y^j?r   (jiuv*stA


   THE BIOLOGICAL CLASSIFICATION OF
        INFLUENZA BACILLI

        By T. M. Rivers, M. D.,
          Baltimore, Md.

{From the Department of Pathology and Bacteriology, The Johns
         Hopkins University) 
 
[From The Johns Hopkins Hospital Bulletin, Vol. XXXI, No. 348,
                       February, 1920]
      THE BIOLOGICAL  CLASSIFICATION  OF
                 INFLUENZA BACILLI

                 By T.  M. Rivers, M. D.,
                      Baltimore, Md.
{From the Department of Pathology and Bacteriology,  The Johns
                    Hopkins University)

   Since the discovery  of B. influenzce medical literature has [50]
become burdened  with discussions about its  relationship to
epidemic and sporadic influenza, pneumonia, pertussis, diph-
theria, measles, scarlet fever,  conjunctivitis, and meningitis,
whether  it is  really  hemoglobinophilic,  and  whether  all
strains  are alike  morphologically, culturally and  immuno-
logically.   It is very interesting  reading, but, after  all, the
answer to many questions seems no nearer than it was almost
30 years ago.  A brief review is necessary to show the exact
status of B. influenzce  and the difficulties presented  in under-
taking a study of this kind.
   Pfeiffer1 described  a small  Gram-negative  bacillus  in
smears from the pharynx and the sputum of influenza patients.
Later he cultivated the bacillus, finding hemoglobin essential
for its growth on artificial media, and further characterized
it as non-motile, aerobic, not found in the blood  of influenza
patients, and not  very pathogenic for  animals.  Afterwards
he found in three  cases of bronchopneumonia following diph-
theria in children a bacillus in every respect the  same as the
one spoken of as the true influenza bacillus, with the exception
that it was larger and developed many thread-like forms.
This organism  he called the pseudo-influenza bacillus.  The
differentiation was made on  morphology alone.   Wolff,2 in
1903 stated that Pfeiffer believed all influenza bacilli were the
same, and that his original idea of a pseudo-influenza  bacillus
was incorrect.
(D 
[5<>]  Cantani * thought that the pseudo-influenza bacillus and tin'
    true one were identical, and that neither of them were hemo-
    globinophilic, as they could be grown on media, enriched with
    spermatic fluid, which did not give the spectroscopic band of
    hemoglobin.   Ghon and Prevss * considered hemoglobin neces-
    sary even though it were present in such small quantities  that
[511 it failed to give a band with the spectroscope unless hydrazin
    were  added.   Xeisser " was able to grow an  influenza-like
    bacillus, isolated from  a  case  of purulent conjunctivitis, on
    plain agar for  20 generations in symbiosis with a  xerosis
    bacillus.  Davis'  thought hemoglobin, acting as a catalytic
    agent, was necessary for growth, and showed that a very small
    amount was  required (1  part in 180,000  parts of medium).
    According to him, growth will take place in  the  presence of
    coagulated hemoglobin, but if the hemoglobin  be broken up
    by excessive   heating into  hematin  and  globin,  no growth
    occurs.
       The difficulties which arise in the exact identification of an
    influenza bacillus have caused many mistakes  in the past  and
    probably will continue  to cause them.   Spengler,7 Luzzatto"
    and Jochmann and Krause" have described at different times
    a bacillus which they considered the cause of pertussis.  It is
    now believed  they were dealing with the  group of influenza
    bacilli and not  the organism later described  by Bordet and
    Gengou.
       Gram-negative hemoglobinophilic bacilli have been recov-
    ered from different parts of the human body, also from ani-
    mals, under  a variety of  conditions.  Rosenthal10 considered
    them ordinary members of the mouth flora without pathogenic
    significance.   Auerbach,11 Davis,12  and many others have  iso-
    lated them from  the throats of patients  with  measles, per-
    tussis, diphtheria and scarlet fever.  Klieneberger u described
    a  pseudo-influenza bacillus which he isolated from pus in a
    gall-bladder.   Cohn14 recovered B. influenzce from a case of
    acute urethritis.   It is  generally known that these organisms
    may  cause  sinusitis, pneumonia,  septicemia, endocarditis,
    arthritis, otitis media, and meningitis.  Davis " isolated from
    the urine of three patients a small, Gram-negative, hemolytic,
    anaerobic, hemoglobinophilic bacillus.   Moon"  recovered a
(2) 
small,  Gram-negative anaerobic,  hemoglobinophilic bacillus
from a chronic ethmoid infection.   Friedberger,17 working in
Pfeiffer's clinic,  isolated from the  preputial secretions of a
