescence
   minus one (FMO) controls. We observed that only intermediate monocytes
   (CD14^+CD16^+) showed significant intracellular IL‐6 that decreased
   when cells were cultured in high glucose compared to normal glucose
   (Fig. [70]2C).

   To investigate the mechanism of hyperglycemic‐induced IL‐6 suppression,
   we studied the p38 mitogen‐activated protein kinase (p38MAPK) given its
   importance in regulating IL‐6 expression [71]23, [72]24, [73]25. A
   previous study had shown that hyperglycemic environments impair p38MAPK
   activity in keratinocytes [74]26. A modest reduction in the mRNA of the
   p38 gene, MAPK11, was seen in the RNA‐seq data from the in vivo
   clinical study (data not shown). No other isoform of p38 was detectable
   in the RNA‐seq analysis. Blockade of p38 activity with the inhibitor
   SB203580 decreased in vitro IL‐6 secretion from whole PBMCs after 5 h
   confirming the importance of p38 in IL‐6 expression (Fig. [75]2D).
   SB203580 only inhibits the kinase activity of p38MAPK but does not
   prevent its phosphorylation [76]27. Using an electrochemiluminescence
   assay to quantify total and phosphorylated p38 protein, we observed
   that hyperglycemia did not alter the phosphorylation status of p38MAPK
   but did reduce total amounts of p38 (Fig. [77]2E and 2F). These data
   suggested that hyperglycemia reduces p38 expression, but not
   phosphorylation, leading to reduced IL‐6 expression.

   Finally, we sought to determine if hyperglycemia‐mediated IL‐6
   suppression could alter downstream adaptive immunity given the
   importance of IL‐6 in differentiating naïve T‐cells into TH17 cells
   [78]28. We hypothesized that a high‐glucose environment would reduce
   IL‐17A expression in an IL‐6‐dependent fashion. PBMC samples from
   another eight control individuals were equally divided into normal and
   high‐glucose wells for 5 h followed by non‐specific stimulation with
   PMA/ionomycin for an additional 12 h. Exposure of primary PBMCs to a
   high‐glucose environment reduced IL‐17A expression by ELISA compared to
   a normal glucose environment (Fig. [79]3A and 3B).

Figure 3.

   Figure 3
   [80]Open in a new tab

   Acute hyperglycemia decreases IL‐17A expression. (a) IL‐17A measured by
   ELISA from cell culture supernatants in PBMC samples (n = 8) with in
   vitro acute hyperglycemia. (b) Number of samples that decreased IL‐17A
   expression with hyperglycemia.

Discussion

   Together, these data characterize the immune responses to short‐term
   acute hyperglycemia in humans. We show that short‐term hyperglycemia
   decreases IL‐6 expression in CD14^++CD16^+ intermediate monocytes and
   IL‐17A expression. These findings may have clinical relevance to the
   “immune paralysis” associated with critical illness and the development
   of type 1 diabetes.

   These findings also add to our growing knowledge that intermediate
   monocytes have a distinct metabolism from other monocyte subsets and
   may serve as circulating immuno‐glucose sensors [81]29. Future studies
   will evaluate inherent metabolic differences between human monocyte
   subsets and mechanisms of hyperglycemic suppression of p38 expression.
   Also, while our in vitro data explored the inhibition of IL‐6 during of
   acute hyperglycemia, our in vivo RNA‐seq data from the
   glucose‐octreotide protocol suggest that multiple other immune
   transcripts increase with hyperglycemia including IL‐12A, which is
   important in TH1 development. It is probable that the effects of
   hyperglycemia on adaptive immune cells are complex and will require
   additional studies to fully elucidate them.

   Our observation of reduced IL‐6 expression with hyperglycemia differs
   from previous publications. Methodological differences such as samples
   types and culture conditions used most likely account for these
   differences. We believe, however, that our in vivo RNA‐seq data
   combined with our in vitro data from primary human cells are valid
   models of immune responses during acute hyperglycemia. It is not clear
   if these data are readily applicable to chronic forms of hyperglycemia
   such as types 1 and 2 diabetes but an ongoing clinical trial with a
   monoclonal antibody against IL‐6 in type 1 diabetes ([82]NCT02293837)
   may be informative.

   Despite these limitations, we believe these data provide valuable
   mechanistic data relevant to immune function. It adds to our knowledge
   of human monocyte subsets and improves our understanding of the
   interaction between glucose metabolism and immunity.

Conflict of Interest

   The authors report no conflicts of interest relevant to this
   publication.

Supporting information

   Additional supporting information may be found in the online version of
   this article at the publisher's web‐site.

   Figure S1. Increased osmolarity with mannitol does not affect IL‐6
   secretion from whole PBMCs after 5‐h culture (n = 3).

   Figure S2. Monocytes and T‐cells from two individuals were isolated
   from whole PBMC samples using magnetic bead separation.
   [83]Click here for additional data file.^ (332.2KB, docx)

Acknowledgments