ORIGINAL RESEARCH

Possible Long-Term Efficacy of Sitagliptin,a Dipeptidyl Peptidase-4 Inhibitor, for SlowlyProgressive Type 1 Diabetes (SPIDDM) in the Stageof Non-Insulin-Dependency: An Open-LabelRandomized Controlled Pilot Trial (SPAN-S)

Takuya Awata . Akira Shimada . Taro Maruyama . Yoichi Oikawa . Nobuyuki Yasukawa . Susumu Kurihara .

Yumi Miyashita . Masako Hatano . Yuichi Ikegami . Masafumi Matsuda . Masataka Niwa . Youichiro Kazama .

Shoichiro Tanaka . Tetsuro Kobayashi

Received: July 24, 2017 / Published online: September 19, 2017� The Author(s) 2017. This article is an open access publication

ABSTRACT

Introduction: We tested the hypothesis thatdipeptidyl peptidase-4 (DPP-4) inhibitors areeffective in preserving the b-cell function forlong-term periods in patients with slowly pro-gressive type 1 diabetes (SPIDDM) or latentautoimmune diabetes in adults (LADA).

Methods: In the present open-label, random-ized, controlled trial, 14 non-insulin-requiringdiabetic patients with glutamic acid decar-boxylase autoantibodies (GADAb) were ran-domly assigned to receive either sitagliptin (Sgroup) or pioglitazone (P group). As a historicalcontrol, the Tokyo Study, in which non-in-sulin-dependent patients with SPIDDM wereassigned to receive treatment by either insulinor sulfonylurea (SU), was used.Results: On average, the

PC-peptide values

during the oral glucose tolerance test throughthe follow-up periods showed a nonsignificantincrease in the S group (n = 6, n = 5 at48 months) compared to the P group (n = 5,n = 2 at 48 months). In comparison to the datain the Tokyo Study, treatment by sitagliptinsignificantly influenced the longitudinal

Taro Maruyama: deceased before study completion.

Enhanced content To view enhanced content for thisarticle go to http://www.medengine.com/Redeem/18FBF0607C87BD83.

Electronic supplementary material The onlineversion of this article (doi:10.1007/s13300-017-0299-7)contains supplementary material, which is available toauthorized users.

T. Awata (&) � N. YasukawaDepartment of Diabetes, Endocrinology andMetabolism, International University of Health andWelfare Hospital, Iguchi, Nasushiobara-shi, Tochigi,Japane-mail: awatatakuya@gmail.com

A. Shimada � Y. Oikawa � S. Kurihara � M. Hatano �Y. IkegamiDepartment of Endocrinology and Diabetes, Facultyof Medicine, Saitama Medical University, Saitama,Japan

T. MaruyamaDepartment of Internal Medicine, Saitama SocialInsurance Hospital, Saitama, Japan

Y. MiyashitaDivision of RI Laboratory, Biomedical ResearchCenter, Saitama Medical University, Saitama, Japan

M. MatsudaDepartment of Endocrinology and Diabetes,Saitama Medical Center, Saitama MedicalUniversity, Saitama, Japan

M. NiwaIidabashi Medical Clinic, Tokyo, Japan

Y. KazamaKazama Medical Clinic, Yamanashi, Japan

Diabetes Ther (2017) 8:1123–1134

DOI 10.1007/s13300-017-0299-7

changes in theP

C-peptide values with a moreincreased direction than insulin or SU, espe-cially in patients with 48 months of follow-up(p = 0.014 and p = 0.007, respectively).Although the titers of GADAb were not signifi-cantly different between the S and P groupsduring the study, the change ratio of theGADAb titers from baseline was significantlyinversely correlated with the change ratio of theP

C-peptide values from baseline in the S group(p = 0.003); in particular, when the GADAbtiters decreased from baseline, the

PC-peptide

values frequently increased.Conclusion: The present pilot trial suggeststhat treatment of SPIDDM/LADA by sitagliptin,a DPP-4 inhibitor, may be more effective inpreserving the b-cell function than insulintreatment for at least 4 years, possibly throughthe immune modulatory effects of DPP-4inhibitors.Clinical trial registration: Japanese ClinicalTrials Registry UMIN000003693.

Keywords: Dipeptidyl peptidase-4 (DPP-4)inhibitor; Intervention; Latent autoimmunediabetes in adults (LADA); Prevention;Sitagliptin; Slowly progressive type 1 diabetes(SPIDDM); Type 1 diabetes

INTRODUCTION

Type 1 diabetes (T1D) is caused by the autoim-mune destruction of pancreatic b-cells, but ithas been revealed that human T1D is substan-tially heterogeneous in its pathogenesis andnatural history [1]. Slowly progressiveinsulin-dependent (type 1) diabetes mellitus(SPIDDM) [2, 3] or latent autoimmune diabetesin adults (LADA) [4–6] is characterized by the

clinical phenotype of type 2 diabetes (T2D),such as adult onset, non-insulin independenceat diagnosis, obesity and features of metabolicdisorders, but has glutamic acid decarboxylaseautoantibodies (GADAb) and/or islet cell anti-bodies (ICA); other antibodies such as insuli-noma-associated antigen-2 autoantibodies(IA-2Ab) and insulin autoantibodies (IAA) arerather infrequent. Patients with SPIDDM/LADAfrequently progress to an insulin-dependentstate but with a slower rate of b-cell loss than inacute-onset T1D, probably because of the mildautoimmune attack. The prevalence ofSPIDDM/LADA has been reported to range from3% to 12% [7–10] of patients initially diagnosedwith T2D in Western and other countries.

