Intralymphatic Immunotherapy: A 3-year Randomized, Double-blind Study in 72 Patients With Allergic Rhinitis Due to Birch and Grass

Background: There is need for a fast, ecient, and safe way to induce tolerance in patients with severe allergic rhinitis. Methods: Patients with severe birch and timothy allergy were randomized and received three doses of 0.1 ml of birch and 5-grass allergen extracts (10,000 SQ units/ml, ALK-Abelló), or birch and placebo or 5-grass and placebo by ultrasound-guided injections into inguinal lymph nodes at monthly intervals. We have no clean placebo group but the pollen seasons are mostly divided by time. Rhinoconjunctivitis Total Symptom Score, Medication Score and Rhinoconjunctivitis Quality of Life Questionnaire were evaluated before treatment and after each birch and grass pollen season during three subsequent years. Circulating proportions of T helper subsets and allergen-induced cytokine and chemokine production were analyzed by ow cytometry and Luminex. Results: The three groups reported fewer symptoms, lower use of medication and improved quality of life during the birch and grass pollen seasons each year after treatment at an almost similar rate independently of treatment. Mild local pain was the most common adverse event. IgE levels to birch decreased, whereas birch-induced IL-10 secretion increased in all three groups. IgG4 levels to birch and timothy and skin prick test reactivity remained mainly unchanged. Conjunctival challenge tests with timothy extract showed a higher threshold for allergen. In all three groups, regulatory T cell frequencies were increased three years after treatment. Conclusion: Intralymphatic immunotherapy against grass and birch pollen allergy was safe, seemed to be effective, and to be associated with bystander immune modulatory responses.


Background
Nearly 30% of the adult population of Sweden report allergic rhinitis (1). In addition, the prevalence of allergic sensitization is up to 50% in Sweden and in many other European countries (2,3). The total cost of allergic rhinitis in Sweden, with a population of 9.5 million (in 2014), has been estimated at €1.3 billion annually (4). Treating these patients with allergen immunotherapy (AIT) is cost-effective (5). To date, AIT is the only treatment that affects the long-term development of allergic rhinoconjunctivitis. It induces clinical tolerance primarily by stimulating regulatory T (Treg) cells, attenuating T helper 2 (Th2) responses and inducing blocking antibodies (6). Conventional AIT with subcutaneous injections is effective, but consumes time and resources (7,8). Therefore, despite its advantages in allergic rhinitis, AIT has not achieved its full potential. In Sweden, only 15/10,000 receive this treatment, and gures vary worldwide (9) .
AIT improves symptom, medication and combined symptom and medication scores (MS) in patients with allergic rhinoconjunctivitis (10). A randomized placebo-controlled trial with subcutaneous immunotherapy (SCIT) with birch pollen allergen reduced the symptom score by 40% over placebo (11).
In a review article of SCIT and sublingual immunotherapy (SLIT) trials, SCIT was reported to reduce nasal and ocular symptoms by 32-36% compared with placebo, whereas SLIT produced a reduction of 26-36% compared with placebo (12). In an open intralymphatic immunotherapy (ILIT) study, hay-fever patients received only three monthly allergen injections directly into the inguinal lymph nodes and an accumulated dose of only 3,000 standardized quanti ed units (SQ-U) in contrast to approximately 3,000,000 SQ-U with SCIT. Greater e cacy and safety, and faster relief of symptoms were observed after ILIT than after SCIT in an open study (13). The study was followed by a few smaller studies of which ve (14)(15)(16)(17)(18) con rmed the conclusion from the original trial, whereas one showed no bene t of ILIT (19).
Recently we performed a small open pilot study of 10 patients treated with ILIT for birch or grass pollen allergy. We concluded that ILIT was associated with improved quality of life, reduced symptoms, and bene cial immunological changes (18). However, further studies are required to determine whether ILIT can induce clinical effects similar to those of SCIT concerning rhinitis symptoms and how ILIT affects immune responses (20).

Aim
The objective was to evaluate whether ILIT can be an alternative to SCIT and SLIT. We also aimed to determine its safety, e cacy and immune modulatory effects, including changes in spontaneous and allergen-induced cytokine and chemokine production, and proportions of circulating T helper cell subsets.

Study design
A 3-year double-blind randomized clinical trial in 72 patients with rhinoconjunctivitis due to sensitization with birch and grass pollen allergens. The patients were given active treatment with birch or grass in one inguinal lymph node and active treatment with the other allergen or placebo in an inguinal lymph node on the other side. The study was without a clean placebo group as the pollen seasons are clearly separated in Sweden.