dog  a  small,  Gram-negative, non-motile, hemoglobinophilic
bacillus.   Wolff,2 working in the same  clinic, isolated from
the lungs of a  rat  a  bacillus  that remained  exactly  like
B. influenzce for three months, after which time it acquired
the property of growing on ordinary media.  Pritchett and
Stillman18  have recovered  from throat  cultures  a   Gram-
negative, non-motile, aerobic, hemolytic, hemoglobinophilic
bacillus  which makes milk alkaline  (growth takes place when
a little blood is added to the milk).
   In 1905  Wollstein,19 working on  pertussis, isolated  bacilli
considered as belonging to the influenza group, but from which
they could  be differentiated by agglutination and absorption
tests.  In 1906,20 not feeling so positive about this difference,
she says:  " The similarity  of cultural characteristics of all
the influenza bacilli has been emphasized by Neisser, and my
experience  with  the  agglutination reactions leads  me to re-
gard all strains as belonging to one family."
   Meunier21 (1897)  reported 10 cases of broncho-pneumonia
in young children caused by B. influenzce. In several instances
the bacillus was recovered from the blood both before and after
death.   At  this time, contrary to Pfeiffer's statement, it was
noted that these organisms were highly pathogenic for rabbits.
 Slawyk,22 in 1899, isolated an influenza bacillus from the blood
and spinal  fluid of a child with meningitis.  This organism
was seen by Pfeiffer who  agreed that it was  B.  influenzce.
Cohen23 (1909)  recovered from the spinal  fluid  and blood of
patients with meningitis an  organism similar to  the influenza
bacillus, except that  it  caused septicemia in rabbits, and be-
cause of this difference and evidence obtained by  protection
experiments, he was inclined to believe that the two  organisms
were not the same.  Thursfield,24  in 1910, reported two cases
of  B.  influenzal septicemia  without meningitis.   He  agreed
 with Cohen and concluded by saying,  " Organisms hitherto
described as B.  influenza} are not all identical, but like the
coli-typhoid family, belong to a group the various members of
which possess very different pathogenic  powers."  AVollstein,25,
                             (3) 
[51] in 1911. working with respiratory and meningeal strains, con-
    sidered them identical, and varying only in virulence.  Davis."
    in 1911.. found  no difference in  the virulence of respiratory
    and meningeal strains.  Wollstein,27 in 1915, still regarded all
    influenza bacilli as more  or  less  identical regardless of their
    own origin or virulence.  Davis,12 in 1915, suggested dividing
    them into two groups, one showing the phenomenon of symbi-
    osis, the other lacking it.  Williams,28 in 1919, working with a
    number  of strains reported  a few  crosses by agglutination,
    but none by absorption tests.  Huntoon  and Hannum,2* in
    1919, say, '* We have found  no strains among our collection
    which do not show relationship  either directly or indirectly
    through absorptions."  Gay  and Harris,30 in 1919,  found in
    their  serological  work  on influenza evidence  that  influenza
    bacilli probably could be divided into groups.
      In  spite of the vast amount of work  done, very little is
    known  about B.  influenzce and its  biological activities.   At
    present it is described as a small, (J ram-negative, aerobic, non-
    motile,  hemoglobinophilic bacillus.   For  all  that is known,
    there  may be a number of different kinds of bacilli answering
    that description, or there may be only one.  In reality there is
    one true B. influenza' existing in name only, and  that is the
    first one grown and described by Pfeiffer, as he did not and
    could not prove  any of the subsequent strains to be  identical
    with the first.
      Jordann has  made  the best  contribution  lately to our
    knowledge of influenza bacilli by showing that 10 of 13 strains
    formed indol.  Owing to this discovery a study of the cul-
    tural  characteristics of different  strains of B. influenzce was
    undertaken.  The work has  not been completed, and what
    follows is merely a preliminary report made at this time with
    the hope that others will become interested and assist in the
    solution of the problem.
      The bacilli for  study were obtained from  normal throats
    since  the epidemic, 32 strains; from influenza meningitis, 5
    strains  (4 coming from Dr. Howland's clinic this  year and 1
    isolated by Dr. Wollstein in 1917) ; and  from cases of epi-
    demic influenza,  14 strains (supplied by  Drs. Parker, Woll-
    stein  and Stillman).   Two strains of B.  pertussis, one  from
                                (4) 
 