The generally unsatisfactory results of recentprevention/intervention trials for T1D havehighlighted its heterogeneity [11, 12], and it hastherefore been proposed that SPIDDM/LADA isan attractive target for immune interventionapproaches [13, 14], given its milder autoim-mune process and the longer preservation of theb-cell function than in acute-onset T1D. Wepreviously conducted the Tokyo Study in Japa-nese patients with SPIDDM [15], which showedthat, compared with sulfonylurea (SU), treat-ment with insulin slowed progression to aninsulin-dependent state. The effect of earlyinsulin therapy on SPIDDM/LADA has beenconfirmed [16].

Subsequent to the Tokyo study, we focusedon dipeptidyl peptidase-4 (DPP-4) inhibitors asa promising treatment option for SPIDDM,since several studies have reported the benefi-cial effects of DPP-4 inhibitors in the NODmouse, a mouse model of T1D [17–19]. Weconducted the ‘‘study for the prevention ofautoimmune non-insulin-dependent diabetesmellitus with sitagliptin’’ (SPAN-S) to evaluatethe effects of sitagliptin, the first DPP-4 inhi-bitor released in Japan, on the glycemic control,preservation of b-cell function and longitudinalchanges in islet cell antibodies. Although sev-eral studies in patients with SPIDDM/LADAhave reported favorable results with DPP-4inhibitors since the launch of this study[20–22], we continued our study for a longerperiod to assess the long-term efficacy of sita-gliptin for SPIDDM/LADA.

S. TanakaAi Home Clinic Toshima, Tokyo, Japan

T. KobayashiDivision of Immunology and Molecular Medicine,Okinaka Memorial Institute for Medical Research,Tokyo, Japan

1124 Diabetes Ther (2017) 8:1123–1134

METHODS

Patients

This study was conducted in accordance withthe International Conference on Harmoniza-tion Guidelines for Good Clinical Practice andthe Declaration of Helsinki. The Saitama Medi-cal University Ethics Committee approved thisstudy, and the protocol was approved by theinstitutional review boards of the collaboratinghospitals or the directors of the clinics. All ofthe procedures followed were in accordancewith the ethical standards of the responsiblecommittee on human experimentation (insti-tutional and national) and with the Helsinki

Declaration of 1964, as revised in 2013. Writteninformed consent was obtained from allpatients for being included in the study.

This was an open-label, prospective, ran-domized, controlled trial (Fig. 1). The protocolof this trial and supporting CONSORT checklistare available as Protocol S1 (original language),S2 (English translation) and Checklist S1. Par-ticipants were recruited from 10 June 2010, andfollow-up data were collected until 23 February2017. Fourteen SPIDDM patients were enrolled,and 11 of them were followed for at least12 months (Fig. 1b). The inclusion criteria wereas follows: (1) diabetes diagnosis and HbA1clevel between 6.9% and 8.4%, (2) GADAb-posi-tive, (3) fasting serum C-peptide level C1.0 ng/ml, (4) age 20–79 years of age, (5) no treatmentwith oral glucose-lowering drugs except met-formin within 2 months before enrollment and(6) no history of insulin or GLP-1 receptorantagonist treatment. The exclusion criteriawere: (1) women who were pregnant or plan-ning on becoming pregnant, (2) renal dysfunc-tion with serum creatinine C1.5 mg/dl for malesand C1.3 mg/dl for females, or creatinineclearance B50 ml/min, and (3) patients whowere deemed unsuitable for participation bytheir attending physician because of associateddisease, complications or any other reasons. Allsubjects included in the study were Japanese.

The eligible patients, who were recruitedfrom the collaborating hospitals or clinics, wereall randomly assigned using a centralizedinternet web system to receive either sitagliptinor pioglitazone based on minimization adjust-ing for age, sex, HbA1c level, GADAb titer andmetformin use before entry. Because the inferi-ority of SU compared to insulin treatment hadbeen revealed in the Tokyo study, pioglitazone,a thiazolidinedione (TZD) approved in Japan,was selected as the comparator. The efficacy ofTZD on SPIDDM/LADA has been evaluated inseveral studies with conflicting results [23, 24].In addition, from an ethical point of view,patients under treatment with metformin wereallowed to be enrolled because a possiblefavorable effect of metformin on SPIDDM/LADA was suggested in our previous study [23].

Fig. 1 Trial design and participant flow. a Schematicillustration of the trial design. b Participant flow. Data arepresented as n. a-GI alpha-glucosidase inhibitor, Lablaboratory test, M month, OGTT 75-g oral glucosetolerance test

Diabetes Ther (2017) 8:1123–1134 1125

Study Design

Patients were scheduled for visits at 2-monthintervals for at least 12 months of follow-up.The dosage for the sitagliptin group was startedat 50 mg once per day and that for the piogli-tazone group was 15 mg once per day; if met-formin was used at enrollment, the dosage wasnot changed. The target for glycemic controlwas set at an HbA1c level \7.0%. If glycemiccontrol was not achieved during each visit, theregimens were intensified by either (1) increas-ing the dosage of sitagliptin or pioglitazone to100 or 30 mg, respectively, and/or (2) addingmetformin at 250–2250 mg or increasing thedosage up to 2250 mg and/or adding analpha-glucosidase inhibitor (acarbose, vogliboseor miglitol). Notably, SU agents were neverused. The patients in both groups were switchedto an insulin injection regimen when theHbA1c levels became C9.4% despite the aboveprotocol of medication.