Study population eligibility criteria
In all, 126 patients were assessed for eligibility. Forty-four did not meet the inclusion criteria, 7 withdrew consent before treatment and 1 was excluded for unknown reasons (Fig. 1). Fifty-seven patients with allergic rhinoconjunctivitis due to birch and timothy pollen allergens were randomized in 2014 and 17 patients 2015. The 74 patients, including 35 females, were 19-53 years old and had seasonal allergic symptoms to birch and grass (Table 1), whereof 28 were randomized and followed in the Department of Medicine, County Hospital Ryhov, Jönköping. All participants were given ILIT at Allergy Center, University Hospital, Linköping, Sweden. Their skin prick test was >3 mm and displayed IgE to birch and timothy >0.35 kU/L. Exclusion criteria were pulmonary disease, < 80 % of predicted forced expiratory volume at the end of the rst second (FEV1), use of more than 800 µg inhaled budesonide (or equivalent) per day, pregnancy, severe arterial hypertension, autoimmunity, cardiovascular, hepatic, renal, upper airway or metabolic disease, mental incapacity, alcohol abuse, medication interfering with immune response or beta-blockers. From earlier studies we expected 8 out of 9 patients would improve at least 40%. With 40 active treated and 20 in the placebo group, with an alpha of 0.05 the power was calculated to 92%.

Intralymphatic immunotherapy
The patients were randomized into three groups receiving three doses at four-week intervals of 0.1 ml of birch pollen allergen on aluminum hydroxide (10,000 SQ-U/ml; ALK-Abelló, Hørsholm, Denmark) and/or 0.1 ml of 5-grass pollen allergen on aluminum hydroxide (10,000 SQ-U/ml; ALK-Abelló, Hørsholm, Denmark. 5-grass is a mix of equal amounts of SQ-U of Alopecurus pratensis (meadow foxtale), Dactylis glomerata (cocks's foot), Festuca pratensis (meadow fescue), Lolium perenne (English ryegrass), and Phleum pratense (timothy). Each allergen dose was 1.000 SQ-U. A diluent from ALK was used as placebo. Thus the paricipants received two injections, one in each groin on three occasions. Patients were randomized in blocks of six, facilitated by Forum Östergötland. An unblinded nurse prepared and marked each syringe with a label providing randomization number, injection number and injection site. ILIT was administered by three clinicians (LA, PR, and UN). Ultrasound-led technique was used whereby a lymph node was punctured with a 27G (0.4 x 40 mm) needle. Histamine-1 blocker desloratadine tablet 5 mg was given 15 minutes prior to the injections.

Primary outcome measures
Symptoms and drug consumption were primary outcome measures. Symptoms were validated based on the rhinoconjunctivitis total symptom score (RTSS) questionnaire (21). Drug consumption was measured using an MS questionnaire (see Additional le 1). The RTSS and MS were recorded by the patients at the end of the birch pollen season (approximately June 1 st ) and at the end of the grass pollen season (approximately September 1 st ) before treatment and for the following three and four year's altogether.
The birch-and grass pollen seasons are quite separate in Sweden. Also, when we planned the study, no daily symptom score was recommended.

Safety assessment
Safety was assessed as the recording of adverse events from the time of the rst injection to three years after the last injection. A research nurse called the patients to assess adverse events during the rst 5 days after each injection. Safety laboratory parameters were assessed at screening, after the third ILIT injections and after the rst pollen season following treatment (Table 2).

Secondary outcome measures
Effects on quality of life were measured using the rhinoconjunctivitis quality of life questionnaire (RQLQ) (22) recorded as RTSS and MS, see above. Skin prick test reactions (Soluprick SQ Birch and Timothy, ALK-Abelló), allergen-speci c IgE and allergen-speci c IgG4 levels were analyzed (ImmunoCAP ThermoFisher, Uppsala, Sweden) before ILIT and in the fall the following three years. Conjunctival allergen provocation tests (CAPT) (23) were performed with timothy (Aquagen SQ Timothy, ALK-Abelló) before treatment and after the rst pollen season after treatment. Due to lack of extract from the company planned CAPT were not performed after the third pollen season ( Table 2).

Immune laboratory methods
Flow cytometry was used to analyze the CD4+ Th cell population in whole blood from the patients at randomization, and one and three years after completed ILIT. Peripheral blood mononuclear cells obtained from the patients at randomization and one year after ILIT were stimulated in vitro with birch and timothy allergen. Levels of IL-5, IL-10, IL-13, IFN-g, CCL17 and CXCL10 were quanti ed using Luminex. For logistic reasons immune tests were only analyzed from the 45 participants from the Allergy Center, University Hospital, Linköping, Sweden. For detailed methods, experimental protocols, and statistical analyses, see the Methods section in this article's supplementary information.