<s
-»     %  a
ft
4  s
 
Dr. W. W. Ford, the other from the X. Y. Board  of Health  [51]
were studied at the same time for comparison.  In isolating the  [52]
organisms  from normal throats, all Gram-negative,  aerobic,
hemoglobinophilic bacilli were included in  the series regard-
less of morphology and the kind of colony formed, since these
latter vary so much under different conditions that they are
unreliable criteria.
   In a study of this kind it is necessary to be able to obtain
at all times a  vigorous growth of the organisms both on  solid
and in liquid media and  describe their  colony formation,
morphology and biological activities under as favorable con-
ditions as possible.   Meat-infusion agar and broth,  Ph 7.5,
with from 1 to 2 per cent  rabbit blood, furnish good media
for growth and to which other ingredients  may be added for
study.
   All the strain are hemoglobinophilic after at least  two
months of artificial cultivation, except in a few instances, in
which they were grown on hemoglobin-free  media  in  sym-
biosis with other bacteria for several generations.  They are
Gram-negative, non-motile,  do  not liquefy blood-gelatin, do
not ferment  glucose, lactose, saccharose, maltose, mannite,
salicin, inulin, glycerol  and  xylose in one week, and, most of
them, including the three hemolytic ones, have at some time
shown the phenomenon  of symbiosis.  Three strains hemolyse
 rabbit blood both in liquid and in solid media.
   In working with the morphology and colony formation, it is
necessary to grow the various strains side by  side, generation
after generation, on exactly the same medium, and then it is   -
possible to notice in a general way differences between them.
 Fig. 1  (magnification  8)  shows  colonies of four different
 organisms 36 hours old, grown  under the conditions named
 above, the upper row as seen by direct light, the lower by top
 light.  No. 1 forms  indol, No. 2 forms amylase, No. 3 is hemo-
 lytic, No. 4 is B. pertussis.   Nos. 1 and 3 are moist, become
 very granular and  dark in the center,  and when  one to five
 clays old put out daughter colonies.  No. 2 is tough, holds its
 shape, and puts out very few daughter  colonies.  All pit and
 turn  the  medium  a dirty  brownish  color.   All are  tan-
 colored after  -18 hours.   Older observers attributed this color
(5) 
[52] to the hemoglobin in  the medium,  but this is imlikely as it
    occurs when they  are grown  on hemoglobin-free  media  in
    symbiosis with other bacteria.   Figs. 2 and  3  (magnification
    20)  are copies from Pfeiffer1 and Grassberger32 showing the
    type of colonies with which they were working.
      When a solid medium is inoculated with the various strains,
    there is seen a  distinct and constant difference between the
    edges formed.   Cultures of B. pertussis  (Fig. 4)  have ele-
    vated, smooth edges;  some of the influenza bacilli  (Figs. 5
    and 6), elevated, lobate edges; others (Figs. 7 and 8), slightly
    elevated, finely irregular edges; and others which grow very
    delicately have  no sharply  defined  edges.  (See  lower  right
    corner of Fig. 9 which also shows two other  types of cultures
    mentioned above,  and  finally B. pertussis in the upper  right
    corner).   Smears  from cultures with the lobate edges usually
    show the typical small bipolar-staining bacilli  called the true
    influenza bacillus, whereas  smears from the ones with finely
    irregular  edges  and no sharply defined edges usually  show
    large, more easily stained bacilli,  bizarre shapes, and thread-
    like  forms.   This  difference  is  well  shown  by  copies   of
    Pfeiffer's1 true  (Fig. 10) and  his pseudo-influenza  (Fig. 11)
    bacilli.
       One of the first  things noticed when working with influenza
    bacilli was that some strains both on solid and in liquid media
    soon  developed a  perfectly characteristic fresh  fecal  odor.
    Later, when cultures of the different strains were tested for
    indol formation, it was found that  the  ones which produced
    this  odor gave positive tests for indol when the ether extracts
    were layered with Ehrlich's reagent. Cultures have been dis-
    tilled and  ether extracts of the water-clear distillate gave the
    same results as were  obtained before distillation.  Repeated
    tests have  been  made  on the various strains grown from two
    days to two weeks, and indol has been produced constantly  by
    some,  and not,  just as constantly, by others.   Thirty of the
    51 strains  form indol  and the  five meningeal ones are in the
    positive group.  The spinal fluids of patients with influenza
    meningitis do not give a positive test when first  drawn, but,
    if a tube of it be placed in an incubator for from 12  to  24
                                16) 
hours, a definite test is obtained.  The fluid from a meningo- [52]
coccus meningitis gave a negative test when thus treated.
  The ability of the various influenza bacilli to form amylase
has been studied in the following manner.  Defibrinated rabbit
blood was added to melted meat-infusion agar which was kept
at 95° C. long enough to destroy the amylase present  in the
blood. Then to 100 c. c. of this mixture were added from 10
to 15 c. c. of a 2 per cent sterile soluble starch solution.  Plates
were  poured, allowed to cool and inoculated at various  points
with  different  influenza bacilli, and after from  3 to 5 days
incubations were covered with  a  weak Lugol's solution.  If
the starch were not split, the medium was a dark blue up to
the edge  of  the cultures, but if it were changed, the  iodine
reaction was absent  around them, a colorless zone  of from
4 to  8 mm.  in width being left.  A positive and a  negative
reaction are  shown  in Fig. 12.   Nine strains produce amylase
in small amounts, and none of these produce indol.
   Tubes  of  potassium nitrate blood-broth  were inoculated
with  influenza bacilli, incubated for 5  days, and then  tested
with  the  sulphanilic acid  and  naphthylamin  reagent for
nitrites.   Thirty-three of the  51  strains have at some time
given a positive reaction.   Some  always reduce  the nitrates,
others do so  irregularly, while still others have never done so.
This characteristic  is displayed by some  indol formers,  by
some amylase formers and by two  of the three hemolytic ones.
Further work may show that all can reduce nitrates under the
proper conditions, making this a common characteristic of the
whole group, just as hemoglobin is essential for their growth.
   Flasks  of 100 c. c. of fat-free milk with brom-cresol purple
for an indicator, and flasks of 100 c. c. meat-infusion broth
with  the same indicator were autoclaved separately.  (If auto-
claved together a precipitate forms which interferes.)  After
cooling,  equal quantities of the  milk and the  broth were
mixed and from 1 to 2 per cent rabbit blood was added.  This
is a  good'medium for the growth of  influenza  bacilli, and
accurate readings can be made  when  an uninoculated tube is [53]
kept  for  a control.  When these tubes of  blood-broth-milk
were inoculated and  incubated, it was obvious very  soon that
all influenza bacilli  are not the same culturally,  as  some,
(7) 
within 48 hours, made the milk slightly but definitely acid.
nthers slightly but definitely alkaline, while still others gave
doubtful results.  This difference persisted for a week.  Read-
ings made at intervals longer than a week cannot be given at
this  time.  The  source of the acid  is unknown, since lactose
is not fermented.