Follow-up Assessment and Endpoint

The patients’ body weight, blood pressure,plasma glucose (PG) and HbA1c levels weremeasured every 2 months. All of the patientsreceived an annual 75-g oral glucose tolerancetest (OGTT) without receiving their morningdose of medication, and their islet antibodies(GADAb, IA-2Ab and IAA) were determinedannually. The primary endpoints were theHbA1c values, the proportion of patients whoseHbA1c values increased to C9.4%, and the per-iod from enrollment in which their HbA1cvalues increased to C9.4%. The secondary end-points included the sum of the serum C-peptidelevels at 0, 30, 60, 90 and 120 min during theOGTT (

PC-peptide), body mass index (BMI)

and islet autoantibodies.

Autoantibodies

GADAb, IA-2Ab and IAA levels were determinedin this study via radioimmunoassay (RSR Ltd.,Cardiff, UK). With regard to the thyroid autoan-tibodies (TAb), thyroid peroxidase autoantibod-ies (TPOAb) and thyroglobulin autoantibodies

(TGAb) were determined by an electrochemilu-minescence immunoassay (Roche DiagnosticsGmbH, Mannheim, Germany). Regarding thepresent GADAb assay, 10 U/ml corresponded to180 U/ml (WHO units), the cutoff value thatpredicted further progression of b-cell dysfunc-tion in the Tokyo study [25].

Safety Assessments

Any adverse events associated with the medi-cations used in this study were evaluated andtreated properly by physicians. Hypoglycemiawas defined as a blood glucose level\70 mg/dl(3.9 mmol/l), irrespective of hypoglycemicsymptoms.

Sample Size

Based on our previous intervention trial (TokyoStudy) [15], in which 60 SPIDDM patients (30each in the insulin and SU groups) were fol-lowed and the primary endpoint was the time atwhich the patient reached an insulin-depen-dent state (i.e.,

PC-peptide\4 ng/ml), and on

the more preserved b-cell function in theinclusion criteria of this trial, which wouldshow deleterious progression of the functionmore clearly, we anticipated that a total of 40patients (20 each in the S and P groups) wouldbe a sufficient population in our initial plan forthe SPAN-S. Although the primary endpoint inthe present study was defined by HbA1c values,worsening of HbA1c values was expected tocorrespond well to worsening of

PC-peptide

levels.

Statistical Analysis

Continuous data at baseline (n = 11) were firstchecked for their normality by the D’Agostinoand Pearson test, revealing that the age, GADAbtiter and change ratio of the GADAb titer frombaseline significantly deviated from a normaldistribution. Thus, the Mann-Whitney U testand Wilcoxon signed rank test were applied tothese data for comparisons between groups andcomparison from the baseline within the samegroup, respectively. An unpaired t test and

1126 Diabetes Ther (2017) 8:1123–1134

paired t-test were used for duration, BMI,HbA1c,

PPG,

PC-peptide and change ratio of

PC-peptide from baseline, respectively. Cate-

gorical variables were compared using Fisher’sexact test. To evaluate the difference in thelongitudinal

PC-peptide data or their change

ratios between the treatment groups, the inter-action between time and treatment assignmentat entry was examined using a repeated mea-sures analysis of variance (ANOVA). The corre-lation between the change ratio of the GADAbtiters and that of the

PC-peptide values was

assessed using a nonparametric Spearman’scorrelation coefficient. Statistical significancewas defined as p\0.05. The GraphPad Prismsoftware program ver. 6 (GraphPad Software,Inc., La Jolla, CA, USA) was used for these tests.

RESULTS

Patients and Follow-Up

As shown in Fig. 1b, 14 patients were allocated:7 were assigned sitagliptin (S group) and 7pioglitazone (P group). Among them, sixpatients in the S group and five patients in the Pgroup were followed for more than 1 year andanalyzed.

The baseline characteristics of these 11patients are shown in Table 1. There were nosignificant differences between the two groupsin age, sex, duration, metformin use beforeentry, BMI, HbA1c levels, GADAb titers or pos-itivity of other autoantibodies (IA2Ab, IAA andTAb).

Six and five patients in the S group werefollowed for 24 and 48 months, respectively,while five, four and two patients in the P groupwere followed for 12, 24 and 48 months,respectively (Table S1). The intensification ofthe regimens after entry is shown in Table S2.Briefly, the sitagliptin dosage was not increasedin any patients, and the pioglitazone dosagewas increased in three patients (cases P2, P4 andP5). Metformin was added to two patients (caseS5 and P5) among those who had not beentreated by metformin at entry (cases S5, P1 andP5), and an alpha-glucosidase inhibitor (vogli-bose) was added to one patient (case S4). The

metformin dosage was changed in four patients(cases S2, S4, P3 and P4).

Longitudinal Changes in the BMI, BloodPressure and HbA1c

The BMI (or body weight) of the S group did notsignificantly differ from that at baseline; theBMI (or body weight) of the P group showed an

Table 1 Baseline characteristics of study patients withSPIDDM

Sitagliptingroup(n5 6)

Pioglitazonegroup (n5 5)

p value

Age (years) 56 ± 16 57 ± 8 0.93

Sex

(male:female)

2:4 2:3 0.85

Duration of

diabetes

(years)

4.3 ± 2.9 3.8 ± 4.1 0.81

Metformin use

before entry,

n (%)

5 (83) 3 (60) 0.48

BMI (kg/m2) 23.2 ± 6.2 23.9 ± 3.9 0.73

HbA1c (%) 7.5 ± 0.5 7.3 ± 0.4 0.59

GADAb (U/ml) 2.65

(1.6–250)

8.4

(1.8–14,000)

0.93

IA-2Ab positive,

n (%)

2 (33) 1 (20) 0.70

IAA positive,

n (%)

0 (0) 0 (0) na

TAb (TPOAb

and/or

TGAb)

positive, n (%)