Statistics
Descriptive statistics for RQLQ, RTSS and MS are presented in medians and percentiles (p25 and p75), and in the graphs with medians and 95% con dence interval. Paired comparisons over time were calculated with Friedman's test and adjusted with the Bonferroni correction for multiple comparisons. Descriptive statistics for IgE, IgG4, SPT and CAPT are presented in mean values and standard deviation (SD). Paired comparisons over time were calculated with repeated measures ANOVA with Bonferroni con dence interval adjustment. The answers to the 28 questions in RQLQ were explored with an item analysis, rendering a Cronbach's Alpha at 0.933; thus the changes within the different domains of RQLQ were consistent. All analyses above were performed in SPSS version 25 (IBM Corp. Armonk, NY, USA).
All ow cytometry, cytokine and chemokine data were analyzed using GraphPad Prism, version 8.3.1 (GraphPad software, Inc., La Jolla, CA, USA), and non-parametric tests were used. Comparisons at the different time points within the treatment groups were calculated using paired Wilcoxon signed ranks test. Unpaired Mann Whitney U test was used to compare differences between the treatment groups at the different time points. The signi cance level was set at p < 0.05.

Results
Seventy-four patients were randomized to ILIT with three doses of 0.1 ml of birch-and 5-grass pollen allergen extracts, or either and placebo at monthly intervals. One patient withdrew consent after the rst injections and one was lost to follow-up after two injections. Seventy-two patients received all injections (Fig. 1). One patient was lost to follow-up two years after treatment and his RQLQ and RTSS had been halved and the patient had not required or used anti-allergic medication. Another patient was lost to follow-up three years after ILIT but two years after treatment the patient showed no improvement in RTSS, MS or RQLQ. Hence, 70 patients remained for analyses three years after treatment.

Symptoms and medication.
The symptoms measured by the RTSS and MS were signi cantly reduced three years post ILIT regardless of active allergen during the birch and grass pollen seasons. The reduction was already evident during the rst season after ILIT, and the effect was sustained throughout the following three years. When combining all three groups, RTSS was reduced from 12.2 to 7.4 (-39%) and from 11.2 to 6.5 (-42%) three years after treatment during the birch and grass pollen seasons, respectively (p<0.05, Fig. 2c-d, see Additional le 4). When combining MS data from all three treatment groups together, it was signi cantly reduced from 9.4 before ILIT to 4.9 (-48%) and from 8.6 to 4.4 (-49%) three years after treatment during the birch and grass pollen seasons, respectively (p<0.01, Fig. 2e-f. For levels of RTSS and MS see Additional le 4). There were no sex-related differences in improvement (data not shown). The patients receiving ILIT in 2014 and 2015 responded almost in the same way (data not shown).

Adverse events
A total of 438 injections were given, of which 285 had an active substance containing allergen extract. Mild local pain at the injection site was the most common adverse event (AE) ( Table 3). However, on three occasions, patients recorded severe pain from ILIT. One patient had moderate breathing problems without any fall in peak ow 30 minutes after the second ILIT (birch and 5-grass) and received the third treatment without any AEs. One patient experienced breathing problem two hours after physical activity, four days after the rst injections, and was relieved with salbutamol inhalations, antihistamine and oral corticosteroids. The remaining injections followed without breathing problems. During the follow-up after ILIT, nine patients reported severe AEs, diverticulitis, miscarriage, burn injury, disc hernia, abdominal pain, hysterectomy, concussion, chest pain, and heart failure but none was judged to be related to ILIT. One patient showed hypothyroidism with elevated levels of antithyroid peroxidase, and thyroid-stimulating hormone receptor antibodies diagnosed in 2018, possibly related to the therapy given in 2014.