                       DISCUSSION
  Many  people  have never felt absolutely certain that the
differences  between B. pertussis and B. influenzce were sharp
enough to be beyond doubt in  spite of the serological  proof.
B. pertussis, after  a period of artificial cultivation,  can be
grown on plain media, forms no indol, no nitrites and  makes
milk  very alkaline.  Some of  the influenza bacilli  also form
no indol and no nitrites, but none of them has ever made milk
nearly as alkaline as B. pertussis.
  Time alone will tell whether these cultural characteristics
will  be constant.  While there  are differences in the biological
activities of the various strains of B.  influenzce, at the same
time  there are definite groups, the  individual members of
which are similar culturally.  Only one group will be discussed
at this time.  It consists of 10 strains five from  the  spinal
fluid of patients  with influenza meningitis, two  epidemic
strains from New York, and three from normal throats.  The
growth and morphology of these arc similar, all form indol,
all  reduce  nitrates  to nitrites,  and make  blood-broth-milk
slightly acid within 48 hours.
  Whether the strains of large bacilli  that are amylase form-
ers,  or  the hemolytic ones, should be included in  this big
influenza group is a question  that  will  have to be  decided.
The nine amylase  formers and the three hemolytic strains
have characteristics  in common with the  big group, as shown
by certain  ones forming nitrites and by one of the hemolytic
strains forming both indol and nitrites.  Possibly this is a big
group of organisms, like the  streptococci, which  have  been
divided  into hemolytic and non-hemolytic strains, and fur-
ther  subdivided  by cultural  characteristics  and  serological
tests.  Possibly  the group can  be compared  with  the  Gram-
negative diplococci, Meningococcus, Parameningococci, Micro-
                            (8) 
 
THE JOHNS HOPKINS HOSPITAL BULLETIN, FEBRUARY

■*~
t^^^*""*^
}$r>i


5f» •£
SVvi
Fig. 2.
1920
PLATE XXVM
Fig. 3.
?-ti
-,-N7*\
 
2*c
f/ «R
4j>

£-*L
/
fWJ^1

1h«
*&*** ^
 -J2&
.1  yfp-'^
 
 
coccus catarrhalis, Micrococcus flavus, Micrococcus pharyngis [53]
siccus, Gonococcus and others.