2 (33) 3 (60) 0.46

Data are n (%), mean ± SD or median (range)BMI body mass index, GADAb glutamic acid decarboxy-lase autoantibodies, IA-2Ab insulinoma-associated anti-gen-2 autoantibodies, IAA insulin autoantibodies, TAbthyroid autoantibodies, TPOAb thyroid peroxidaseautoantibodies, TGAb thyroglobulin autoantibodies, nanot applicable

Diabetes Ther (2017) 8:1123–1134 1127

increasing trend from the value at enrollment,but the change was not significant (Table S1).The blood pressure of both groups did not sig-nificantly differ from that at baseline (data notshown). The HbA1c levels were decreased inboth groups, and significant reductions frombaseline were observed at several points in the Sgroup: at 12 months in the patients with both24 and 48 months of follow-up (n = 6 and n = 5,respectively) and at 36 months in those with48 months of follow-up (Table S1). The worstHbA1c levels in the S and P groups were 8.2%and 8.2%, respectively (Table S2). The values ofBMI (or body weight), blood pressure andHbA1c did not significantly differ between thetwo groups during the study.

Longitudinal Changes in the C-peptideResponse to the OGTT

On average, theP

C-peptide values throughoutthe follow-up periods were greater in the S groupthan in the P group, although a repeated-measureANOVA showed that the influence of the treat-ment assignment on the longitudinal changes inthe

PC-peptide values was not significant in any

of the follow-up periods (Table S1). In the sixpatients in the S group who were observed for atleast 24 months, the

PC-peptide values increased

at 12 and 24 months (six patients) and decreasedat 36 and 48months (5 patients) but were stillhigher than those at baseline; furthermore, theincreases at 12 months from baseline were sig-nificant (p = 0.049) in the patients with 12 and24 months of follow-up (Table S1). In the fivepatients in the P group, of whom only two werefollowed more than 36 months, the

PC-peptide

values were mostly unchanged from baseline(Table S1 and Table S2).

Comparison of the b-Cell Functionwith Historical Control

We compared our present data with those in theTokyo Study as a historical control. In the TokyoStudy [15], the GADAb-positive non-insulin-re-quiring diabetic patients (SPIDDM), the sametarget subjects as in the present SPAN-S trial, wererandomly assigned to receive either insulin

(I group) or SU (SU group) and were followed upfor a maximum of 60 months. The longitudinalchanges in the

PC-peptide values of the S and P

groups were compared with those of the I and SUgroups in the Tokyo Study. Of note, no signifi-cant differenceswere observed between the S or Pgroup in the present study and the I or SU groupin the Tokyo Study in the baseline characteristicsof age, sex, duration, BMI, HbA1c levels, GADAbtiters and

PC-peptide values. Although the I

group demonstrated better preservation ofP

C-peptide values than the SU group, as repor-ted previously, the preservation of

PC-peptide

values in the S group in the present studyappeared to be even better. As shown in Fig. 2a, arepeated-measures ANOVA revealed a significantinteraction between time and treatment assign-ment (sitagliptin or insulin) as well as betweentime and treatment assignment (sitagliptin orSU) in the patients with 48 months of follow-up(p = 0.030 and p = 0.00004, respectively), indi-cating that treatment by sitagliptin significantlyinfluenced the longitudinal changes of the C--peptide responses to the OGTT with a moreincreased direction than insulin and SU. Differ-ences between the S group and the I and SUgroupswere further assessed based on the changeratios of the

PC-peptide values from baseline

(Fig. 2b). The significant superiority of treatmentby sitagliptin to that by insulin or SU in thepreservation of the

PC-peptide values appeared

to be more evident in the patients with48 months of follow-up (p = 0.014 andp = 0.007, respectively), although the significantsuperiority of sitagliptin to insulin or SUwas alsoobserved in the patients with 24 and 36 monthsof follow-up (data not shown). As for the P group,no significant superiority of pioglitazone com-pared to insulin or SU was observed in arepeated-measures ANOVA (data not shown).

Individual LongitudinalP

C-peptideValues, GADAb Titers and IA2Ab Titersof the Patients in the S Group with 48Months of Follow-Up

We were able to observe the clinical course for48 months in 5 patients of the S group (casesS1, S2, S3, S4 and S5 in Table S2). Individual

1128 Diabetes Ther (2017) 8:1123–1134

longitudinal change ratios of theirP

C-peptidevalues, GADAb and IA-2Ab titers from baselineare shown in Fig. S1. Among these patients,four were heterozygous for the susceptible andneutral HLA haplotypes (S/N), and the otherone (case S1) was homozygous for the neutralhaplotype (N/N) [26]. Regarding TAb (TPOAbor TGAb), two patients (cases S3 and S5) werepositive for TPOAb. The

PC-peptide values of

all five patients were preserved during the48 months of follow-up. Except for cases S2and S5 whose

PC-peptide values remained

almost the same throughout the study (S2) orslightly decreased from 24 to 48 months (S5),the values of the other three patients showed atendency to increase from 12 to 24 months andthen decreased from 36 to 48 months (Fig. S1).In comparison to cases S2, S3 and S5, who allhad GADAb [10 U/ml and two of whom (S2and S3) were IA2Ab-positive, the

PC-peptide

values of cases S1 and S4, who both showedGADAb \10 U/ml and were IA2Ab-negative,were increased to a greater degree from base-line, even at 48 months from entry. The rela-tive increase from baseline was highest in caseS1, who carried the HLA N/N genotype as wellas both GADAb\10 U/ml and IA2Ab-negativestatus.