Health-related quality of life
The impact on health-related quality of life as measured by RQLQ was signi cantly reduced during the birch and grass pollen seasons, regardless of active ILIT with birch, 5-grass or both allergens. When combining all three groups, the RQLQ score was signi cantly reduced from 3.15 to 1.50 (-52%) and from 2.82 to 1.25 (-56%) three years after treatment during the birch and grass pollen seasons, respectively (p<0.01, Fig. 2a-b, see Additional le 4).
Conjunctival allergen provocation tests, IgE, IgG4 and skin prick tests CAPT with timothy were performed before ILIT and the rst year post ILIT, and showed a higher tolerance threshold in patients receiving ILIT with birch and 5-grass (p<0.05) and 5-grass and placebo (p<0.05), but not signi cant in patients who received ILIT with birch and placebo (p=0.19, see Additional le 4). The IgE levels to birch decreased from baseline to three years post ILIT from 25.45 to 17.71 kU/L (p<0.01 in the group treated with birch and placebo; in the group treated with 5-grass and placebo levels went from 33.18 to 24.42 (p<0.01). The levels of IgE to timothy decreased after ILIT with birch and placebo from 16.60 to 11.00 kU/L (p<0.05); similar, but non-signi cant, differences were determined after ILIT with 5grass and placebo. IgE levels to both birch and timothy decreased slightly after ILIT with both birch and 5grass. There were no signi cant decrease in the levels of total IgE (see Additional le 4). Levels of IgG4 antibodies to birch and timothy remained unchanged in all three treatment groups, except for IgG4 levels to timothy, which increased from in mean 0.36 to 0.44 mg/L (p<0.05) after ILIT with birch and 5-grass (see Additional le 4). Skin prick test for reactivity to birch and timothy allergens remained unchanged during the study period (see Additional le 4).

Circulating T helper cell subsets
Flow cytometry data revealed that the proportion of Th1 cells, de ned as CD3 + CD4 + CD45RA -Tbet + cells, decreased between baseline and three years after treatment in the groups receiving ILIT with birch and placebo and 5-grass and placebo (p<0.05, Fig. 3a-b). An increase was observed after one year in the group receiving birch and 5-grass treatment (p<0.05, Fig. 3c). The proportion of Th2 cells, de ned as CD3 + CD4 + CD45RA -GATA3 + cells, increased three years after treatment in the group receiving birch and placebo ILIT (Fig. 3d, p<0.01) and in the group receiving both treatments (Fig. 3f, p<0.05). No other changes were observed in the group that received 5-grass and placebo ILIT (Fig. 3e). The proportion of Th17 cells, de ned as CD3 + CD4 + CD45RA -RORC + cells, decreased from baseline to three years post ILIT, independent of treatment ( Fig. 3g-i, p<0.05). A signi cant reduction in Th17 cell frequencies was also observed from one year to three years after ILIT (p<0.01). The Th17 memory population was signi cantly higher at baseline in the birch and placebo group than in the other treatment groups (p<0.001, see Additional le 7, a). One year after treatment, a higher proportion of Th17 memory cells was observed in patients receiving 5-grass and placebo ILIT than in the birch and placebo group (p<0.05, ee Additional le 7, b). No signi cant differences between the treatment groups were observed at three years (Fig. S2c in the Supplementary information). The proportion of CD4 dim CD25 hi FoxP3 + Tregs (Fig. 4a-c) and activated Tregs, de ned as CD3 + CD4 + CD45RA -FoxP3 ++ cells (Fig. 4d-f), signi cantly increased between baseline and three years, independent of treatment. In contrast, the proportion of resting Tregs, de ned as CD3 + CD4 + CD45RA + FoxP3 + cells, was not affected by ILIT (data not shown).

Allergen-induced cytokine and chemokine production
At baseline and one year post ILIT, peripheral blood mononuclear cells were harvested and stimulated with birch and grass allergen in vitro to measure cytokine and chemokine secretion (see Additional le 5 and 6). Both birch and grass allergens induced an increase in IL-5 production after birch and placebo ILIT (p<0.05, Fig. 5a and b, while IL-5 secretion did not change in the other treatment groups. Increased birch allergen induced IL-10 secretion was also observed after birch and placebo ILIT and 5-grass and placebo ILIT, whereas no signi cant change occurred in the birch and 5-grass group (Fig. 5c). ILIT did not affect grass-allergen IL-10 production (Fig. 5d). The spontaneous production of the CCL17 chemokine decreased after 5-grass and placebo ILIT and birch and 5-grass ILIT, but not after birch and placebo ILIT (p<0.05 Fig. 5e).