                       CONCLUSIONS

   1. The   Gram-negative,  non-motile,   hemoglobinophilic
bacilli can be classified biologically by reactions which admit
of subdivisions of the group.
   2. In working with a  suspected B. influenzae, the following
routine should be followed:
     (a) Determination  of hemoglobinophilic qualities.
     (b) Colony formation.
     (c) Hemolytic  test.
     (d) Gram stain.
     (e) Morphology.
     (f) Motility.
     (g) Indol formation.
     (h) Eeduction  of nitrates to nitrites.
     (i) Amylase formation.
     (j) Eeaction in blood-broth-milk.
   3. B. pertussis  can be differentiated  from  the  group of
B. influenzce by cultural characteristics.

                       BIBLIOGRAPHY
   1. Pfeiffer, R.: Ztschr.  f. Hyg., 1893, XIII, 357.
   2. Wolff, A.:  Centralbl. f. Bakteriol., 1903,1; Orig., XXXIII, 407.
   3. Cantani, A.:  Ztschr. f. Hyg., 1901, XXXVI, 29.
   4. Ghon, A., and v.  Preyss, W.:  Centralbl. f. Bakteriol., 1902, I;
Orig., XXXII, 90.
   5. Neisser, M.:  Deutsche med. Wchnschr., 1903, XXIX, 462.
   6. Davis, D. J.:  Jour. Infect. Dis., 1907, IV, 73.
   7. Spengler, C:   Deutsche med. Wchnschr., 1897, XXIII, 830.
   8. Luzzatto, A.:   Centralbl. f.  Bakteriol., 1900, I;  Orig., XXVII,
817.
   9. Jochmann, G., and Krause, P.:  Ztschr. f. Hyg., 1901, XXXVI,
193.
   10. Rosenthal, G.:  Comp. rend. Soc. de biol., 1900, LII, 266.
   11. Auerbach, M.:   Ztschr. f. Hyg., 1904, XLVII, 259.
   12. Davis, D. J.:  Jour. Am. Med. Assn., 1915, LXIV, 1814.
   13. Klieneberger, C:   Deutsche med. Wchnschr., 1905, XXXI,
575.
   14. Cohn, P.:  Ibid., 1152.
   15. Davis, D. J.:  Jour.  Infect. Dis., 1910, VII,  599.
(«) 
  16. Moon, V. H.:  Tr. Chicago Path. Soc, 1913,  IX, 51.
  17. Friedberger, E.:  Centralbl. f.  Bakteriol., 1903, I;  Orig.,
XXXIII, 401.
  18. Pritchett, I. W., and Stillman, E. G.:   Jour. Exper. Med.,
1919, XXIX,  259.
  19. Wollstein, M.:  Jour. Exper. Med., 1905, VII, 335.
  20. Wollstein, M.:  Ibid., 1906, VIII, 681.
  21. Meunier, H.:  Arch. gen. de m§d., 1897, I, 129.
  22. Slawyk:  Ztschr. f. Hyg., 1899, XXXII, 443.
  23. Cohen:  Ann.  de l'Inst. Pasteur, 1909, XXIII, 273.
  24. Thursfield, H.:  Quart. Jour. Med., 1910,  IV, 7.
  25. Wollstein, M.:  Am. Jour. Dis. Child., 1911, I, 42.
  26. Davis, D. J.:  Ibid., 249.
  27. Wollstein, M.:  Jour. Exper. Med., 1915, XXII, 445.
  28. Williams, A.:  Arch. Pediat, 1919, XXXVI, 107.
  29. Huntoon, F. M., and Hannum, S.:  Jour. Immunol., 1919, IV,
167.
  30. Gay, F. P., and Harris, D. H.:   Jour. Infect. Dis., 1919, XXV,
414.
  31. Jordan, E. O.:   Jour.  Am. Med. Assn.,  1919, LXXII, 1542.
  32. Grassberger, R.:  Ztschr. f. Hyg., 1897, XXV, 453.
(10) 
 
 
 
tS
r.'.
;a^:'