GADAb Titers

The titers of GADAb were not significantly differ-entbetween theSandPgroups at any timepoint inthe present study (Table S1). However, as shown inFig. 3, the change ratios of the GADAb titers frombaseline were significantly inversely correlatedwith the change ratios of the

PC-peptide values

from baseline in the S group by a nonparametricanalysis (Spearman’s rank correlation coeffi-cient=-0.600,p= 0.003). Inparticular,when thechange ratios of the GADAb titers from baselinewere negative, those of the

PC-peptide values

were mostly positive, i.e., when the GADAb titersdecreased from baseline, the

PC-peptide values

frequently increased. In contrast,when the changeratios of the GADAb titers from baseline were pos-itive, those of the

PC-peptide values tended to

show values around zero (Fig. 3). In the P group, anonsignificant inverse correlation was observed(Spearman’s rankcorrelationcoefficient=-0.478,p= 0.101; Fig. S2).

Hypoglycemia and Adverse Events

No drug-related adverse events, includinghypoglycemia, were observed.

Fig. 2 Longitudinal changes in the C-peptide response tothe OGTT for 48 months in the patients with sitagliptintreatment and in the Tokyo study. Patients with48 months of follow-up are shown. The data are expressedas the mean ± SEM. In both the

PC-peptide values

(a) and change ratios from baseline (b), a

repeated-measures ANOVA revealed a significant interac-tion between time and treatment assignment (sitagliptin orinsulin; p = 0.030 and p = 0.014, respectively) as well asbetween time and treatment assignment (sitagliptin or SU;p = 0.00004 and p = 0.007, respectively) in the longitu-dinal changes

Diabetes Ther (2017) 8:1123–1134 1129

DISCUSSION

In the present report, although the initial out-comes based on comparisons between sitaglip-tin and comparator (pioglitazone) were notevaluated enough—mostly because of therecruited sample size with low power—we pre-sented our study as a pilot trial focusing onsitagliptin-treated patients with long-term(4 years) follow-up and further compared theeffects of sitagliptin treatment on preserving theb-cell function with those of insulin or SUtreatment in the patients enrolled in the TokyoStudy.

The present findings suggested that treat-ment by sitagliptin for SPIDDM/LADA is effec-tive with respect to the preservation of the b-cellfunction as well as glycemic control. On

average, theP

C-peptide values of the sitaglip-tin group were consistently higher than base-line for 48 months, with an apparent peak at24 months. Compared to the data of the TokyoStudy as a historical control, in which SPIDDMpatients were randomly assigned to treatmentby either insulin or SU, the preservation of theb-cell function following treatment by sitaglip-tin was significantly superior to that achievedwith insulin and SU therapy, with the differ-ences being evident from 24 months (Fig. 2b).We should be cautious of bias associated withusing a historical control. Nevertheless, asdescribed in the results section, no significantdifferences in the baseline characteristics wereobserved between the present study and theTokyo Study. Furthermore, in the Tokyo Study,several patients showed a low b-cell function atbaseline (defined as

PC-peptide \10 ng/ml),

possibly because the enrolled patients in theTokyo Study were recruited among those trea-ted by SU, while none of the patients in thesitagliptin group of the present study showedsuch a low function. However, even whenpatients with a low b-cell function at baselinewere excluded, the longitudinal changes in theP

C-peptide values in patients with 48 monthsof follow-up were significantly higher in thesitagliptin group than in the insulin and SUgroups (p = 0.019 and p = 0.014 byrepeated-measures ANOVA, respectively; datanot shown).

The favorable effect of sitagliptin on theb-cell function may be because of the reducedglucotoxicity due to the improvement of gly-cemic control, as shown by the decline in theHbA1c values from baseline (Table S1). How-ever, DPP-4 inhibitors may also have roles in theimmune system. In the NOD mouse model ofT1D, sitagliptin improved the rate of the isletgraft survival partially by reducing the CD4 Tcell numbers [19]. In another study, twice-dailytreatment with NVP-DPP, a short-acting DPP-4inhibitor, was able to reverse new-onset dia-betes in NOD mice by increasing the number ofCD4?CD25?FoxP3? regulatory T cells [18].Recently, Alonso et al. reported that a DPP-4inhibitor, MK626, prevented T1D in NOD micein association with changes in the CD8? Teffector memory lymphocyte subset [27]. The

Fig. 3 The correlation between the change ratios of theGADAb titers and of the

PC-peptide values from

baseline in the S group. Spearman correlation calculationsrevealed a significant inverse correlation between thechange ratios of the GADAb titers and those of theP

C-peptide values from baseline in the patients withsitagliptin treatment (Spearman’s rank correlation coeffi-cient = -0.600, p = 0.003)

1130 Diabetes Ther (2017) 8:1123–1134

superiority of sitagliptin to insulin in thepreservation of the b-cell function might comefrom the immune regulatory effects of DPP-4inhibitors on the autoimmune process.

Intriguingly, in the present study, we alsoobserved that the change ratio of GADAb titersfrom baseline was significantly inversely corre-lated with that of

PC-peptide in the S group

(Fig. 3). A higher GADAb titer in SPIDDM is amarker of both an activated T cell response tob-cell destruction and a high risk for progressionto insulin dependence [28, 29]. Therefore, theinverse correlation observed in the presentstudy, especially the obvious associationbetween the increase in the

PC-peptide values

from baseline and the decrease in the GADAbtiters (Fig. 3), may reflect changes in the activeautoimmune process of the patients influencedby the treatment with sitagliptin, which resul-ted in the alteration of the b-cell function in aninverse direction. As demonstrated in the indi-vidual longitudinal change ratios of theirP

C-peptide values from baseline (Fig. S1), thelong-term preservation of the b-cell function(P

C-peptide values) of patients in the S groupappeared to be associated with both GADAb\10 U/ml and IA2Ab-negative status (cases S1and S4) or the absence of the HLA class II sus-ceptibility haplotype (case S1). However, furthervalidation in a larger population of patients willbe required to determine the immunologic andgenetic factors influencing the efficacy of sita-gliptin on the b-cell function.