Discussion
The current study is the largest double-blind randomized clinical ILIT trial to date. Patients with hay fever due to sensitization to both grass and birch pollen allergens received 5-grass and birch ILIT, 5-grass ILIT (with birch placebo), or birch ILIT (with grass placebo), thus with "active allergen placebo" as suggested in the ARIA-GA 2 LEN statement 2011 (24). The study revealed statistically signi cant clinical e cacy of ILIT in all three treatment groups. Participants reported improvements in RTSS, RQLQ as well as reduced MS in the rst pollen season following ILIT. The effects were sustained throughout the following three follow-up seasons. We expected that the clinical e cacy of ILIT would mostly be observed for the targeted allergen, i.e. that patients receiving birch ILIT would not respond well during the grass pollen season and vice versa for patients with 5-grass ILIT. Surprisingly, patients receiving birch ILIT or 5-grass ILIT reported improved clinical symptoms during pollen seasons for which they had received no treatment. The clinical data were accompanied and supported by immunological changes such as an increase in the proportion of Treg cells and birch allergen included secretion of IL-10 after the treatment, as well as decreased spontaneous production of the Th2 chemokine CCL17. In contrast, the increase in the number of Th2 cells and the increased level of allergen-speci c IL-5 production after birch ILIT were unexpected. As clinical improvement was sustained throughout our study, we suggest the results are not only due to a placebo effect, although placebo effects may be very strong in AIT-clinical trials (25). The clinical effect did not seem to be mediated by allergen-speci c IgG4, as these antibody levels were not clearly elevated between pre-ILIT baseline and any of the time points thereafter. AIT e cacy has, however, often been reported to correlate with increased IgG4 in SCIT and SLIT (6). Moreover, in previous grass-or birch-pollen ILIT studies, only moderate or no changes in IgG4 were determined (15)(16)(17)(18)(19), whereas a signi cant increase in allergen-speci c IgG4 was determined after ILIT with cat dander allergen (26). In contrast, the bene cial clinical responses correlated with an increase in the frequency of Tregs and increased secretion of IL-10. This potentially inhibitory bystander effect of Tregs and IL-10 might explain why the improved clinical outcome was not dependent on the ILIT allergen.
The birch pollen counts in south east Sweden in 2014 and 2018 were higher than in 2015, 2016, and 2017 (see Additional le 8). As fty-ve study subjects received ILIT in 2014 and seventeen received ILIT in 2015, differences in pollen count were leveled out. Patients who received ILIT in 2015 compared well in the following seasons to patients that received ILIT in 2014. Recall bias is limited as the patients recorded their RQLQ, RTSS and MS directly after the birch and grass pollen seasons, which are quite separate in Sweden, after enrollment, before ILIT and three years after, i.e. four years in all. Of seventy-two patients receiving ILIT, only two was lost to follow up after three years.
Most of the AEs were judged as mild or moderate, generally at the allergen injection site (Table 3). These are also a normal reactions after SCIT (27,28). However, whereas patients receiving ILIT in the current study reported between none to three AEs, the number of AEs after SCIT can be up to 40-50. In a large study that analyzed AEs in 1,700 patients who had received SCIT, systemic AEs were reported in 3.3% of the injections (28). Edema and pruritus at the injection site, ush, urticaria, wheezing, dyspnea, eye pruritus, headache, and abdominal pain are common (1-10%) or very common (>10%) with SCIT, whereas oral pruritus, oral edema, rhinitis, headache, ear pruritus, throat irritation, asthma, abdominal pain, urticaria, and fatigue are common or very common with SLIT (29). ILIT seems to give AEs similar to those with SCIT, but because ILIT only needs three injections, the AEs can be reduced by up to 90% compared with SCIT.
One limitation of this study is that it had no true placebo group because all the participants received ILIT against, at least, one pollen. This was for ethical reasons as we did not want any patient with rather severe ARC to be without immunotherapy for four years. The approach with "active allergen placebo" was suggested in an ARIA-GA 2 LEN statement in 2011 (24) and supported by a recent SCIT study where dualallergic patients were randomized to receive either grass or birch SCIT. Here, targeted but not untargeted allergen rhinoconjunctivitis symptoms were reduced in contrast to our results (30).
AIT renders signi cant improvements in rhinoconjunctivitis and conjunctival sensitivity that persist at least seven years after termination of treatment (7). In addition, AIT can prevent development of asthma (31,32) and new sensitizations in mono-sensitized patients (12,33). 5582% of patients were reported to abandon the treatment before completing the recommended course of therapy (34). With only three injections over eight weeks, ILIT may overcome the disadvantages of long duration of treatment and poor compliance.
The results of this study add to the hitherto positive studies and suggest that ILIT is effective and safe as treatment for pollen allergy. We will follow the patients in this study for symptoms, health-related quality of life and adverse events in an open study until 2024 (EPN 2017/302-31).

Conclusions
ILIT is an opportunity to make AIT more easily accessible to patients at a lower cost and less risk. There is a need for further studies to establish the optimal dose for e cacy and side effects (20