Recently, based on the potential beneficialeffects of DPP-4 inhibitors and the promisingresults in animal studies, DPP-4 inhibitors havebeen clinically tested in patients with T1D orSPIDDM/LADA. Favorable effects of sitagliptinwere suggested in several cases with T1D [30] orSPIDDM/LADA [31, 32]; very recently, 4-yearclinical remission of T1D in two young patientstreated with sitagliptin and vitamin D3 werereported [33]. However, the results of clinicaltrials for T1D patients have been conflicting andappear to be inconclusive [34–36]. In contrast,the trials for patients with SPIDDM/LADA, inwhom the effects of sitagliptin for 1 year [20],linagliptin for a maximum of 2 years [21] andsaxagliptin for 6 months [22] have been evalu-ated, all reported favorable results concerning

the preservation of the b-cell function, sug-gesting that DPP-4 inhibitors may be effective inSPIDDM/LADA patients as a class effect and thatSPIDDM/LADA might be a good target in pre-vention/intervention trials. The present find-ings lend support to the previous reports onSPIDDM/LADA and further suggest that thebeneficial immunologic effects of DPP-4 inhi-bitors may persist for a longer period innon-insulin-dependent patients with SPIDDM/LADA than was observed in previous studies.

The major limitation of the present study isits small sample size, which substantiallyreduced the power of the study. The low num-ber of enrolled patients may be due in part tothe inclusion criteria of certain HbA1c valuesand no history of insulin treatment, sinceJapanese patients with SPIDDM under subopti-mal glycemic control tend to be treated byinsulin based on the results of the Tokyo Study.It should be noted that due to the small samplesize, the present study lacks sufficient power todiscriminate between the effects of sitagliptinand pioglitazone on SPIDDM and also insuffi-ciently characterized the patients with SPIDDM/LADA suitable for treatment with DPP-4 inhi-bitors. However, the observed efficacy of sita-gliptin treatment on glycemic control and thepreservation of the b-cell function appeared tobe notable, particularly in the two patients whohad GADAb[10 U/ml and were IA2Ab-positiveand heterozygous for susceptible and neutralHLA DRB1-DQB1 haplotypes [26] (Fig. S1B andS1C).

Another limitation is that metformin use wasnot prohibited in the present study from anethical point of view. Metformin was used totreat five of the six patients in the S group atbaseline; during the study period, the dosagewas increased in two patients and was added inone patient (Table S2). We previously reportedthat metformin may suppress the disease courseof SPIDDM/LADA more effectively than piogli-tazone [23], and we subsequently experienced apatient with SPIDDM/LADA whose b-cell func-tion was maintained for 5 years with treatmentof metformin alone [37]. Metformin increasesthe GLP-1 level [38], which regulates lympho-cyte proliferation and the maintenance ofperipheral regulatory T cells [39]; given that

Diabetes Ther (2017) 8:1123–1134 1131

metformin was recently reported to improve gutmicrobiome flora in type 2 diabetes patients[40], metformin treatment for non-insulin-de-pendent SPIDDM might have exertedimmunomodulatory effects through gut floramodification. As such, metformin may beeffective in preserving the b-cell function inaddition to its glucose-lowering effects and thusthe usage of metformin may have influencedthe favorable results in the S group. However,possible favorable effects of metformin onSPIDDM/LADA need to be verified in extensiveclinical studies. Finally, although the use of analpha-glucosidase inhibitor was also allowed inthis trial, its influence would be minimal sinceonly one patient received the medication andlittle is known of its role in SPIDDM/LADA.

CONLUSION

The present study suggested that treatment ofSPIDDM/LADA by sitagliptin, possibly as a classeffect of DPP-4 inhibitors, may be more effectivein the preservation of the b-cell function thaninsulin treatment for at least 4 years, possiblythrough the immune modulatory effects ofDPP-4 inhibitors. We therefore believe that thisclass of drugs may be an attractive choice in thetreatment of SPIDDM/LADA, especially in thestage of non-insulin-dependency. Further lar-ge-scale and long-term studies are necessary toclarify the details of the beneficial effects ofDPP-4 inhibitors in SPIDDM/LADA patients andpossibly in acute-onset T1D patients with aresidual b-cell function.

ACKNOWLEDGEMENTS

This work was funded via a grant from theWaksman Foundation of Japan, Inc. The funderhad no role in any aspect of the study beyondfunding. All authors made significant contri-butions to the study design, analysis and inter-pretation of data, as well as preparing andreviewing the manuscript. All named authorsmeet the International Committee of MedicalJournal Editors (ICMJE) criteria for authorshipfor this manuscript, take responsibility for the

integrity of the work as a whole and have givenfinal approval for the version to be published.Previous presentation: some of the findingsfrom this study were presented at the 58thAnnual Meeting of The Japan Diabetes Society,Shimonoseki, Japan, 21–24 May 2015. Theauthors thank all of the members of the insti-tutions joining the SPAN-S trial.

Disclosures. Takuya Awata, Akira Shimada,Taro Maruyama, Yoichi Oikawa, NobuyukiYasukawa, Susumu Kurihara, Yumi Miyashita,Masako Hatano, Yuichi Ikegami, MasafumiMatsuda, Masataka Niwa, Youichiro Kazama,Shoichiro Tanaka and Tetsuro Kobayashi havenothing to disclose.

Compliance with Ethics Guidelines. Allprocedures followed were in accordance withthe ethical standards of the responsible com-mittee on human experimentation (institu-tional and national) and with the HelsinkiDeclaration of 1964, as revised in 2013. Writteninformed consent was obtained from allpatients for being included in the present study.

Data Availability. The data sets during thecurrent study are available from the corre-sponding author on reasonable request.

Open Access. This article is distributedunder the terms of the Creative CommonsAttribution-NonCommercial 4.0 InternationalLicense (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercialuse, distribution, and reproduction in anymedium, provided you give appropriate credit tothe original author(s) and the source, provide alink to the Creative Commons license, andindicate if changes were made.

REFERENCES

1. Atkinson MA, von Herrath M, Powers AC,Clare-Salzler M. Current concepts on the patho-genesis of type 1 diabetes–considerations forattempts to prevent and reverse the disease. Dia-betes Care. 2015;38(6):979–88.

1132 Diabetes Ther (2017) 8:1123–1134

2. Kobayashi T, Tamemoto K, Nakanishi K, et al.Immunogenetic and clinical characterization ofslowly progressive IDDM. Diabetes Care.1993;16(5):780–8.

3. Tanaka S, Ohmori M, Awata T, et al. Diagnotic cri-teria for slowly progressive insulin-dependent (type1) diabetes mellitus (SPIDDM) (2012): report by thecomitteeon slowly progressive insulin-dependent(type 1) diabetes mellitus of the Japan DiabetesSociety. Diabetol Int. 2015;6(1):1–7.

4. Tuomi T, Groop LC, Zimmet PZ, Rowley MJ,Knowles W, Mackay IR. Antibodies to glutamic aciddecarboxylase reveal latent autoimmune diabetesmellitus in adults with a non-insulin-dependentonset of disease. Diabetes. 1993;42(2):359–62.

5. Pozzilli P, Di Mario U. Autoimmune diabetes notrequiring insulin at diagnosis (latent autoimmunediabetes of the adult): definition, characterization,and potential prevention. Diabetes Care.2001;24(8):1460–7.

6. Stenstrom G, Gottsater A, Bakhtadze E, Berger B,Sundkvist G. Latent autoimmune diabetes in adults:definition, prevalence, beta-cell function, andtreatment. Diabetes. 2005;54(Suppl 2):S68–72.

7. Tanaka S, Awata T, Shimada A, et al. Clinical char-acteristics of slowly progressive insulin-dependent(type 1) diabetes mellitus (SPIDDM): 1st subcom-mittee report on SPIDDM, comittee on type 1 dia-betes, Japan Diabetes Society. J Japan Diab Soc (inJapanese). 2011;54(1):65–75.

8. Tuomi T, Carlsson A, Li H, et al. Clinical and geneticcharacteristics of type 2 diabetes with and withoutGAD antibodies. Diabetes. 1999;48(1):150–7.

9. Zhou Z, Xiang Y, Ji L, et al. Frequency, immuno-genetics, and clinical characteristics of latentautoimmune diabetes in China (LADA Chinastudy): a nationwide, multicenter, clinic-basedcross-sectional study. Diabetes. 2013;62(2):543–50.

10. Maddaloni E, Lessan N, Al Tikriti A, Buzzetti R,Pozzilli P, Barakat MT. Latent autoimmune diabetesin adults in the United Arab Emirates: clinical fea-tures and factors related to insulin-requirement.PLoS ONE. 2015;10(8):e0131837.

11. Pozzilli P. Type 1 diabetes mellitus in 2011:heterogeneity of T1DM raises questions for therapy.Nat Rev Endocrinol. 2012;8(2):78–80.

12. Atkinson MA, Eisenbarth GS, Michels AW. Type 1diabetes. Lancet. 2014;383(9911):69–82.

13. Kobayashi T, Nakanishi K, Murase T, Kosaka K.Small doses of subcutaneous insulin as a strategy forpreventing slowly progressive beta-cell failure in

islet cell antibody-positive patients with clinicalfeatures of NIDDM. Diabetes. 1996;45(5):622–6.

14. Pozzilli P, Maddaloni E, Buzzetti R. Combinationimmunotherapies for type 1 diabetes mellitus. NatRev Endocrinol. 2015;11(5):289–97.

15. Maruyama T, Tanaka S, Shimada A, et al. Insulinintervention in slowly progressive insulin-depen-dent (type 1) diabetes mellitus. J Clin EndocrinolMetab. 2008;93(6):2115–21.

16. Brophy S, Davies H, Mannan S, Brunt H, WilliamsR. Interventions for latent autoimmune diabetes(LADA) in adults. Cochrane Database Syst Rev.2011;9:CD006165.

17. Suarez-Pinzon WL, Cembrowski GS, Rabinovitch A.Combination therapy with a dipeptidyl peptidase-4inhibitor and a proton pump inhibitor restoresnormoglycaemia in non-obese diabetic mice. Dia-betologia. 2009;52(8):1680–2.

18. Tian L, Gao J, Hao J, et al. Reversal of new-onsetdiabetes through modulating inflammation andstimulating beta-cell replication in nonobese dia-betic mice by a dipeptidyl peptidase IV inhibitor.Endocrinology. 2010;151(7):3049–60.

19. Kim SJ, Nian C, McIntosh CH. Sitagliptin (MK0431)inhibition of dipeptidyl peptidase IV decreasesnonobese diabetic mouse CD4 ? T-cell migrationthrough incretin-dependent and -independentpathways. Diabetes. 2010;59(7):1739–50.

20. Zhao Y, Yang L, Xiang Y, et al. Dipeptidyl peptidase4 inhibitor sitagliptin maintains beta-cell functionin patients with recent-onset latent autoimmunediabetes in adults: one year prospective study. J ClinEndocrinol Metab. 2014;99(5):E876–80.

21. Johansen OE, Boehm BO, Grill V, et al. C-peptidelevels in latent autoimmune diabetes in adults trea-ted with linagliptin versus glimepiride: exploratoryresults from a 2-year double-blind, randomized,controlled study. Diabetes Care. 2014;37(1):e11–2.

22. Buzzetti R, Pozzilli P, Frederich R, Iqbal N, Hirsh-berg B. Saxagliptin improves glycaemic control andC-peptide secretion in latent autoimmune diabetesin adults (LADA). Diabetes Metab Res Rev.2016;32(3):289–96.

23. Shimada A, Shigihara T, Okubo Y, Katsuki T,Yamada Y, Oikawa Y. Pioglitazone may acceleratedisease course of slowly progressive type 1 diabetes.Diabetes Metab Res Rev. 2011;27(8):951–3.

24. Zhou Z, Li X, Huang G, et al. Rosiglitazone com-bined with insulin preserves islet beta cell functionin adult-onset latent autoimmune diabetes (LADA).Diabetes Metab Res Rev. 2005;21(2):203–8.

Diabetes Ther (2017) 8:1123–1134 1133

25. Kobayashi T, Tanaka S, Shimada A, Maruyama T,et al. High titer of autoantibodies to GAD identifiesa specific phenotype of adult-onset autoimmunediabetes: response to Buzzetti et al. Diabetes Care.2007;30(11):e126.

26. Awata T, Kuzuya T, Matsuda A, Iwamoto Y, Kana-zawa Y. Genetic analysis of HLA class II alleles andsusceptibility to type 1 (insulin-dependent) diabetesmellitus in Japanese subjects. Diabetologia.1992;35(5):419–24.

27. Alonso N, Julian MT, Carrascal J, et al. Type 1 dia-betes prevention in NOD Mice by targeting DPPIV/CD26 is associated with changes in CD8(?)T effec-tor memory subset. PLoS ONE.2015;10(11):e0142186.

28. Kasuga A, Maruyama T, Nakamoto S, Ozawa Y,Suzuki Y, Saruta T. High-titer autoantibodiesagainst glutamic acid decarboxylase plus autoanti-bodies against insulin and IA-2 predicts insulinrequirement in adult diabetic patients. J Autoim-mun. 1999;12(2):131–5.

29. Shimada A, Maruyama T, Suzuki R, et al. Anti--GAD65 antibody titer may be important in assess-ing T-cell response in anti-GAD65 ? diabetes withresidual beta-cell function. Diabetes Care.1999;22(10):1759.

30. Lima-Martinez MM, Guerra-Alcala E, Contreras M,Nastasi J, Noble JA, Polychronakos C. One yearremission of type 1 diabetes mellitus in a patienttreated with sitagliptin. Endocrinol Diabetes Meta-bol Case Rep. 2014;2014:140072.

31. Kandasamy N, Lennox G, Annamalai AK, MaguireG, Adler AI. Sitagliptin in glutamic acid decar-boxylase antibody-positive diabetes mellitus.Endocr Pract. 2012;18(4):e65–8.

32. Rapti E, Karras S, Grammatiki M, et al. Combinedtreatment with sitagliptin and vitamin D in apatient with latent autoimmune diabetes in adults.Endocrinol Diabetes Metabol Case Rep.2016;2016:150136.

33. Pinheiro MM, Pinheiro FM, Torres MA. Four-yearclinical remission of type 1 diabetes mellitus in twopatients treated with sitagliptin and vitamin D3.Endocrinol Diabetes Metabol Case Rep.2016;2016:16–0099.

34. Ellis SL, Moser EG, Snell-Bergeon JK, Rodionova AS,Hazenfield RM, Garg SK. Effect of sitagliptin onglucose control in adult patients with Type 1 dia-betes: a pilot, double-blind, randomized, crossovertrial. Diabet Med. 2011;28(10):1176–81.

35. Garg SK, Moser EG, Bode BW, et al. Effect of sita-gliptin on post-prandial glucagon and GLP-1 levelsin patients with type 1 diabetes: investigator-initi-ated, double-blind, randomized, placebo-controlledtrial. Endocr Pract. 2013;19(1):19–28.

36. Griffin KJ, Thompson PA, Gottschalk M, Kyllo JH,Rabinovitch A. Combination therapy with sita-gliptin and lansoprazole in patients withrecent-onset type 1 diabetes (REPAIR-T1D): 12--month results of a multicentre, randomised,placebo-controlled, phase 2 trial. Lancet DiabetesEndocrinol. 2014;2(9):710–8.

37. Hirata T, Shimada A, Morimoto J, Maruyama T.Slowly progressive type 1 diabetes treated withmetformin for five years after onset. Intern Med.2013;52(23):2635–7.

38. Bahne E, Hansen M, Bronden A, Sonne DP, VilsbollT, Knop FK. Involvement of glucagon-like peptide-1in the glucose-lowering effect of metformin. Dia-betes Obes Metab. 2016;18(10):955–61.

39. Hadjiyanni I, Siminovitch KA, Danska JS, DruckerDJ. Glucagon-like peptide-1 receptor signallingselectively regulates murine lymphocyte prolifera-tion and maintenance of peripheral regulatory Tcells. Diabetologia. 2010;53(4):730–40.

40. Forslund K, Hildebrand F, Nielsen T, et al. Disen-tangling type 2 diabetes and metformin treatmentsignatures in the human gut microbiota. Nature.2015;528(7581):262–6.

1134 Diabetes Ther (2017) 8:1123–1134